diff options
Diffstat (limited to 'string.c')
| -rw-r--r-- | string.c | 7151 |
1 files changed, 3738 insertions, 3413 deletions
@@ -23,7 +23,6 @@ #include "debug_counter.h" #include "encindex.h" -#include "gc.h" #include "id.h" #include "internal.h" #include "internal/array.h" @@ -79,25 +78,41 @@ VALUE rb_cString; VALUE rb_cSymbol; -/* FLAGS of RString +/* Flags of RString * * 1: RSTRING_NOEMBED - * 2: STR_SHARED (== ELTS_SHARED) - * 2-6: RSTRING_EMBED_LEN (5 bits == 32) - * 5: STR_SHARED_ROOT (RSTRING_NOEMBED==1 && STR_SHARED == 0, there may be - * other strings that rely on this string's buffer) - * 6: STR_BORROWED (when RSTRING_NOEMBED==1 && klass==0, unsafe to recycle - * early, specific to rb_str_tmp_frozen_{acquire,release}) - * 7: STR_TMPLOCK (set when a pointer to the buffer is passed to syscall - * such as read(2). Any modification and realloc is prohibited) - * - * 8-9: ENC_CODERANGE (2 bits) - * 10-16: ENCODING (7 bits == 128) + * The string is not embedded. When a string is embedded, the contents + * follow the header. When a string is not embedded, the contents is + * on a separately allocated buffer. + * 2: STR_SHARED (equal to ELTS_SHARED) + * The string is shared. The buffer this string points to is owned by + * another string (the shared root). + * 3: STR_CHILLED (will be frozen in a future version) + * The string appears frozen but can be mutated with a warning. + * 5: STR_SHARED_ROOT + * Other strings may point to the contents of this string. When this + * flag is set, STR_SHARED must not be set. + * 6: STR_BORROWED + * When RSTRING_NOEMBED is set and klass is 0, this string is unsafe + * to be unshared by rb_str_tmp_frozen_release. + * 7: STR_TMPLOCK + * The pointer to the buffer is passed to a system call such as + * read(2). Any modification and realloc is prohibited. + * 8-9: ENC_CODERANGE + * Stores the coderange of the string. + * 10-16: ENCODING + * Stores the encoding of the string. * 17: RSTRING_FSTR - * 18: STR_NOFREE (do not free this string's buffer when a String is freed. - * used for a string object based on C string literal) - * 19: STR_FAKESTR (when RVALUE is not managed by GC. Typically, the string - * object header is temporarily allocated on C stack) + * The string is a fstring. The string is deduplicated in the fstring + * table. + * 18: STR_NOFREE + * Do not free this string's buffer when the string is reclaimed + * by the garbage collector. Used for when the string buffer is a C + * string literal. + * 19: STR_FAKESTR + * The string is not allocated or managed by the garbage collector. + * Typically, the string object header (struct RString) is temporarily + * allocated on C stack. */ #define RUBY_MAX_CHAR_LEN 16 @@ -109,54 +124,35 @@ VALUE rb_cSymbol; #define STR_SET_NOEMBED(str) do {\ FL_SET((str), STR_NOEMBED);\ - if (USE_RVARGC) {\ - FL_UNSET((str), STR_SHARED | STR_SHARED_ROOT | STR_BORROWED);\ - }\ - else {\ - STR_SET_EMBED_LEN((str), 0);\ - }\ -} while (0) -#define STR_SET_EMBED(str) FL_UNSET((str), (STR_NOEMBED|STR_NOFREE)) -#if USE_RVARGC -# define STR_SET_EMBED_LEN(str, n) do { \ - assert(str_embed_capa(str) > (n));\ - RSTRING(str)->as.embed.len = (n);\ -} while (0) -#else -# define STR_SET_EMBED_LEN(str, n) do { \ - long tmp_n = (n);\ - RBASIC(str)->flags &= ~RSTRING_EMBED_LEN_MASK;\ - RBASIC(str)->flags |= (tmp_n) << RSTRING_EMBED_LEN_SHIFT;\ + FL_UNSET((str), STR_SHARED | STR_SHARED_ROOT | STR_BORROWED);\ } while (0) -#endif +#define STR_SET_EMBED(str) FL_UNSET((str), STR_NOEMBED | STR_SHARED | STR_NOFREE) #define STR_SET_LEN(str, n) do { \ - if (STR_EMBED_P(str)) {\ - STR_SET_EMBED_LEN((str), (n));\ - }\ - else {\ - RSTRING(str)->as.heap.len = (n);\ - }\ + RSTRING(str)->len = (n); \ } while (0) -#define STR_DEC_LEN(str) do {\ - if (STR_EMBED_P(str)) {\ - long n = RSTRING_LEN(str);\ - n--;\ - STR_SET_EMBED_LEN((str), n);\ - }\ - else {\ - RSTRING(str)->as.heap.len--;\ - }\ -} while (0) +static inline bool +str_enc_fastpath(VALUE str) +{ + // The overwhelming majority of strings are in one of these 3 encodings. + switch (ENCODING_GET_INLINED(str)) { + case ENCINDEX_ASCII_8BIT: + case ENCINDEX_UTF_8: + case ENCINDEX_US_ASCII: + return true; + default: + return false; + } +} -#define TERM_LEN(str) rb_enc_mbminlen(rb_enc_get(str)) +#define TERM_LEN(str) (str_enc_fastpath(str) ? 1 : rb_enc_mbminlen(rb_enc_from_index(ENCODING_GET(str)))) #define TERM_FILL(ptr, termlen) do {\ char *const term_fill_ptr = (ptr);\ const int term_fill_len = (termlen);\ *term_fill_ptr = '\0';\ if (UNLIKELY(term_fill_len > 1))\ - memset(term_fill_ptr, 0, term_fill_len);\ + memset(term_fill_ptr, 0, term_fill_len);\ } while (0) #define RESIZE_CAPA(str,capacity) do {\ @@ -165,33 +161,33 @@ VALUE rb_cSymbol; } while (0) #define RESIZE_CAPA_TERM(str,capacity,termlen) do {\ if (STR_EMBED_P(str)) {\ - if (str_embed_capa(str) < capacity + termlen) {\ - char *const tmp = ALLOC_N(char, (size_t)(capacity) + (termlen));\ - const long tlen = RSTRING_LEN(str);\ - memcpy(tmp, RSTRING_PTR(str), tlen);\ - RSTRING(str)->as.heap.ptr = tmp;\ - RSTRING(str)->as.heap.len = tlen;\ + if (str_embed_capa(str) < capacity + termlen) {\ + char *const tmp = ALLOC_N(char, (size_t)(capacity) + (termlen));\ + const long tlen = RSTRING_LEN(str);\ + memcpy(tmp, RSTRING_PTR(str), tlen);\ + RSTRING(str)->as.heap.ptr = tmp;\ + RSTRING(str)->len = tlen;\ STR_SET_NOEMBED(str);\ - RSTRING(str)->as.heap.aux.capa = (capacity);\ - }\ + RSTRING(str)->as.heap.aux.capa = (capacity);\ + }\ }\ else {\ - assert(!FL_TEST((str), STR_SHARED)); \ - SIZED_REALLOC_N(RSTRING(str)->as.heap.ptr, char, \ - (size_t)(capacity) + (termlen), STR_HEAP_SIZE(str)); \ - RSTRING(str)->as.heap.aux.capa = (capacity);\ + RUBY_ASSERT(!FL_TEST((str), STR_SHARED)); \ + SIZED_REALLOC_N(RSTRING(str)->as.heap.ptr, char, \ + (size_t)(capacity) + (termlen), STR_HEAP_SIZE(str)); \ + RSTRING(str)->as.heap.aux.capa = (capacity);\ }\ } while (0) #define STR_SET_SHARED(str, shared_str) do { \ if (!FL_TEST(str, STR_FAKESTR)) { \ - assert(RSTRING_PTR(shared_str) <= RSTRING_PTR(str)); \ - assert(RSTRING_PTR(str) <= RSTRING_PTR(shared_str) + RSTRING_LEN(shared_str)); \ - RB_OBJ_WRITE((str), &RSTRING(str)->as.heap.aux.shared, (shared_str)); \ - FL_SET((str), STR_SHARED); \ + RUBY_ASSERT(RSTRING_PTR(shared_str) <= RSTRING_PTR(str)); \ + RUBY_ASSERT(RSTRING_PTR(str) <= RSTRING_PTR(shared_str) + RSTRING_LEN(shared_str)); \ + RB_OBJ_WRITE((str), &RSTRING(str)->as.heap.aux.shared, (shared_str)); \ + FL_SET((str), STR_SHARED); \ FL_SET((shared_str), STR_SHARED_ROOT); \ - if (RBASIC_CLASS((shared_str)) == 0) /* for CoW-friendliness */ \ - FL_SET_RAW((shared_str), STR_BORROWED); \ + if (RBASIC_CLASS((shared_str)) == 0) /* for CoW-friendliness */ \ + FL_SET_RAW((shared_str), STR_BORROWED); \ } \ } while (0) @@ -214,27 +210,43 @@ VALUE rb_cSymbol; static inline long str_embed_capa(VALUE str) { -#if USE_RVARGC return rb_gc_obj_slot_size(str) - offsetof(struct RString, as.embed.ary); -#else - return RSTRING_EMBED_LEN_MAX + 1; -#endif +} + +bool +rb_str_reembeddable_p(VALUE str) +{ + return !FL_TEST(str, STR_NOFREE|STR_SHARED_ROOT|STR_SHARED); } static inline size_t -str_embed_size(long capa) +rb_str_embed_size(long capa) { return offsetof(struct RString, as.embed.ary) + capa; } +size_t +rb_str_size_as_embedded(VALUE str) +{ + size_t real_size; + if (STR_EMBED_P(str)) { + real_size = rb_str_embed_size(RSTRING(str)->len) + TERM_LEN(str); + } + /* if the string is not currently embedded, but it can be embedded, how + * much space would it require */ + else if (rb_str_reembeddable_p(str)) { + real_size = rb_str_embed_size(RSTRING(str)->as.heap.aux.capa) + TERM_LEN(str); + } + else { + real_size = sizeof(struct RString); + } + return real_size; +} + static inline bool STR_EMBEDDABLE_P(long len, long termlen) { -#if USE_RVARGC - return rb_gc_size_allocatable_p(str_embed_size(len + termlen)); -#else - return len <= RSTRING_EMBED_LEN_MAX + 1 - termlen; -#endif + return rb_gc_size_allocatable_p(rb_str_embed_size(len + termlen)); } static VALUE str_replace_shared_without_enc(VALUE str2, VALUE str); @@ -265,59 +277,49 @@ rb_str_make_independent(VALUE str) } void +rb_str_make_embedded(VALUE str) +{ + RUBY_ASSERT(rb_str_reembeddable_p(str)); + RUBY_ASSERT(!STR_EMBED_P(str)); + + char *buf = RSTRING(str)->as.heap.ptr; + long len = RSTRING(str)->len; + + STR_SET_EMBED(str); + STR_SET_LEN(str, len); + + if (len > 0) { + memcpy(RSTRING_PTR(str), buf, len); + ruby_xfree(buf); + } + + TERM_FILL(RSTRING(str)->as.embed.ary + len, TERM_LEN(str)); +} + +void rb_debug_rstring_null_ptr(const char *func) { fprintf(stderr, "%s is returning NULL!! " "SIGSEGV is highly expected to follow immediately.\n" "If you could reproduce, attach your debugger here, " "and look at the passed string.\n", - func); + func); } /* symbols for [up|down|swap]case/capitalize options */ static VALUE sym_ascii, sym_turkic, sym_lithuanian, sym_fold; static rb_encoding * -get_actual_encoding(const int encidx, VALUE str) -{ - const unsigned char *q; - - switch (encidx) { - case ENCINDEX_UTF_16: - if (RSTRING_LEN(str) < 2) break; - q = (const unsigned char *)RSTRING_PTR(str); - if (q[0] == 0xFE && q[1] == 0xFF) { - return rb_enc_get_from_index(ENCINDEX_UTF_16BE); - } - if (q[0] == 0xFF && q[1] == 0xFE) { - return rb_enc_get_from_index(ENCINDEX_UTF_16LE); - } - return rb_ascii8bit_encoding(); - case ENCINDEX_UTF_32: - if (RSTRING_LEN(str) < 4) break; - q = (const unsigned char *)RSTRING_PTR(str); - if (q[0] == 0 && q[1] == 0 && q[2] == 0xFE && q[3] == 0xFF) { - return rb_enc_get_from_index(ENCINDEX_UTF_32BE); - } - if (q[3] == 0 && q[2] == 0 && q[1] == 0xFE && q[0] == 0xFF) { - return rb_enc_get_from_index(ENCINDEX_UTF_32LE); - } - return rb_ascii8bit_encoding(); - } - return rb_enc_from_index(encidx); -} - -static rb_encoding * get_encoding(VALUE str) { - return get_actual_encoding(ENCODING_GET(str), str); + return rb_enc_from_index(ENCODING_GET(str)); } static void mustnot_broken(VALUE str) { if (is_broken_string(str)) { - rb_raise(rb_eArgError, "invalid byte sequence in %s", rb_enc_name(STR_ENC_GET(str))); + rb_raise(rb_eArgError, "invalid byte sequence in %s", rb_enc_name(STR_ENC_GET(str))); } } @@ -326,7 +328,7 @@ mustnot_wchar(VALUE str) { rb_encoding *enc = STR_ENC_GET(str); if (rb_enc_mbminlen(enc) > 1) { - rb_raise(rb_eArgError, "wide char encoding: %s", rb_enc_name(enc)); + rb_raise(rb_eArgError, "wide char encoding: %s", rb_enc_name(enc)); } } @@ -354,51 +356,51 @@ fstr_update_callback(st_data_t *key, st_data_t *value, st_data_t data, int exist VALUE str = (VALUE)*key; if (existing) { - /* because of lazy sweep, str may be unmarked already and swept - * at next time */ + /* because of lazy sweep, str may be unmarked already and swept + * at next time */ - if (rb_objspace_garbage_object_p(str)) { - arg->fstr = Qundef; - return ST_DELETE; - } + if (rb_objspace_garbage_object_p(str)) { + arg->fstr = Qundef; + return ST_DELETE; + } - arg->fstr = str; - return ST_STOP; + arg->fstr = str; + return ST_STOP; } else { - if (FL_TEST_RAW(str, STR_FAKESTR)) { - if (arg->copy) { - VALUE new_str = str_new(rb_cString, RSTRING(str)->as.heap.ptr, RSTRING(str)->as.heap.len); - rb_enc_copy(new_str, str); - str = new_str; - } - else { - str = str_new_static(rb_cString, RSTRING(str)->as.heap.ptr, - RSTRING(str)->as.heap.len, - ENCODING_GET(str)); - } - OBJ_FREEZE_RAW(str); - } - else { - if (!OBJ_FROZEN(str)) - str = str_new_frozen(rb_cString, str); - if (STR_SHARED_P(str)) { /* str should not be shared */ - /* shared substring */ + if (FL_TEST_RAW(str, STR_FAKESTR)) { + if (arg->copy) { + VALUE new_str = str_new(rb_cString, RSTRING(str)->as.heap.ptr, RSTRING(str)->len); + rb_enc_copy(new_str, str); + str = new_str; + } + else { + str = str_new_static(rb_cString, RSTRING(str)->as.heap.ptr, + RSTRING(str)->len, + ENCODING_GET(str)); + } + OBJ_FREEZE(str); + } + else { + if (!OBJ_FROZEN(str) || CHILLED_STRING_P(str)) { + str = str_new_frozen(rb_cString, str); + } + if (STR_SHARED_P(str)) { /* str should not be shared */ + /* shared substring */ str_make_independent(str); - assert(OBJ_FROZEN(str)); - } - if (!BARE_STRING_P(str)) { - str = str_new_frozen(rb_cString, str); - } - } - RBASIC(str)->flags |= RSTRING_FSTR; + RUBY_ASSERT(OBJ_FROZEN(str)); + } + if (!BARE_STRING_P(str)) { + str = str_new_frozen(rb_cString, str); + } + } + RBASIC(str)->flags |= RSTRING_FSTR; - *key = *value = arg->fstr = str; - return ST_CONTINUE; + *key = *value = arg->fstr = str; + return ST_CONTINUE; } } -RUBY_FUNC_EXPORTED VALUE rb_fstring(VALUE str) { @@ -408,29 +410,30 @@ rb_fstring(VALUE str) Check_Type(str, T_STRING); if (FL_TEST(str, RSTRING_FSTR)) - return str; + return str; bare = BARE_STRING_P(str); if (!bare) { if (STR_EMBED_P(str)) { - OBJ_FREEZE_RAW(str); + OBJ_FREEZE(str); return str; } - if (FL_TEST_RAW(str, STR_NOEMBED|STR_SHARED_ROOT|STR_SHARED) == (STR_NOEMBED|STR_SHARED_ROOT)) { - assert(OBJ_FROZEN(str)); + + if (FL_TEST_RAW(str, STR_SHARED_ROOT | STR_SHARED) == STR_SHARED_ROOT) { + RUBY_ASSERT(OBJ_FROZEN(str)); return str; } } - if (!OBJ_FROZEN(str)) + if (!FL_TEST_RAW(str, FL_FREEZE | STR_NOFREE | STR_CHILLED)) rb_str_resize(str, RSTRING_LEN(str)); fstr = register_fstring(str, FALSE); if (!bare) { - str_replace_shared_without_enc(str, fstr); - OBJ_FREEZE_RAW(str); - return str; + str_replace_shared_without_enc(str, fstr); + OBJ_FREEZE(str); + return str; } return fstr; } @@ -447,14 +450,15 @@ register_fstring(VALUE str, bool copy) do { args.fstr = str; st_update(frozen_strings, (st_data_t)str, fstr_update_callback, (st_data_t)&args); - } while (args.fstr == Qundef); + } while (UNDEF_P(args.fstr)); } RB_VM_LOCK_LEAVE(); - assert(OBJ_FROZEN(args.fstr)); - assert(!FL_TEST_RAW(args.fstr, STR_FAKESTR)); - assert(!FL_TEST_RAW(args.fstr, FL_EXIVAR)); - assert(RBASIC_CLASS(args.fstr) == rb_cString); + RUBY_ASSERT(OBJ_FROZEN(args.fstr)); + RUBY_ASSERT(!FL_TEST_RAW(args.fstr, STR_FAKESTR)); + RUBY_ASSERT(!FL_TEST_RAW(args.fstr, FL_EXIVAR)); + RUBY_ASSERT(RBASIC_CLASS(args.fstr) == rb_cString); + return args.fstr; } @@ -465,14 +469,14 @@ setup_fake_str(struct RString *fake_str, const char *name, long len, int encidx) /* SHARED to be allocated by the callback */ if (!name) { - RUBY_ASSERT_ALWAYS(len == 0); - name = ""; + RUBY_ASSERT_ALWAYS(len == 0); + name = ""; } ENCODING_SET_INLINED((VALUE)fake_str, encidx); RBASIC_SET_CLASS_RAW((VALUE)fake_str, rb_cString); - fake_str->as.heap.len = len; + fake_str->len = len; fake_str->as.heap.ptr = (char *)name; fake_str->as.heap.aux.capa = len; return (VALUE)fake_str; @@ -492,7 +496,7 @@ rb_setup_fake_str(struct RString *fake_str, const char *name, long len, rb_encod * shared string which refers a static string literal. `ptr` must * point a constant string. */ -MJIT_FUNC_EXPORTED VALUE +VALUE rb_fstring_new(const char *ptr, long len) { struct RString fake_str; @@ -527,8 +531,8 @@ fstring_cmp(VALUE a, VALUE b) RSTRING_GETMEM(a, aptr, alen); RSTRING_GETMEM(b, bptr, blen); return (alen != blen || - ENCODING_GET(a) != ENCODING_GET(b) || - memcmp(aptr, bptr, alen) != 0); + ENCODING_GET(a) != ENCODING_GET(b) || + memcmp(aptr, bptr, alen) != 0); } static inline int @@ -576,23 +580,23 @@ search_nonascii(const char *p, const char *e) if (UNALIGNED_WORD_ACCESS || e - p >= SIZEOF_VOIDP) { #if !UNALIGNED_WORD_ACCESS - if ((uintptr_t)p % SIZEOF_VOIDP) { - int l = SIZEOF_VOIDP - (uintptr_t)p % SIZEOF_VOIDP; - p += l; - switch (l) { - default: UNREACHABLE; + if ((uintptr_t)p % SIZEOF_VOIDP) { + int l = SIZEOF_VOIDP - (uintptr_t)p % SIZEOF_VOIDP; + p += l; + switch (l) { + default: UNREACHABLE; #if SIZEOF_VOIDP > 4 - case 7: if (p[-7]&0x80) return p-7; - case 6: if (p[-6]&0x80) return p-6; - case 5: if (p[-5]&0x80) return p-5; - case 4: if (p[-4]&0x80) return p-4; + case 7: if (p[-7]&0x80) return p-7; + case 6: if (p[-6]&0x80) return p-6; + case 5: if (p[-5]&0x80) return p-5; + case 4: if (p[-4]&0x80) return p-4; #endif - case 3: if (p[-3]&0x80) return p-3; - case 2: if (p[-2]&0x80) return p-2; - case 1: if (p[-1]&0x80) return p-1; - case 0: break; - } - } + case 3: if (p[-3]&0x80) return p-3; + case 2: if (p[-2]&0x80) return p-2; + case 1: if (p[-1]&0x80) return p-1; + case 0: break; + } + } #endif #if defined(HAVE_BUILTIN___BUILTIN_ASSUME_ALIGNED) &&! UNALIGNED_WORD_ACCESS #define aligned_ptr(value) \ @@ -600,19 +604,19 @@ search_nonascii(const char *p, const char *e) #else #define aligned_ptr(value) (uintptr_t *)(value) #endif - s = aligned_ptr(p); - t = (uintptr_t *)(e - (SIZEOF_VOIDP-1)); + s = aligned_ptr(p); + t = (uintptr_t *)(e - (SIZEOF_VOIDP-1)); #undef aligned_ptr - for (;s < t; s++) { - if (*s & NONASCII_MASK) { + for (;s < t; s++) { + if (*s & NONASCII_MASK) { #ifdef WORDS_BIGENDIAN - return (const char *)s + (nlz_intptr(*s&NONASCII_MASK)>>3); + return (const char *)s + (nlz_intptr(*s&NONASCII_MASK)>>3); #else - return (const char *)s + (ntz_intptr(*s&NONASCII_MASK)>>3); + return (const char *)s + (ntz_intptr(*s&NONASCII_MASK)>>3); #endif - } - } - p = (const char *)s; + } + } + p = (const char *)s; } switch (e - p) { @@ -669,42 +673,42 @@ rb_str_coderange_scan_restartable(const char *s, const char *e, rb_encoding *enc const char *p = s; if (*cr == ENC_CODERANGE_BROKEN) - return e - s; + return e - s; if (rb_enc_to_index(enc) == rb_ascii8bit_encindex()) { - /* enc is ASCII-8BIT. ASCII-8BIT string never be broken. */ - if (*cr == ENC_CODERANGE_VALID) return e - s; - p = search_nonascii(p, e); + /* enc is ASCII-8BIT. ASCII-8BIT string never be broken. */ + if (*cr == ENC_CODERANGE_VALID) return e - s; + p = search_nonascii(p, e); *cr = p ? ENC_CODERANGE_VALID : ENC_CODERANGE_7BIT; - return e - s; + return e - s; } else if (rb_enc_asciicompat(enc)) { - p = search_nonascii(p, e); - if (!p) { - if (*cr != ENC_CODERANGE_VALID) *cr = ENC_CODERANGE_7BIT; - return e - s; - } - for (;;) { - int ret = rb_enc_precise_mbclen(p, e, enc); - if (!MBCLEN_CHARFOUND_P(ret)) { - *cr = MBCLEN_INVALID_P(ret) ? ENC_CODERANGE_BROKEN: ENC_CODERANGE_UNKNOWN; - return p - s; - } - p += MBCLEN_CHARFOUND_LEN(ret); - if (p == e) break; - p = search_nonascii(p, e); - if (!p) break; - } + p = search_nonascii(p, e); + if (!p) { + if (*cr != ENC_CODERANGE_VALID) *cr = ENC_CODERANGE_7BIT; + return e - s; + } + for (;;) { + int ret = rb_enc_precise_mbclen(p, e, enc); + if (!MBCLEN_CHARFOUND_P(ret)) { + *cr = MBCLEN_INVALID_P(ret) ? ENC_CODERANGE_BROKEN: ENC_CODERANGE_UNKNOWN; + return p - s; + } + p += MBCLEN_CHARFOUND_LEN(ret); + if (p == e) break; + p = search_nonascii(p, e); + if (!p) break; + } } else { - while (p < e) { - int ret = rb_enc_precise_mbclen(p, e, enc); - if (!MBCLEN_CHARFOUND_P(ret)) { - *cr = MBCLEN_INVALID_P(ret) ? ENC_CODERANGE_BROKEN: ENC_CODERANGE_UNKNOWN; - return p - s; - } - p += MBCLEN_CHARFOUND_LEN(ret); - } + while (p < e) { + int ret = rb_enc_precise_mbclen(p, e, enc); + if (!MBCLEN_CHARFOUND_P(ret)) { + *cr = MBCLEN_INVALID_P(ret) ? ENC_CODERANGE_BROKEN: ENC_CODERANGE_UNKNOWN; + return p - s; + } + p += MBCLEN_CHARFOUND_LEN(ret); + } } *cr = ENC_CODERANGE_VALID; return e - s; @@ -716,6 +720,20 @@ str_enc_copy(VALUE str1, VALUE str2) rb_enc_set_index(str1, ENCODING_GET(str2)); } +/* Like str_enc_copy, but does not check frozen status of str1. + * You should use this only if you're certain that str1 is not frozen. */ +static inline void +str_enc_copy_direct(VALUE str1, VALUE str2) +{ + int inlined_encoding = RB_ENCODING_GET_INLINED(str2); + if (inlined_encoding == ENCODING_INLINE_MAX) { + rb_enc_set_index(str1, rb_enc_get_index(str2)); + } + else { + ENCODING_SET_INLINED(str1, inlined_encoding); + } +} + static void rb_enc_cr_str_copy_for_substr(VALUE dest, VALUE src) { @@ -724,25 +742,25 @@ rb_enc_cr_str_copy_for_substr(VALUE dest, VALUE src) */ str_enc_copy(dest, src); if (RSTRING_LEN(dest) == 0) { - if (!rb_enc_asciicompat(STR_ENC_GET(src))) - ENC_CODERANGE_SET(dest, ENC_CODERANGE_VALID); - else - ENC_CODERANGE_SET(dest, ENC_CODERANGE_7BIT); - return; + if (!rb_enc_asciicompat(STR_ENC_GET(src))) + ENC_CODERANGE_SET(dest, ENC_CODERANGE_VALID); + else + ENC_CODERANGE_SET(dest, ENC_CODERANGE_7BIT); + return; } switch (ENC_CODERANGE(src)) { case ENC_CODERANGE_7BIT: - ENC_CODERANGE_SET(dest, ENC_CODERANGE_7BIT); - break; + ENC_CODERANGE_SET(dest, ENC_CODERANGE_7BIT); + break; case ENC_CODERANGE_VALID: - if (!rb_enc_asciicompat(STR_ENC_GET(src)) || - search_nonascii(RSTRING_PTR(dest), RSTRING_END(dest))) - ENC_CODERANGE_SET(dest, ENC_CODERANGE_VALID); - else - ENC_CODERANGE_SET(dest, ENC_CODERANGE_7BIT); - break; + if (!rb_enc_asciicompat(STR_ENC_GET(src)) || + search_nonascii(RSTRING_PTR(dest), RSTRING_END(dest))) + ENC_CODERANGE_SET(dest, ENC_CODERANGE_VALID); + else + ENC_CODERANGE_SET(dest, ENC_CODERANGE_7BIT); + break; default: - break; + break; } } @@ -754,21 +772,15 @@ rb_enc_cr_str_exact_copy(VALUE dest, VALUE src) } static int -enc_coderange_scan(VALUE str, rb_encoding *enc, int encidx) +enc_coderange_scan(VALUE str, rb_encoding *enc) { - if (rb_enc_mbminlen(enc) > 1 && rb_enc_dummy_p(enc) && - rb_enc_mbminlen(enc = get_actual_encoding(encidx, str)) == 1) { - return ENC_CODERANGE_BROKEN; - } - else { - return coderange_scan(RSTRING_PTR(str), RSTRING_LEN(str), enc); - } + return coderange_scan(RSTRING_PTR(str), RSTRING_LEN(str), enc); } int rb_enc_str_coderange_scan(VALUE str, rb_encoding *enc) { - return enc_coderange_scan(str, enc, rb_enc_to_index(enc)); + return enc_coderange_scan(str, enc); } int @@ -777,9 +789,7 @@ rb_enc_str_coderange(VALUE str) int cr = ENC_CODERANGE(str); if (cr == ENC_CODERANGE_UNKNOWN) { - int encidx = ENCODING_GET(str); - rb_encoding *enc = rb_enc_from_index(encidx); - cr = enc_coderange_scan(str, enc, encidx); + cr = enc_coderange_scan(str, get_encoding(str)); ENC_CODERANGE_SET(str, cr); } return cr; @@ -792,7 +802,7 @@ rb_enc_str_asciionly_p(VALUE str) if (!rb_enc_asciicompat(enc)) return FALSE; - else if (rb_enc_str_coderange(str) == ENC_CODERANGE_7BIT) + else if (is_ascii_string(str)) return TRUE; return FALSE; } @@ -801,7 +811,7 @@ static inline void str_mod_check(VALUE s, const char *p, long len) { if (RSTRING_PTR(s) != p || RSTRING_LEN(s) != len){ - rb_raise(rb_eRuntimeError, "string modified"); + rb_raise(rb_eRuntimeError, "string modified"); } } @@ -809,17 +819,13 @@ static size_t str_capacity(VALUE str, const int termlen) { if (STR_EMBED_P(str)) { -#if USE_RVARGC return str_embed_capa(str) - termlen; -#else - return (RSTRING_EMBED_LEN_MAX + 1 - termlen); -#endif } else if (FL_TEST(str, STR_SHARED|STR_NOFREE)) { - return RSTRING(str)->as.heap.len; + return RSTRING(str)->len; } else { - return RSTRING(str)->as.heap.aux.capa; + return RSTRING(str)->as.heap.aux.capa; } } @@ -833,34 +839,30 @@ static inline void must_not_null(const char *ptr) { if (!ptr) { - rb_raise(rb_eArgError, "NULL pointer given"); + rb_raise(rb_eArgError, "NULL pointer given"); } } static inline VALUE -str_alloc(VALUE klass, size_t size) -{ - assert(size > 0); - RVARGC_NEWOBJ_OF(str, struct RString, klass, - T_STRING | (RGENGC_WB_PROTECTED_STRING ? FL_WB_PROTECTED : 0), size); - return (VALUE)str; -} - -static inline VALUE str_alloc_embed(VALUE klass, size_t capa) { - size_t size = str_embed_size(capa); - assert(rb_gc_size_allocatable_p(size)); -#if !USE_RVARGC - assert(size <= sizeof(struct RString)); -#endif - return str_alloc(klass, size); + size_t size = rb_str_embed_size(capa); + RUBY_ASSERT(size > 0); + RUBY_ASSERT(rb_gc_size_allocatable_p(size)); + + NEWOBJ_OF(str, struct RString, klass, + T_STRING | (RGENGC_WB_PROTECTED_STRING ? FL_WB_PROTECTED : 0), size, 0); + + return (VALUE)str; } static inline VALUE str_alloc_heap(VALUE klass) { - return str_alloc(klass, sizeof(struct RString)); + NEWOBJ_OF(str, struct RString, klass, + T_STRING | STR_NOEMBED | (RGENGC_WB_PROTECTED_STRING ? FL_WB_PROTECTED : 0), sizeof(struct RString), 0); + + return (VALUE)str; } static inline VALUE @@ -878,7 +880,7 @@ str_new0(VALUE klass, const char *ptr, long len, int termlen) VALUE str; if (len < 0) { - rb_raise(rb_eArgError, "negative string size (or size too big)"); + rb_raise(rb_eArgError, "negative string size (or size too big)"); } RUBY_DTRACE_CREATE_HOOK(STRING, len); @@ -891,16 +893,15 @@ str_new0(VALUE klass, const char *ptr, long len, int termlen) } else { str = str_alloc_heap(klass); - RSTRING(str)->as.heap.aux.capa = len; + RSTRING(str)->as.heap.aux.capa = len; /* :FIXME: @shyouhei guesses `len + termlen` is guaranteed to never * integer overflow. If we can STATIC_ASSERT that, the following * mul_add_mul can be reverted to a simple ALLOC_N. */ RSTRING(str)->as.heap.ptr = rb_xmalloc_mul_add_mul(sizeof(char), len, sizeof(char), termlen); - STR_SET_NOEMBED(str); } if (ptr) { - memcpy(RSTRING_PTR(str), ptr, len); + memcpy(RSTRING_PTR(str), ptr, len); } STR_SET_LEN(str, len); TERM_FILL(RSTRING_PTR(str) + len, termlen); @@ -980,7 +981,7 @@ rb_enc_str_new_cstr(const char *ptr, rb_encoding *enc) { must_not_null(ptr); if (rb_enc_mbminlen(enc) != 1) { - rb_raise(rb_eArgError, "wchar encoding given"); + rb_raise(rb_eArgError, "wchar encoding given"); } return rb_enc_str_new(ptr, strlen(ptr), enc); } @@ -991,21 +992,20 @@ str_new_static(VALUE klass, const char *ptr, long len, int encindex) VALUE str; if (len < 0) { - rb_raise(rb_eArgError, "negative string size (or size too big)"); + rb_raise(rb_eArgError, "negative string size (or size too big)"); } if (!ptr) { - rb_encoding *enc = rb_enc_get_from_index(encindex); - str = str_new0(klass, ptr, len, rb_enc_mbminlen(enc)); + rb_encoding *enc = rb_enc_get_from_index(encindex); + str = str_new0(klass, ptr, len, rb_enc_mbminlen(enc)); } else { - RUBY_DTRACE_CREATE_HOOK(STRING, len); + RUBY_DTRACE_CREATE_HOOK(STRING, len); str = str_alloc_heap(klass); - RSTRING(str)->as.heap.len = len; - RSTRING(str)->as.heap.ptr = (char *)ptr; - RSTRING(str)->as.heap.aux.capa = len; - STR_SET_NOEMBED(str); - RBASIC(str)->flags |= STR_NOFREE; + RSTRING(str)->len = len; + RSTRING(str)->as.heap.ptr = (char *)ptr; + RSTRING(str)->as.heap.aux.capa = len; + RBASIC(str)->flags |= STR_NOFREE; } rb_enc_associate_index(str, encindex); return str; @@ -1036,16 +1036,16 @@ rb_enc_str_new_static(const char *ptr, long len, rb_encoding *enc) } static VALUE str_cat_conv_enc_opts(VALUE newstr, long ofs, const char *ptr, long len, - rb_encoding *from, rb_encoding *to, - int ecflags, VALUE ecopts); + rb_encoding *from, rb_encoding *to, + int ecflags, VALUE ecopts); static inline bool is_enc_ascii_string(VALUE str, rb_encoding *enc) { int encidx = rb_enc_to_index(enc); if (rb_enc_get_index(str) == encidx) - return is_ascii_string(str); - return enc_coderange_scan(str, enc, encidx) == ENC_CODERANGE_7BIT; + return is_ascii_string(str); + return enc_coderange_scan(str, enc) == ENC_CODERANGE_7BIT; } VALUE @@ -1059,27 +1059,27 @@ rb_str_conv_enc_opts(VALUE str, rb_encoding *from, rb_encoding *to, int ecflags, if (!from) from = rb_enc_get(str); if (from == to) return str; if ((rb_enc_asciicompat(to) && is_enc_ascii_string(str, from)) || - to == rb_ascii8bit_encoding()) { - if (STR_ENC_GET(str) != to) { - str = rb_str_dup(str); - rb_enc_associate(str, to); - } - return str; + rb_is_ascii8bit_enc(to)) { + if (STR_ENC_GET(str) != to) { + str = rb_str_dup(str); + rb_enc_associate(str, to); + } + return str; } RSTRING_GETMEM(str, ptr, len); newstr = str_cat_conv_enc_opts(rb_str_buf_new(len), 0, ptr, len, - from, to, ecflags, ecopts); + from, to, ecflags, ecopts); if (NIL_P(newstr)) { - /* some error, return original */ - return str; + /* some error, return original */ + return str; } return newstr; } VALUE rb_str_cat_conv_enc_opts(VALUE newstr, long ofs, const char *ptr, long len, - rb_encoding *from, int ecflags, VALUE ecopts) + rb_encoding *from, int ecflags, VALUE ecopts) { long olen; @@ -1088,14 +1088,14 @@ rb_str_cat_conv_enc_opts(VALUE newstr, long ofs, const char *ptr, long len, rb_raise(rb_eIndexError, "index %ld out of string", ofs); if (ofs < 0) ofs += olen; if (!from) { - STR_SET_LEN(newstr, ofs); - return rb_str_cat(newstr, ptr, len); + STR_SET_LEN(newstr, ofs); + return rb_str_cat(newstr, ptr, len); } rb_str_modify(newstr); return str_cat_conv_enc_opts(newstr, ofs, ptr, len, from, - rb_enc_get(newstr), - ecflags, ecopts); + rb_enc_get(newstr), + ecflags, ecopts); } VALUE @@ -1109,8 +1109,8 @@ rb_str_initialize(VALUE str, const char *ptr, long len, rb_encoding *enc) static VALUE str_cat_conv_enc_opts(VALUE newstr, long ofs, const char *ptr, long len, - rb_encoding *from, rb_encoding *to, - int ecflags, VALUE ecopts) + rb_encoding *from, rb_encoding *to, + int ecflags, VALUE ecopts) { rb_econv_t *ec; rb_econv_result_t ret; @@ -1131,35 +1131,36 @@ str_cat_conv_enc_opts(VALUE newstr, long ofs, const char *ptr, long len, sp = (unsigned char*)ptr; start = sp; while ((dest = (unsigned char*)RSTRING_PTR(newstr)), - (dp = dest + converted_output), - (ret = rb_econv_convert(ec, &sp, start + len, &dp, dest + olen, 0)), - ret == econv_destination_buffer_full) { - /* destination buffer short */ - size_t converted_input = sp - start; - size_t rest = len - converted_input; - converted_output = dp - dest; - rb_str_set_len(newstr, converted_output); - if (converted_input && converted_output && - rest < (LONG_MAX / converted_output)) { - rest = (rest * converted_output) / converted_input; - } - else { - rest = olen; - } - olen += rest < 2 ? 2 : rest; - rb_str_resize(newstr, olen); + (dp = dest + converted_output), + (ret = rb_econv_convert(ec, &sp, start + len, &dp, dest + olen, 0)), + ret == econv_destination_buffer_full) { + /* destination buffer short */ + size_t converted_input = sp - start; + size_t rest = len - converted_input; + converted_output = dp - dest; + rb_str_set_len(newstr, converted_output); + if (converted_input && converted_output && + rest < (LONG_MAX / converted_output)) { + rest = (rest * converted_output) / converted_input; + } + else { + rest = olen; + } + olen += rest < 2 ? 2 : rest; + rb_str_resize(newstr, olen); } DATA_PTR(econv_wrapper) = 0; + RB_GC_GUARD(econv_wrapper); rb_econv_close(ec); switch (ret) { case econv_finished: - len = dp - (unsigned char*)RSTRING_PTR(newstr); - rb_str_set_len(newstr, len); - rb_enc_associate(newstr, to); - return newstr; + len = dp - (unsigned char*)RSTRING_PTR(newstr); + rb_str_set_len(newstr, len); + rb_enc_associate(newstr, to); + return newstr; default: - return Qnil; + return Qnil; } } @@ -1182,7 +1183,7 @@ rb_external_str_new_with_enc(const char *ptr, long len, rb_encoding *eenc) /* ASCII-8BIT case, no conversion */ if ((eidx == rb_ascii8bit_encindex()) || - (eidx == rb_usascii_encindex() && search_nonascii(ptr, ptr + len))) { + (eidx == rb_usascii_encindex() && search_nonascii(ptr, ptr + len))) { return rb_str_new(ptr, len); } /* no default_internal or same encoding, no conversion */ @@ -1193,8 +1194,8 @@ rb_external_str_new_with_enc(const char *ptr, long len, rb_encoding *eenc) /* ASCII compatible, and ASCII only string, no conversion in * default_internal */ if ((eidx == rb_ascii8bit_encindex()) || - (eidx == rb_usascii_encindex()) || - (rb_enc_asciicompat(eenc) && !search_nonascii(ptr, ptr + len))) { + (eidx == rb_usascii_encindex()) || + (rb_enc_asciicompat(eenc) && !search_nonascii(ptr, ptr + len))) { return rb_enc_str_new(ptr, len, ienc); } /* convert from the given encoding to default_internal */ @@ -1202,7 +1203,7 @@ rb_external_str_new_with_enc(const char *ptr, long len, rb_encoding *eenc) /* when the conversion failed for some reason, just ignore the * default_internal and result in the given encoding as-is. */ if (NIL_P(rb_str_cat_conv_enc_opts(str, 0, ptr, len, eenc, 0, Qnil))) { - rb_str_initialize(str, ptr, len, eenc); + rb_str_initialize(str, ptr, len, eenc); } return str; } @@ -1212,9 +1213,9 @@ rb_external_str_with_enc(VALUE str, rb_encoding *eenc) { int eidx = rb_enc_to_index(eenc); if (eidx == rb_usascii_encindex() && - rb_enc_str_coderange(str) != ENC_CODERANGE_7BIT) { - rb_enc_associate_index(str, rb_ascii8bit_encindex()); - return str; + !is_ascii_string(str)) { + rb_enc_associate_index(str, rb_ascii8bit_encindex()); + return str; } rb_enc_associate_index(str, eidx); return rb_str_conv_enc(str, eenc, rb_default_internal_encoding()); @@ -1284,10 +1285,9 @@ str_replace_shared_without_enc(VALUE str2, VALUE str) RSTRING_GETMEM(str, ptr, len); if (str_embed_capa(str2) >= len + termlen) { char *ptr2 = RSTRING(str2)->as.embed.ary; - STR_SET_EMBED(str2); - memcpy(ptr2, RSTRING_PTR(str), len); - STR_SET_EMBED_LEN(str2, len); - TERM_FILL(ptr2+len, termlen); + STR_SET_EMBED(str2); + memcpy(ptr2, RSTRING_PTR(str), len); + TERM_FILL(ptr2+len, termlen); } else { VALUE root; @@ -1299,7 +1299,8 @@ str_replace_shared_without_enc(VALUE str2, VALUE str) root = rb_str_new_frozen(str); RSTRING_GETMEM(root, ptr, len); } - assert(OBJ_FROZEN(root)); + RUBY_ASSERT(OBJ_FROZEN(root)); + if (!STR_EMBED_P(str2) && !FL_TEST_RAW(str2, STR_SHARED|STR_NOFREE)) { if (FL_TEST_RAW(str2, STR_SHARED_ROOT)) { rb_fatal("about to free a possible shared root"); @@ -1309,11 +1310,13 @@ str_replace_shared_without_enc(VALUE str2, VALUE str) ruby_sized_xfree(ptr2, STR_HEAP_SIZE(str2)); } } - FL_SET(str2, STR_NOEMBED); - RSTRING(str2)->as.heap.len = len; - RSTRING(str2)->as.heap.ptr = ptr; + FL_SET(str2, STR_NOEMBED); + RSTRING(str2)->as.heap.ptr = ptr; STR_SET_SHARED(str2, root); } + + STR_SET_LEN(str2, len); + return str2; } @@ -1340,7 +1343,7 @@ rb_str_new_shared(VALUE str) VALUE rb_str_new_frozen(VALUE orig) { - if (OBJ_FROZEN(orig)) return orig; + if (RB_FL_TEST_RAW(orig, FL_FREEZE | STR_CHILLED) == FL_FREEZE) return orig; return str_new_frozen(rb_obj_class(orig), orig); } @@ -1358,33 +1361,69 @@ rb_str_tmp_frozen_acquire(VALUE orig) return str_new_frozen_buffer(0, orig, FALSE); } +VALUE +rb_str_tmp_frozen_no_embed_acquire(VALUE orig) +{ + if (OBJ_FROZEN_RAW(orig) && !STR_EMBED_P(orig) && !rb_str_reembeddable_p(orig)) return orig; + if (STR_SHARED_P(orig) && !STR_EMBED_P(RSTRING(orig)->as.heap.aux.shared)) return rb_str_tmp_frozen_acquire(orig); + + VALUE str = str_alloc_heap(0); + OBJ_FREEZE(str); + /* Always set the STR_SHARED_ROOT to ensure it does not get re-embedded. */ + FL_SET(str, STR_SHARED_ROOT); + + size_t capa = str_capacity(orig, TERM_LEN(orig)); + + /* If the string is embedded then we want to create a copy that is heap + * allocated. If the string is shared then the shared root must be + * embedded, so we want to create a copy. If the string is a shared root + * then it must be embedded, so we want to create a copy. */ + if (STR_EMBED_P(orig) || FL_TEST_RAW(orig, STR_SHARED | STR_SHARED_ROOT)) { + RSTRING(str)->as.heap.ptr = rb_xmalloc_mul_add_mul(sizeof(char), capa, sizeof(char), TERM_LEN(orig)); + memcpy(RSTRING(str)->as.heap.ptr, RSTRING_PTR(orig), capa); + } + else { + /* orig must be heap allocated and not shared, so we can safely transfer + * the pointer to str. */ + RSTRING(str)->as.heap.ptr = RSTRING(orig)->as.heap.ptr; + RBASIC(str)->flags |= RBASIC(orig)->flags & STR_NOFREE; + RBASIC(orig)->flags &= ~STR_NOFREE; + STR_SET_SHARED(orig, str); + } + + RSTRING(str)->len = RSTRING(orig)->len; + RSTRING(str)->as.heap.aux.capa = capa; + + return str; +} + void rb_str_tmp_frozen_release(VALUE orig, VALUE tmp) { if (RBASIC_CLASS(tmp) != 0) - return; + return; if (STR_EMBED_P(tmp)) { - assert(OBJ_FROZEN_RAW(tmp)); + RUBY_ASSERT(OBJ_FROZEN_RAW(tmp)); } else if (FL_TEST_RAW(orig, STR_SHARED) && - !FL_TEST_RAW(orig, STR_TMPLOCK|RUBY_FL_FREEZE)) { - VALUE shared = RSTRING(orig)->as.heap.aux.shared; + !FL_TEST_RAW(orig, STR_TMPLOCK|RUBY_FL_FREEZE)) { + VALUE shared = RSTRING(orig)->as.heap.aux.shared; - if (shared == tmp && !FL_TEST_RAW(tmp, STR_BORROWED)) { - assert(RSTRING(orig)->as.heap.ptr == RSTRING(tmp)->as.heap.ptr); - assert(RSTRING(orig)->as.heap.len == RSTRING(tmp)->as.heap.len); + if (shared == tmp && !FL_TEST_RAW(tmp, STR_BORROWED)) { + RUBY_ASSERT(RSTRING(orig)->as.heap.ptr == RSTRING(tmp)->as.heap.ptr); + RUBY_ASSERT(RSTRING_LEN(orig) == RSTRING_LEN(tmp)); /* Unshare orig since the root (tmp) only has this one child. */ - FL_UNSET_RAW(orig, STR_SHARED); - RSTRING(orig)->as.heap.aux.capa = RSTRING(tmp)->as.heap.aux.capa; - RBASIC(orig)->flags |= RBASIC(tmp)->flags & STR_NOFREE; - assert(OBJ_FROZEN_RAW(tmp)); + FL_UNSET_RAW(orig, STR_SHARED); + RSTRING(orig)->as.heap.aux.capa = RSTRING(tmp)->as.heap.aux.capa; + RBASIC(orig)->flags |= RBASIC(tmp)->flags & STR_NOFREE; + RUBY_ASSERT(OBJ_FROZEN_RAW(tmp)); /* Make tmp embedded and empty so it is safe for sweeping. */ STR_SET_EMBED(tmp); - STR_SET_EMBED_LEN(tmp, 0); - } + STR_SET_LEN(tmp, 0); + } } } @@ -1397,12 +1436,11 @@ str_new_frozen(VALUE klass, VALUE orig) static VALUE heap_str_make_shared(VALUE klass, VALUE orig) { - assert(!STR_EMBED_P(orig)); - assert(!STR_SHARED_P(orig)); + RUBY_ASSERT(!STR_EMBED_P(orig)); + RUBY_ASSERT(!STR_SHARED_P(orig)); VALUE str = str_alloc_heap(klass); - STR_SET_NOEMBED(str); - RSTRING(str)->as.heap.len = RSTRING_LEN(orig); + STR_SET_LEN(str, RSTRING_LEN(orig)); RSTRING(str)->as.heap.ptr = RSTRING_PTR(orig); RSTRING(str)->as.heap.aux.capa = RSTRING(orig)->as.heap.aux.capa; RBASIC(str)->flags |= RBASIC(orig)->flags & STR_NOFREE; @@ -1423,45 +1461,42 @@ str_new_frozen_buffer(VALUE klass, VALUE orig, int copy_encoding) if (STR_EMBED_P(orig) || STR_EMBEDDABLE_P(len, termlen)) { str = str_new0(klass, RSTRING_PTR(orig), len, termlen); - assert(STR_EMBED_P(str)); + RUBY_ASSERT(STR_EMBED_P(str)); } else { - if (FL_TEST_RAW(orig, STR_SHARED)) { - VALUE shared = RSTRING(orig)->as.heap.aux.shared; + if (FL_TEST_RAW(orig, STR_SHARED)) { + VALUE shared = RSTRING(orig)->as.heap.aux.shared; long ofs = RSTRING(orig)->as.heap.ptr - RSTRING_PTR(shared); - long rest = RSTRING_LEN(shared) - ofs - RSTRING(orig)->as.heap.len; - assert(ofs >= 0); - assert(rest >= 0); - assert(ofs + rest <= RSTRING_LEN(shared)); -#if !USE_RVARGC - assert(!STR_EMBED_P(shared)); -#endif - assert(OBJ_FROZEN(shared)); - - if ((ofs > 0) || (rest > 0) || - (klass != RBASIC(shared)->klass) || - ENCODING_GET(shared) != ENCODING_GET(orig)) { - str = str_new_shared(klass, shared); - assert(!STR_EMBED_P(str)); - RSTRING(str)->as.heap.ptr += ofs; - RSTRING(str)->as.heap.len -= ofs + rest; - } - else { - if (RBASIC_CLASS(shared) == 0) - FL_SET_RAW(shared, STR_BORROWED); - return shared; - } - } + long rest = RSTRING_LEN(shared) - ofs - RSTRING_LEN(orig); + RUBY_ASSERT(ofs >= 0); + RUBY_ASSERT(rest >= 0); + RUBY_ASSERT(ofs + rest <= RSTRING_LEN(shared)); + RUBY_ASSERT(OBJ_FROZEN(shared)); + + if ((ofs > 0) || (rest > 0) || + (klass != RBASIC(shared)->klass) || + ENCODING_GET(shared) != ENCODING_GET(orig)) { + str = str_new_shared(klass, shared); + RUBY_ASSERT(!STR_EMBED_P(str)); + RSTRING(str)->as.heap.ptr += ofs; + STR_SET_LEN(str, RSTRING_LEN(str) - (ofs + rest)); + } + else { + if (RBASIC_CLASS(shared) == 0) + FL_SET_RAW(shared, STR_BORROWED); + return shared; + } + } else if (STR_EMBEDDABLE_P(RSTRING_LEN(orig), TERM_LEN(orig))) { str = str_alloc_embed(klass, RSTRING_LEN(orig) + TERM_LEN(orig)); - STR_SET_EMBED(str); - memcpy(RSTRING_PTR(str), RSTRING_PTR(orig), RSTRING_LEN(orig)); - STR_SET_EMBED_LEN(str, RSTRING_LEN(orig)); - TERM_FILL(RSTRING_END(str), TERM_LEN(orig)); - } - else { + STR_SET_EMBED(str); + memcpy(RSTRING_PTR(str), RSTRING_PTR(orig), RSTRING_LEN(orig)); + STR_SET_LEN(str, RSTRING_LEN(orig)); + TERM_FILL(RSTRING_END(str), TERM_LEN(orig)); + } + else { str = heap_str_make_shared(klass, orig); - } + } } if (copy_encoding) rb_enc_cr_str_exact_copy(str, orig); @@ -1484,9 +1519,6 @@ str_new_empty_String(VALUE str) } #define STR_BUF_MIN_SIZE 63 -#if !USE_RVARGC -STATIC_ASSERT(STR_BUF_MIN_SIZE, STR_BUF_MIN_SIZE > RSTRING_EMBED_LEN_MAX); -#endif VALUE rb_str_buf_new(long capa) @@ -1497,12 +1529,6 @@ rb_str_buf_new(long capa) VALUE str = str_alloc_heap(rb_cString); -#if !USE_RVARGC - if (capa < STR_BUF_MIN_SIZE) { - capa = STR_BUF_MIN_SIZE; - } -#endif - FL_SET(str, STR_NOEMBED); RSTRING(str)->as.heap.aux.capa = capa; RSTRING(str)->as.heap.ptr = ALLOC_N(char, (size_t)capa + 1); RSTRING(str)->as.heap.ptr[0] = '\0'; @@ -1532,7 +1558,7 @@ void rb_str_free(VALUE str) { if (FL_TEST(str, RSTRING_FSTR)) { - st_data_t fstr = (st_data_t)str; + st_data_t fstr = (st_data_t)str; RB_VM_LOCK_ENTER(); { @@ -1543,26 +1569,26 @@ rb_str_free(VALUE str) } if (STR_EMBED_P(str)) { - RB_DEBUG_COUNTER_INC(obj_str_embed); + RB_DEBUG_COUNTER_INC(obj_str_embed); } else if (FL_TEST(str, STR_SHARED | STR_NOFREE)) { - (void)RB_DEBUG_COUNTER_INC_IF(obj_str_shared, FL_TEST(str, STR_SHARED)); - (void)RB_DEBUG_COUNTER_INC_IF(obj_str_shared, FL_TEST(str, STR_NOFREE)); + (void)RB_DEBUG_COUNTER_INC_IF(obj_str_shared, FL_TEST(str, STR_SHARED)); + (void)RB_DEBUG_COUNTER_INC_IF(obj_str_shared, FL_TEST(str, STR_NOFREE)); } else { - RB_DEBUG_COUNTER_INC(obj_str_ptr); - ruby_sized_xfree(STR_HEAP_PTR(str), STR_HEAP_SIZE(str)); + RB_DEBUG_COUNTER_INC(obj_str_ptr); + ruby_sized_xfree(STR_HEAP_PTR(str), STR_HEAP_SIZE(str)); } } -RUBY_FUNC_EXPORTED size_t +size_t rb_str_memsize(VALUE str) { if (FL_TEST(str, STR_NOEMBED|STR_SHARED|STR_NOFREE) == STR_NOEMBED) { - return STR_HEAP_SIZE(str); + return STR_HEAP_SIZE(str); } else { - return 0; + return 0; } } @@ -1594,48 +1620,46 @@ str_shared_replace(VALUE str, VALUE str2) str_discard(str); termlen = rb_enc_mbminlen(enc); + STR_SET_LEN(str, RSTRING_LEN(str2)); + if (str_embed_capa(str) >= RSTRING_LEN(str2) + termlen) { - STR_SET_EMBED(str); - memcpy(RSTRING_PTR(str), RSTRING_PTR(str2), (size_t)RSTRING_LEN(str2) + termlen); - STR_SET_EMBED_LEN(str, RSTRING_LEN(str2)); + STR_SET_EMBED(str); + memcpy(RSTRING_PTR(str), RSTRING_PTR(str2), (size_t)RSTRING_LEN(str2) + termlen); rb_enc_associate(str, enc); ENC_CODERANGE_SET(str, cr); } else { -#if USE_RVARGC if (STR_EMBED_P(str2)) { - assert(!FL_TEST(str2, STR_SHARED)); - long len = RSTRING(str2)->as.embed.len; - assert(len + termlen <= str_embed_capa(str2)); + RUBY_ASSERT(!FL_TEST(str2, STR_SHARED)); + long len = RSTRING_LEN(str2); + RUBY_ASSERT(len + termlen <= str_embed_capa(str2)); char *new_ptr = ALLOC_N(char, len + termlen); memcpy(new_ptr, RSTRING(str2)->as.embed.ary, len + termlen); RSTRING(str2)->as.heap.ptr = new_ptr; - RSTRING(str2)->as.heap.len = len; + STR_SET_LEN(str2, len); RSTRING(str2)->as.heap.aux.capa = len; STR_SET_NOEMBED(str2); } -#endif - STR_SET_NOEMBED(str); - FL_UNSET(str, STR_SHARED); - RSTRING(str)->as.heap.ptr = RSTRING_PTR(str2); - RSTRING(str)->as.heap.len = RSTRING_LEN(str2); + STR_SET_NOEMBED(str); + FL_UNSET(str, STR_SHARED); + RSTRING(str)->as.heap.ptr = RSTRING_PTR(str2); - if (FL_TEST(str2, STR_SHARED)) { - VALUE shared = RSTRING(str2)->as.heap.aux.shared; - STR_SET_SHARED(str, shared); - } - else { - RSTRING(str)->as.heap.aux.capa = RSTRING(str2)->as.heap.aux.capa; - } + if (FL_TEST(str2, STR_SHARED)) { + VALUE shared = RSTRING(str2)->as.heap.aux.shared; + STR_SET_SHARED(str, shared); + } + else { + RSTRING(str)->as.heap.aux.capa = RSTRING(str2)->as.heap.aux.capa; + } - /* abandon str2 */ - STR_SET_EMBED(str2); - RSTRING_PTR(str2)[0] = 0; - STR_SET_EMBED_LEN(str2, 0); - rb_enc_associate(str, enc); - ENC_CODERANGE_SET(str, cr); + /* abandon str2 */ + STR_SET_EMBED(str2); + RSTRING_PTR(str2)[0] = 0; + STR_SET_LEN(str2, 0); + rb_enc_associate(str, enc); + ENC_CODERANGE_SET(str, cr); } } @@ -1645,17 +1669,17 @@ rb_obj_as_string(VALUE obj) VALUE str; if (RB_TYPE_P(obj, T_STRING)) { - return obj; + return obj; } str = rb_funcall(obj, idTo_s, 0); return rb_obj_as_string_result(str, obj); } -MJIT_FUNC_EXPORTED VALUE +VALUE rb_obj_as_string_result(VALUE str, VALUE obj) { if (!RB_TYPE_P(str, T_STRING)) - return rb_any_to_s(obj); + return rb_any_to_s(obj); return str; } @@ -1666,64 +1690,57 @@ str_replace(VALUE str, VALUE str2) len = RSTRING_LEN(str2); if (STR_SHARED_P(str2)) { - VALUE shared = RSTRING(str2)->as.heap.aux.shared; - assert(OBJ_FROZEN(shared)); - STR_SET_NOEMBED(str); - RSTRING(str)->as.heap.len = len; - RSTRING(str)->as.heap.ptr = RSTRING_PTR(str2); - STR_SET_SHARED(str, shared); - rb_enc_cr_str_exact_copy(str, str2); + VALUE shared = RSTRING(str2)->as.heap.aux.shared; + RUBY_ASSERT(OBJ_FROZEN(shared)); + STR_SET_NOEMBED(str); + STR_SET_LEN(str, len); + RSTRING(str)->as.heap.ptr = RSTRING_PTR(str2); + STR_SET_SHARED(str, shared); + rb_enc_cr_str_exact_copy(str, str2); } else { - str_replace_shared(str, str2); + str_replace_shared(str, str2); } return str; } static inline VALUE -ec_str_alloc(struct rb_execution_context_struct *ec, VALUE klass, size_t size) -{ - assert(size > 0); - RB_RVARGC_EC_NEWOBJ_OF(ec, str, struct RString, klass, - T_STRING | (RGENGC_WB_PROTECTED_STRING ? FL_WB_PROTECTED : 0), size); - return (VALUE)str; -} - -static inline VALUE ec_str_alloc_embed(struct rb_execution_context_struct *ec, VALUE klass, size_t capa) { - size_t size = str_embed_size(capa); - assert(rb_gc_size_allocatable_p(size)); -#if !USE_RVARGC - assert(size <= sizeof(struct RString)); -#endif - return ec_str_alloc(ec, klass, size); + size_t size = rb_str_embed_size(capa); + RUBY_ASSERT(size > 0); + RUBY_ASSERT(rb_gc_size_allocatable_p(size)); + + NEWOBJ_OF(str, struct RString, klass, + T_STRING | (RGENGC_WB_PROTECTED_STRING ? FL_WB_PROTECTED : 0), size, ec); + + return (VALUE)str; } static inline VALUE ec_str_alloc_heap(struct rb_execution_context_struct *ec, VALUE klass) { - return ec_str_alloc(ec, klass, sizeof(struct RString)); + NEWOBJ_OF(str, struct RString, klass, + T_STRING | STR_NOEMBED | (RGENGC_WB_PROTECTED_STRING ? FL_WB_PROTECTED : 0), sizeof(struct RString), ec); + + return (VALUE)str; } static inline VALUE str_duplicate_setup(VALUE klass, VALUE str, VALUE dup) { const VALUE flag_mask = -#if !USE_RVARGC - RSTRING_NOEMBED | RSTRING_EMBED_LEN_MASK | -#endif ENC_CODERANGE_MASK | ENCODING_MASK | FL_FREEZE - ; + ; VALUE flags = FL_TEST_RAW(str, flag_mask); int encidx = 0; if (STR_EMBED_P(str)) { - long len = RSTRING_EMBED_LEN(str); + long len = RSTRING_LEN(str); - assert(str_embed_capa(dup) >= len + 1); - STR_SET_EMBED_LEN(dup, len); + RUBY_ASSERT(STR_EMBED_P(dup)); + RUBY_ASSERT(str_embed_capa(dup) >= len + 1); MEMCPY(RSTRING(dup)->as.embed.ary, RSTRING(str)->as.embed.ary, char, len + 1); } else { @@ -1735,27 +1752,20 @@ str_duplicate_setup(VALUE klass, VALUE str, VALUE dup) root = str = str_new_frozen(klass, str); flags = FL_TEST_RAW(str, flag_mask); } - assert(!STR_SHARED_P(root)); - assert(RB_OBJ_FROZEN_RAW(root)); -#if USE_RVARGC - if (1) { -#else - if (STR_EMBED_P(root)) { - MEMCPY(RSTRING(dup)->as.embed.ary, RSTRING(root)->as.embed.ary, - char, RSTRING_EMBED_LEN_MAX + 1); - } - else { -#endif - RSTRING(dup)->as.heap.len = RSTRING_LEN(str); - RSTRING(dup)->as.heap.ptr = RSTRING_PTR(str); - RB_OBJ_WRITE(dup, &RSTRING(dup)->as.heap.aux.shared, root); - flags |= RSTRING_NOEMBED | STR_SHARED; - } + RUBY_ASSERT(!STR_SHARED_P(root)); + RUBY_ASSERT(RB_OBJ_FROZEN_RAW(root)); + + RSTRING(dup)->as.heap.ptr = RSTRING_PTR(str); + FL_SET(root, STR_SHARED_ROOT); + RB_OBJ_WRITE(dup, &RSTRING(dup)->as.heap.aux.shared, root); + flags |= RSTRING_NOEMBED | STR_SHARED; } + STR_SET_LEN(dup, RSTRING_LEN(str)); + if ((flags & ENCODING_MASK) == (ENCODING_INLINE_MAX<<ENCODING_SHIFT)) { - encidx = rb_enc_get_index(str); - flags &= ~ENCODING_MASK; + encidx = rb_enc_get_index(str); + flags &= ~ENCODING_MASK; } FL_SET_RAW(dup, flags & ~FL_FREEZE); if (encidx) rb_enc_associate_index(dup, encidx); @@ -1766,11 +1776,11 @@ static inline VALUE ec_str_duplicate(struct rb_execution_context_struct *ec, VALUE klass, VALUE str) { VALUE dup; - if (!USE_RVARGC || FL_TEST(str, STR_NOEMBED)) { - dup = ec_str_alloc_heap(ec, klass); + if (STR_EMBED_P(str)) { + dup = ec_str_alloc_embed(ec, klass, RSTRING_LEN(str) + TERM_LEN(str)); } else { - dup = ec_str_alloc_embed(ec, klass, RSTRING_EMBED_LEN(str) + TERM_LEN(str)); + dup = ec_str_alloc_heap(ec, klass); } return str_duplicate_setup(klass, str, dup); @@ -1780,11 +1790,11 @@ static inline VALUE str_duplicate(VALUE klass, VALUE str) { VALUE dup; - if (!USE_RVARGC || FL_TEST(str, STR_NOEMBED)) { - dup = str_alloc_heap(klass); + if (STR_EMBED_P(str)) { + dup = str_alloc_embed(klass, RSTRING_LEN(str) + TERM_LEN(str)); } else { - dup = str_alloc_embed(klass, RSTRING_EMBED_LEN(str) + TERM_LEN(str)); + dup = str_alloc_heap(klass); } return str_duplicate_setup(klass, str, dup); @@ -1796,6 +1806,18 @@ rb_str_dup(VALUE str) return str_duplicate(rb_obj_class(str), str); } +/* :nodoc: */ +VALUE +rb_str_dup_m(VALUE str) +{ + if (LIKELY(BARE_STRING_P(str))) { + return str_duplicate(rb_obj_class(str), str); + } + else { + return rb_obj_dup(str); + } +} + VALUE rb_str_resurrect(VALUE str) { @@ -1804,10 +1826,20 @@ rb_str_resurrect(VALUE str) } VALUE -rb_ec_str_resurrect(struct rb_execution_context_struct *ec, VALUE str) +rb_ec_str_resurrect(struct rb_execution_context_struct *ec, VALUE str, bool chilled) { RUBY_DTRACE_CREATE_HOOK(STRING, RSTRING_LEN(str)); - return ec_str_duplicate(ec, rb_cString, str); + VALUE new_str = ec_str_duplicate(ec, rb_cString, str); + if (chilled) { + STR_CHILL_RAW(new_str); + } + return new_str; +} + +bool +rb_str_chilled_p(VALUE str) +{ + return CHILLED_STRING_P(str); } /* @@ -1829,78 +1861,162 @@ rb_str_init(int argc, VALUE *argv, VALUE str) int n; if (!keyword_ids[0]) { - keyword_ids[0] = rb_id_encoding(); - CONST_ID(keyword_ids[1], "capacity"); + keyword_ids[0] = rb_id_encoding(); + CONST_ID(keyword_ids[1], "capacity"); } n = rb_scan_args(argc, argv, "01:", &orig, &opt); if (!NIL_P(opt)) { - rb_get_kwargs(opt, keyword_ids, 0, 2, kwargs); - venc = kwargs[0]; - vcapa = kwargs[1]; - if (venc != Qundef && !NIL_P(venc)) { - enc = rb_to_encoding(venc); - } - if (vcapa != Qundef && !NIL_P(vcapa)) { - long capa = NUM2LONG(vcapa); - long len = 0; - int termlen = enc ? rb_enc_mbminlen(enc) : 1; - - if (capa < STR_BUF_MIN_SIZE) { - capa = STR_BUF_MIN_SIZE; - } - if (n == 1) { - StringValue(orig); - len = RSTRING_LEN(orig); - if (capa < len) { - capa = len; - } - if (orig == str) n = 0; - } - str_modifiable(str); - if (STR_EMBED_P(str)) { /* make noembed always */ - char *new_ptr = ALLOC_N(char, (size_t)capa + termlen); -#if USE_RVARGC - assert(RSTRING(str)->as.embed.len + 1 <= str_embed_capa(str)); - memcpy(new_ptr, RSTRING(str)->as.embed.ary, RSTRING(str)->as.embed.len + 1); -#else - memcpy(new_ptr, RSTRING(str)->as.embed.ary, RSTRING_EMBED_LEN_MAX + 1); -#endif - RSTRING(str)->as.heap.ptr = new_ptr; + rb_get_kwargs(opt, keyword_ids, 0, 2, kwargs); + venc = kwargs[0]; + vcapa = kwargs[1]; + if (!UNDEF_P(venc) && !NIL_P(venc)) { + enc = rb_to_encoding(venc); + } + if (!UNDEF_P(vcapa) && !NIL_P(vcapa)) { + long capa = NUM2LONG(vcapa); + long len = 0; + int termlen = enc ? rb_enc_mbminlen(enc) : 1; + + if (capa < STR_BUF_MIN_SIZE) { + capa = STR_BUF_MIN_SIZE; } - else if (FL_TEST(str, STR_SHARED|STR_NOFREE)) { + if (n == 1) { + StringValue(orig); + len = RSTRING_LEN(orig); + if (capa < len) { + capa = len; + } + if (orig == str) n = 0; + } + str_modifiable(str); + if (STR_EMBED_P(str) || FL_TEST(str, STR_SHARED|STR_NOFREE)) { + /* make noembed always */ const size_t size = (size_t)capa + termlen; const char *const old_ptr = RSTRING_PTR(str); - const size_t osize = RSTRING(str)->as.heap.len + TERM_LEN(str); - char *new_ptr = ALLOC_N(char, (size_t)capa + termlen); + const size_t osize = RSTRING_LEN(str) + TERM_LEN(str); + char *new_ptr = ALLOC_N(char, size); + if (STR_EMBED_P(str)) RUBY_ASSERT((long)osize <= str_embed_capa(str)); memcpy(new_ptr, old_ptr, osize < size ? osize : size); FL_UNSET_RAW(str, STR_SHARED|STR_NOFREE); RSTRING(str)->as.heap.ptr = new_ptr; - } - else if (STR_HEAP_SIZE(str) != (size_t)capa + termlen) { - SIZED_REALLOC_N(RSTRING(str)->as.heap.ptr, char, - (size_t)capa + termlen, STR_HEAP_SIZE(str)); - } - RSTRING(str)->as.heap.len = len; - TERM_FILL(&RSTRING(str)->as.heap.ptr[len], termlen); - if (n == 1) { - memcpy(RSTRING(str)->as.heap.ptr, RSTRING_PTR(orig), len); - rb_enc_cr_str_exact_copy(str, orig); - } - FL_SET(str, STR_NOEMBED); - RSTRING(str)->as.heap.aux.capa = capa; - } - else if (n == 1) { - rb_str_replace(str, orig); - } - if (enc) { - rb_enc_associate(str, enc); - ENC_CODERANGE_CLEAR(str); - } + } + else if (STR_HEAP_SIZE(str) != (size_t)capa + termlen) { + SIZED_REALLOC_N(RSTRING(str)->as.heap.ptr, char, + (size_t)capa + termlen, STR_HEAP_SIZE(str)); + } + STR_SET_LEN(str, len); + TERM_FILL(&RSTRING(str)->as.heap.ptr[len], termlen); + if (n == 1) { + memcpy(RSTRING(str)->as.heap.ptr, RSTRING_PTR(orig), len); + rb_enc_cr_str_exact_copy(str, orig); + } + FL_SET(str, STR_NOEMBED); + RSTRING(str)->as.heap.aux.capa = capa; + } + else if (n == 1) { + rb_str_replace(str, orig); + } + if (enc) { + rb_enc_associate(str, enc); + ENC_CODERANGE_CLEAR(str); + } } else if (n == 1) { - rb_str_replace(str, orig); + rb_str_replace(str, orig); + } + return str; +} + +/* :nodoc: */ +static VALUE +rb_str_s_new(int argc, VALUE *argv, VALUE klass) +{ + if (klass != rb_cString) { + return rb_class_new_instance_pass_kw(argc, argv, klass); + } + + static ID keyword_ids[2]; + VALUE orig, opt, encoding = Qnil, capacity = Qnil; + VALUE kwargs[2]; + rb_encoding *enc = NULL; + + int n = rb_scan_args(argc, argv, "01:", &orig, &opt); + if (NIL_P(opt)) { + return rb_class_new_instance_pass_kw(argc, argv, klass); + } + + keyword_ids[0] = rb_id_encoding(); + CONST_ID(keyword_ids[1], "capacity"); + rb_get_kwargs(opt, keyword_ids, 0, 2, kwargs); + encoding = kwargs[0]; + capacity = kwargs[1]; + + int termlen = 1; + + if (n == 1) { + orig = StringValue(orig); + } + else { + orig = Qnil; + } + + if (UNDEF_P(encoding)) { + if (!NIL_P(orig)) { + encoding = rb_obj_encoding(orig); + } + } + + if (!UNDEF_P(encoding)) { + enc = rb_to_encoding(encoding); + termlen = rb_enc_mbminlen(enc); + } + + // If capacity is nil, we're basically just duping `orig`. + if (UNDEF_P(capacity)) { + if (NIL_P(orig)) { + VALUE empty_str = str_new(klass, "", 0); + if (enc) { + rb_enc_associate(empty_str, enc); + } + return empty_str; + } + VALUE copy = str_duplicate(klass, orig); + rb_enc_associate(copy, enc); + ENC_CODERANGE_CLEAR(copy); + return copy; } + + long capa = 0; + capa = NUM2LONG(capacity); + if (capa < 0) { + capa = 0; + } + + if (!NIL_P(orig)) { + long orig_capa = rb_str_capacity(orig); + if (orig_capa > capa) { + capa = orig_capa; + } + } + + long fake_len = capa - termlen; + if (fake_len < 0) { + fake_len = 0; + } + + VALUE str = str_new0(klass, NULL, fake_len, termlen); + STR_SET_LEN(str, 0); + TERM_FILL(RSTRING_PTR(str), termlen); + + if (enc) { + rb_enc_associate(str, enc); + } + + if (!NIL_P(orig)) { + rb_str_buf_append(str, orig); + } + return str; } @@ -1951,62 +2067,62 @@ enc_strlen(const char *p, const char *e, rb_encoding *enc, int cr) const char *q; if (rb_enc_mbmaxlen(enc) == rb_enc_mbminlen(enc)) { - long diff = (long)(e - p); - return diff / rb_enc_mbminlen(enc) + !!(diff % rb_enc_mbminlen(enc)); + long diff = (long)(e - p); + return diff / rb_enc_mbminlen(enc) + !!(diff % rb_enc_mbminlen(enc)); } #ifdef NONASCII_MASK else if (cr == ENC_CODERANGE_VALID && enc == rb_utf8_encoding()) { - uintptr_t len = 0; - if ((int)sizeof(uintptr_t) * 2 < e - p) { - const uintptr_t *s, *t; - const uintptr_t lowbits = sizeof(uintptr_t) - 1; - s = (const uintptr_t*)(~lowbits & ((uintptr_t)p + lowbits)); - t = (const uintptr_t*)(~lowbits & (uintptr_t)e); - while (p < (const char *)s) { - if (is_utf8_lead_byte(*p)) len++; - p++; - } - while (s < t) { - len += count_utf8_lead_bytes_with_word(s); - s++; - } - p = (const char *)s; - } - while (p < e) { - if (is_utf8_lead_byte(*p)) len++; - p++; - } - return (long)len; + uintptr_t len = 0; + if ((int)sizeof(uintptr_t) * 2 < e - p) { + const uintptr_t *s, *t; + const uintptr_t lowbits = sizeof(uintptr_t) - 1; + s = (const uintptr_t*)(~lowbits & ((uintptr_t)p + lowbits)); + t = (const uintptr_t*)(~lowbits & (uintptr_t)e); + while (p < (const char *)s) { + if (is_utf8_lead_byte(*p)) len++; + p++; + } + while (s < t) { + len += count_utf8_lead_bytes_with_word(s); + s++; + } + p = (const char *)s; + } + while (p < e) { + if (is_utf8_lead_byte(*p)) len++; + p++; + } + return (long)len; } #endif else if (rb_enc_asciicompat(enc)) { c = 0; - if (ENC_CODERANGE_CLEAN_P(cr)) { - while (p < e) { - if (ISASCII(*p)) { - q = search_nonascii(p, e); - if (!q) - return c + (e - p); - c += q - p; - p = q; - } - p += rb_enc_fast_mbclen(p, e, enc); - c++; - } - } - else { - while (p < e) { - if (ISASCII(*p)) { - q = search_nonascii(p, e); - if (!q) - return c + (e - p); - c += q - p; - p = q; - } - p += rb_enc_mbclen(p, e, enc); - c++; - } - } + if (ENC_CODERANGE_CLEAN_P(cr)) { + while (p < e) { + if (ISASCII(*p)) { + q = search_nonascii(p, e); + if (!q) + return c + (e - p); + c += q - p; + p = q; + } + p += rb_enc_fast_mbclen(p, e, enc); + c++; + } + } + else { + while (p < e) { + if (ISASCII(*p)) { + q = search_nonascii(p, e); + if (!q) + return c + (e - p); + c += q - p; + p = q; + } + p += rb_enc_mbclen(p, e, enc); + c++; + } + } return c; } @@ -2034,49 +2150,49 @@ rb_enc_strlen_cr(const char *p, const char *e, rb_encoding *enc, int *cr) *cr = 0; if (rb_enc_mbmaxlen(enc) == rb_enc_mbminlen(enc)) { - long diff = (long)(e - p); - return diff / rb_enc_mbminlen(enc) + !!(diff % rb_enc_mbminlen(enc)); + long diff = (long)(e - p); + return diff / rb_enc_mbminlen(enc) + !!(diff % rb_enc_mbminlen(enc)); } else if (rb_enc_asciicompat(enc)) { - c = 0; - while (p < e) { - if (ISASCII(*p)) { - q = search_nonascii(p, e); - if (!q) { - if (!*cr) *cr = ENC_CODERANGE_7BIT; - return c + (e - p); - } - c += q - p; - p = q; - } - ret = rb_enc_precise_mbclen(p, e, enc); - if (MBCLEN_CHARFOUND_P(ret)) { - *cr |= ENC_CODERANGE_VALID; - p += MBCLEN_CHARFOUND_LEN(ret); - } - else { - *cr = ENC_CODERANGE_BROKEN; - p++; - } - c++; - } - if (!*cr) *cr = ENC_CODERANGE_7BIT; - return c; + c = 0; + while (p < e) { + if (ISASCII(*p)) { + q = search_nonascii(p, e); + if (!q) { + if (!*cr) *cr = ENC_CODERANGE_7BIT; + return c + (e - p); + } + c += q - p; + p = q; + } + ret = rb_enc_precise_mbclen(p, e, enc); + if (MBCLEN_CHARFOUND_P(ret)) { + *cr |= ENC_CODERANGE_VALID; + p += MBCLEN_CHARFOUND_LEN(ret); + } + else { + *cr = ENC_CODERANGE_BROKEN; + p++; + } + c++; + } + if (!*cr) *cr = ENC_CODERANGE_7BIT; + return c; } for (c=0; p<e; c++) { - ret = rb_enc_precise_mbclen(p, e, enc); - if (MBCLEN_CHARFOUND_P(ret)) { - *cr |= ENC_CODERANGE_VALID; - p += MBCLEN_CHARFOUND_LEN(ret); - } - else { - *cr = ENC_CODERANGE_BROKEN; + ret = rb_enc_precise_mbclen(p, e, enc); + if (MBCLEN_CHARFOUND_P(ret)) { + *cr |= ENC_CODERANGE_VALID; + p += MBCLEN_CHARFOUND_LEN(ret); + } + else { + *cr = ENC_CODERANGE_BROKEN; if (p + rb_enc_mbminlen(enc) <= e) p += rb_enc_mbminlen(enc); else p = e; - } + } } if (!*cr) *cr = ENC_CODERANGE_7BIT; return c; @@ -2096,12 +2212,12 @@ str_strlen(VALUE str, rb_encoding *enc) cr = ENC_CODERANGE(str); if (cr == ENC_CODERANGE_UNKNOWN) { - long n = rb_enc_strlen_cr(p, e, enc, &cr); - if (cr) ENC_CODERANGE_SET(str, cr); - return n; + long n = rb_enc_strlen_cr(p, e, enc, &cr); + if (cr) ENC_CODERANGE_SET(str, cr); + return n; } else { - return enc_strlen(p, e, enc, cr); + return enc_strlen(p, e, enc, cr); } } @@ -2133,7 +2249,7 @@ rb_str_length(VALUE str) * */ -static VALUE +VALUE rb_str_bytesize(VALUE str) { return LONG2NUM(RSTRING_LEN(str)); @@ -2161,7 +2277,7 @@ rb_str_empty(VALUE str) * call-seq: * string + other_string -> new_string * - * Returns a new \String containing +other_string+ concatenated to +self+: + * Returns a new +String+ containing +other_string+ concatenated to +self+: * * "Hello from " + self.to_s # => "Hello from main" * @@ -2182,7 +2298,7 @@ rb_str_plus(VALUE str1, VALUE str2) RSTRING_GETMEM(str2, ptr2, len2); termlen = rb_enc_mbminlen(enc); if (len1 > LONG_MAX - len2) { - rb_raise(rb_eArgError, "string size too big"); + rb_raise(rb_eArgError, "string size too big"); } str3 = str_new0(rb_cString, 0, len1+len2, termlen); ptr3 = RSTRING_PTR(str3); @@ -2191,18 +2307,18 @@ rb_str_plus(VALUE str1, VALUE str2) TERM_FILL(&ptr3[len1+len2], termlen); ENCODING_CODERANGE_SET(str3, rb_enc_to_index(enc), - ENC_CODERANGE_AND(ENC_CODERANGE(str1), ENC_CODERANGE(str2))); + ENC_CODERANGE_AND(ENC_CODERANGE(str1), ENC_CODERANGE(str2))); RB_GC_GUARD(str1); RB_GC_GUARD(str2); return str3; } /* A variant of rb_str_plus that does not raise but return Qundef instead. */ -MJIT_FUNC_EXPORTED VALUE +VALUE rb_str_opt_plus(VALUE str1, VALUE str2) { - assert(RBASIC_CLASS(str1) == rb_cString); - assert(RBASIC_CLASS(str2) == rb_cString); + RUBY_ASSERT(RBASIC_CLASS(str1) == rb_cString); + RUBY_ASSERT(RBASIC_CLASS(str2) == rb_cString); long len1, len2; MAYBE_UNUSED(char) *ptr1, *ptr2; RSTRING_GETMEM(str1, ptr1, len1); @@ -2232,7 +2348,7 @@ rb_str_opt_plus(VALUE str1, VALUE str2) * call-seq: * string * integer -> new_string * - * Returns a new \String containing +integer+ copies of +self+: + * Returns a new +String+ containing +integer+ copies of +self+: * * "Ho! " * 3 # => "Ho! Ho! Ho! " * "Ho! " * 0 # => "" @@ -2252,12 +2368,12 @@ rb_str_times(VALUE str, VALUE times) } if (times == INT2FIX(0)) { str2 = str_alloc_embed(rb_cString, 0); - rb_enc_copy(str2, str); - return str2; + rb_enc_copy(str2, str); + return str2; } len = NUM2LONG(times); if (len < 0) { - rb_raise(rb_eArgError, "negative argument"); + rb_raise(rb_eArgError, "negative argument"); } if (RSTRING_LEN(str) == 1 && RSTRING_PTR(str)[0] == 0) { if (STR_EMBEDDABLE_P(len, 1)) { @@ -2268,14 +2384,13 @@ rb_str_times(VALUE str, VALUE times) str2 = str_alloc_heap(rb_cString); RSTRING(str2)->as.heap.aux.capa = len; RSTRING(str2)->as.heap.ptr = ZALLOC_N(char, (size_t)len + 1); - STR_SET_NOEMBED(str2); } STR_SET_LEN(str2, len); rb_enc_copy(str2, str); return str2; } if (len && LONG_MAX/len < RSTRING_LEN(str)) { - rb_raise(rb_eArgError, "argument too big"); + rb_raise(rb_eArgError, "argument too big"); } len *= RSTRING_LEN(str); @@ -2308,7 +2423,7 @@ rb_str_times(VALUE str, VALUE times) * "%05d" % 123 # => "00123" * * If +self+ contains multiple substitutions, +object+ must be - * an \Array or \Hash containing the values to be substituted: + * an Array or Hash containing the values to be substituted: * * "%-5s: %016x" % [ "ID", self.object_id ] # => "ID : 00002b054ec93168" * "foo = %{foo}" % {foo: 'bar'} # => "foo = bar" @@ -2331,7 +2446,7 @@ static inline void rb_check_lockedtmp(VALUE str) { if (FL_TEST(str, STR_TMPLOCK)) { - rb_raise(rb_eRuntimeError, "can't modify string; temporarily locked"); + rb_raise(rb_eRuntimeError, "can't modify string; temporarily locked"); } } @@ -2346,10 +2461,10 @@ static inline int str_dependent_p(VALUE str) { if (STR_EMBED_P(str) || !FL_TEST(str, STR_SHARED|STR_NOFREE)) { - return 0; + return 0; } else { - return 1; + return 1; } } @@ -2370,18 +2485,18 @@ str_make_independent_expand(VALUE str, long len, long expand, const int termlen) if (len > capa) len = capa; if (!STR_EMBED_P(str) && str_embed_capa(str) >= capa + termlen) { - ptr = RSTRING(str)->as.heap.ptr; - STR_SET_EMBED(str); + ptr = RSTRING(str)->as.heap.ptr; + STR_SET_EMBED(str); memcpy(RSTRING(str)->as.embed.ary, ptr, len); TERM_FILL(RSTRING(str)->as.embed.ary + len, termlen); - STR_SET_EMBED_LEN(str, len); - return; + STR_SET_LEN(str, len); + return; } ptr = ALLOC_N(char, (size_t)capa + termlen); oldptr = RSTRING_PTR(str); if (oldptr) { - memcpy(ptr, oldptr, len); + memcpy(ptr, oldptr, len); } if (FL_TEST_RAW(str, STR_NOEMBED|STR_NOFREE|STR_SHARED) == STR_NOEMBED) { xfree(oldptr); @@ -2390,7 +2505,7 @@ str_make_independent_expand(VALUE str, long len, long expand, const int termlen) FL_UNSET(str, STR_SHARED|STR_NOFREE); TERM_FILL(ptr + len, termlen); RSTRING(str)->as.heap.ptr = ptr; - RSTRING(str)->as.heap.len = len; + STR_SET_LEN(str, len); RSTRING(str)->as.heap.aux.capa = capa; } @@ -2398,7 +2513,7 @@ void rb_str_modify(VALUE str) { if (!str_independent(str)) - str_make_independent(str); + str_make_independent(str); ENC_CODERANGE_CLEAR(str); } @@ -2409,17 +2524,17 @@ rb_str_modify_expand(VALUE str, long expand) long len = RSTRING_LEN(str); if (expand < 0) { - rb_raise(rb_eArgError, "negative expanding string size"); + rb_raise(rb_eArgError, "negative expanding string size"); } if (expand >= LONG_MAX - len) { - rb_raise(rb_eArgError, "string size too big"); + rb_raise(rb_eArgError, "string size too big"); } if (!str_independent(str)) { - str_make_independent_expand(str, len, expand, termlen); + str_make_independent_expand(str, len, expand, termlen); } else if (expand > 0) { - RESIZE_CAPA_TERM(str, len + expand, termlen); + RESIZE_CAPA_TERM(str, len + expand, termlen); } ENC_CODERANGE_CLEAR(str); } @@ -2429,10 +2544,10 @@ static void str_modify_keep_cr(VALUE str) { if (!str_independent(str)) - str_make_independent(str); + str_make_independent(str); if (ENC_CODERANGE(str) == ENC_CODERANGE_BROKEN) - /* Force re-scan later */ - ENC_CODERANGE_CLEAR(str); + /* Force re-scan later */ + ENC_CODERANGE_CLEAR(str); } static inline void @@ -2440,9 +2555,9 @@ str_discard(VALUE str) { str_modifiable(str); if (!STR_EMBED_P(str) && !FL_TEST(str, STR_SHARED|STR_NOFREE)) { - ruby_sized_xfree(STR_HEAP_PTR(str), STR_HEAP_SIZE(str)); - RSTRING(str)->as.heap.ptr = 0; - RSTRING(str)->as.heap.len = 0; + ruby_sized_xfree(STR_HEAP_PTR(str), STR_HEAP_SIZE(str)); + RSTRING(str)->as.heap.ptr = 0; + STR_SET_LEN(str, 0); } } @@ -2450,8 +2565,11 @@ void rb_must_asciicompat(VALUE str) { rb_encoding *enc = rb_enc_get(str); + if (!enc) { + rb_raise(rb_eTypeError, "not encoding capable object"); + } if (!rb_enc_asciicompat(enc)) { - rb_raise(rb_eEncCompatError, "ASCII incompatible encoding: %s", rb_enc_name(enc)); + rb_raise(rb_eEncCompatError, "ASCII incompatible encoding: %s", rb_enc_name(enc)); } } @@ -2460,8 +2578,8 @@ rb_string_value(volatile VALUE *ptr) { VALUE s = *ptr; if (!RB_TYPE_P(s, T_STRING)) { - s = rb_str_to_str(s); - *ptr = s; + s = rb_str_to_str(s); + *ptr = s; } return s; } @@ -2477,7 +2595,7 @@ static int zero_filled(const char *s, int n) { for (; n > 0; --n) { - if (*s++) return 0; + if (*s++) return 0; } return 1; } @@ -2488,7 +2606,7 @@ str_null_char(const char *s, long len, const int minlen, rb_encoding *enc) const char *e = s + len; for (; s + minlen <= e; s += rb_enc_mbclen(s, e, enc)) { - if (zero_filled(s, minlen)) return s; + if (zero_filled(s, minlen)) return s; } return 0; } @@ -2500,12 +2618,12 @@ str_fill_term(VALUE str, char *s, long len, int termlen) * is allocated, like many other functions in this file. */ if (str_dependent_p(str)) { - if (!zero_filled(s + len, termlen)) - str_make_independent_expand(str, len, 0L, termlen); + if (!zero_filled(s + len, termlen)) + str_make_independent_expand(str, len, 0L, termlen); } else { - TERM_FILL(s + len, termlen); - return s; + TERM_FILL(s + len, termlen); + return s; } return RSTRING_PTR(str); } @@ -2516,24 +2634,24 @@ rb_str_change_terminator_length(VALUE str, const int oldtermlen, const int terml long capa = str_capacity(str, oldtermlen) + oldtermlen; long len = RSTRING_LEN(str); - assert(capa >= len); + RUBY_ASSERT(capa >= len); if (capa - len < termlen) { - rb_check_lockedtmp(str); - str_make_independent_expand(str, len, 0L, termlen); + rb_check_lockedtmp(str); + str_make_independent_expand(str, len, 0L, termlen); } else if (str_dependent_p(str)) { - if (termlen > oldtermlen) - str_make_independent_expand(str, len, 0L, termlen); + if (termlen > oldtermlen) + str_make_independent_expand(str, len, 0L, termlen); } else { - if (!STR_EMBED_P(str)) { - /* modify capa instead of realloc */ - assert(!FL_TEST((str), STR_SHARED)); - RSTRING(str)->as.heap.aux.capa = capa - termlen; - } - if (termlen > oldtermlen) { - TERM_FILL(RSTRING_PTR(str) + len, termlen); - } + if (!STR_EMBED_P(str)) { + /* modify capa instead of realloc */ + RUBY_ASSERT(!FL_TEST((str), STR_SHARED)); + RSTRING(str)->as.heap.aux.capa = capa - termlen; + } + if (termlen > oldtermlen) { + TERM_FILL(RSTRING_PTR(str) + len, termlen); + } } return; @@ -2548,18 +2666,18 @@ str_null_check(VALUE str, int *w) const int minlen = rb_enc_mbminlen(enc); if (minlen > 1) { - *w = 1; - if (str_null_char(s, len, minlen, enc)) { - return NULL; - } - return str_fill_term(str, s, len, minlen); + *w = 1; + if (str_null_char(s, len, minlen, enc)) { + return NULL; + } + return str_fill_term(str, s, len, minlen); } *w = 0; if (!s || memchr(s, 0, len)) { - return NULL; + return NULL; } if (s[len]) { - s = str_fill_term(str, s, len, minlen); + s = str_fill_term(str, s, len, minlen); } return s; } @@ -2578,10 +2696,10 @@ rb_string_value_cstr(volatile VALUE *ptr) int w; char *s = str_null_check(str, &w); if (!s) { - if (w) { - rb_raise(rb_eArgError, "string contains null char"); - } - rb_raise(rb_eArgError, "string contains null byte"); + if (w) { + rb_raise(rb_eArgError, "string contains null char"); + } + rb_raise(rb_eArgError, "string contains null byte"); } return s; } @@ -2605,14 +2723,14 @@ rb_check_string_type(VALUE str) * call-seq: * String.try_convert(object) -> object, new_string, or nil * - * If +object+ is a \String object, returns +object+. + * If +object+ is a +String+ object, returns +object+. * * Otherwise if +object+ responds to <tt>:to_str</tt>, * calls <tt>object.to_str</tt> and returns the result. * * Returns +nil+ if +object+ does not respond to <tt>:to_str</tt>. * - * Raises an exception unless <tt>object.to_str</tt> returns a \String object. + * Raises an exception unless <tt>object.to_str</tt> returns a +String+ object. */ static VALUE rb_str_s_try_convert(VALUE dummy, VALUE str) @@ -2643,8 +2761,8 @@ str_nth_len(const char *p, const char *e, long *nthp, rb_encoding *enc) if (ISASCII(*p)) { p2 = search_nonascii(p, e2); if (!p2) { - nth -= e2 - p; - *nthp = nth; + nth -= e2 - p; + *nthp = nth; return (char *)e2; } nth -= p2 - p; @@ -2680,9 +2798,9 @@ static char* str_nth(const char *p, const char *e, long nth, rb_encoding *enc, int singlebyte) { if (singlebyte) - p += nth; + p += nth; else { - p = str_nth_len(p, e, &nth, enc); + p = str_nth_len(p, e, &nth, enc); } if (!p) return 0; if (p > e) p = e; @@ -2702,7 +2820,7 @@ long rb_str_offset(VALUE str, long pos) { return str_offset(RSTRING_PTR(str), RSTRING_END(str), pos, - STR_ENC_GET(str), single_byte_optimizable(str)); + STR_ENC_GET(str), single_byte_optimizable(str)); } #ifdef NONASCII_MASK @@ -2711,26 +2829,26 @@ str_utf8_nth(const char *p, const char *e, long *nthp) { long nth = *nthp; if ((int)SIZEOF_VOIDP * 2 < e - p && (int)SIZEOF_VOIDP * 2 < nth) { - const uintptr_t *s, *t; - const uintptr_t lowbits = SIZEOF_VOIDP - 1; - s = (const uintptr_t*)(~lowbits & ((uintptr_t)p + lowbits)); - t = (const uintptr_t*)(~lowbits & (uintptr_t)e); - while (p < (const char *)s) { - if (is_utf8_lead_byte(*p)) nth--; - p++; - } - do { - nth -= count_utf8_lead_bytes_with_word(s); - s++; - } while (s < t && (int)SIZEOF_VOIDP <= nth); - p = (char *)s; + const uintptr_t *s, *t; + const uintptr_t lowbits = SIZEOF_VOIDP - 1; + s = (const uintptr_t*)(~lowbits & ((uintptr_t)p + lowbits)); + t = (const uintptr_t*)(~lowbits & (uintptr_t)e); + while (p < (const char *)s) { + if (is_utf8_lead_byte(*p)) nth--; + p++; + } + do { + nth -= count_utf8_lead_bytes_with_word(s); + s++; + } while (s < t && (int)SIZEOF_VOIDP <= nth); + p = (char *)s; } while (p < e) { - if (is_utf8_lead_byte(*p)) { - if (nth == 0) break; - nth--; - } - p++; + if (is_utf8_lead_byte(*p)) { + if (nth == 0) break; + nth--; + } + p++; } *nthp = nth; return (char *)p; @@ -2751,31 +2869,55 @@ rb_str_sublen(VALUE str, long pos) if (single_byte_optimizable(str) || pos < 0) return pos; else { - char *p = RSTRING_PTR(str); + char *p = RSTRING_PTR(str); return enc_strlen(p, p + pos, STR_ENC_GET(str), ENC_CODERANGE(str)); } } -VALUE -rb_str_subseq(VALUE str, long beg, long len) +static VALUE +str_subseq(VALUE str, long beg, long len) { VALUE str2; - if (!STR_EMBEDDABLE_P(len, TERM_LEN(str)) && - SHARABLE_SUBSTRING_P(beg, len, RSTRING_LEN(str))) { - long olen; - str2 = rb_str_new_shared(rb_str_new_frozen_String(str)); - RSTRING(str2)->as.heap.ptr += beg; - olen = RSTRING(str2)->as.heap.len; - if (olen > len) RSTRING(str2)->as.heap.len = len; + RUBY_ASSERT(beg >= 0); + RUBY_ASSERT(len >= 0); + RUBY_ASSERT(beg+len <= RSTRING_LEN(str)); + + const int termlen = TERM_LEN(str); + if (!SHARABLE_SUBSTRING_P(beg, len, RSTRING_LEN(str))) { + str2 = rb_str_new(RSTRING_PTR(str) + beg, len); + RB_GC_GUARD(str); + return str2; } - else { - str2 = rb_str_new(RSTRING_PTR(str)+beg, len); + + str2 = str_alloc_heap(rb_cString); + if (str_embed_capa(str2) >= len + termlen) { + char *ptr2 = RSTRING(str2)->as.embed.ary; + STR_SET_EMBED(str2); + memcpy(ptr2, RSTRING_PTR(str) + beg, len); + TERM_FILL(ptr2+len, termlen); + + STR_SET_LEN(str2, len); RB_GC_GUARD(str); } + else { + str_replace_shared(str2, str); + RUBY_ASSERT(!STR_EMBED_P(str2)); + ENC_CODERANGE_CLEAR(str2); + RSTRING(str2)->as.heap.ptr += beg; + if (RSTRING_LEN(str2) > len) { + STR_SET_LEN(str2, len); + } + } - rb_enc_cr_str_copy_for_substr(str2, str); + return str2; +} +VALUE +rb_str_subseq(VALUE str, long beg, long len) +{ + VALUE str2 = str_subseq(str, beg, len); + rb_enc_cr_str_copy_for_substr(str2, str); return str2; } @@ -2790,46 +2932,46 @@ rb_str_subpos(VALUE str, long beg, long *lenp) if (len < 0) return 0; if (!blen) { - len = 0; + len = 0; } if (single_byte_optimizable(str)) { - if (beg > blen) return 0; - if (beg < 0) { - beg += blen; - if (beg < 0) return 0; - } - if (len > blen - beg) - len = blen - beg; - if (len < 0) return 0; - p = s + beg; - goto end; + if (beg > blen) return 0; + if (beg < 0) { + beg += blen; + if (beg < 0) return 0; + } + if (len > blen - beg) + len = blen - beg; + if (len < 0) return 0; + p = s + beg; + goto end; } if (beg < 0) { - if (len > -beg) len = -beg; - if (-beg * rb_enc_mbmaxlen(enc) < RSTRING_LEN(str) / 8) { - beg = -beg; - while (beg-- > len && (e = rb_enc_prev_char(s, e, e, enc)) != 0); - p = e; - if (!p) return 0; - while (len-- > 0 && (p = rb_enc_prev_char(s, p, e, enc)) != 0); - if (!p) return 0; - len = e - p; - goto end; - } - else { - slen = str_strlen(str, enc); - beg += slen; - if (beg < 0) return 0; - p = s + beg; - if (len == 0) goto end; - } + if (len > -beg) len = -beg; + if (-beg * rb_enc_mbmaxlen(enc) < RSTRING_LEN(str) / 8) { + beg = -beg; + while (beg-- > len && (e = rb_enc_prev_char(s, e, e, enc)) != 0); + p = e; + if (!p) return 0; + while (len-- > 0 && (p = rb_enc_prev_char(s, p, e, enc)) != 0); + if (!p) return 0; + len = e - p; + goto end; + } + else { + slen = str_strlen(str, enc); + beg += slen; + if (beg < 0) return 0; + p = s + beg; + if (len == 0) goto end; + } } else if (beg > 0 && beg > RSTRING_LEN(str)) { - return 0; + return 0; } if (len == 0) { - if (beg > str_strlen(str, enc)) return 0; /* str's enc */ - p = s + beg; + if (beg > str_strlen(str, enc)) return 0; /* str's enc */ + p = s + beg; } #ifdef NONASCII_MASK else if (ENC_CODERANGE(str) == ENC_CODERANGE_VALID && @@ -2840,23 +2982,23 @@ rb_str_subpos(VALUE str, long beg, long *lenp) } #endif else if (rb_enc_mbmaxlen(enc) == rb_enc_mbminlen(enc)) { - int char_sz = rb_enc_mbmaxlen(enc); + int char_sz = rb_enc_mbmaxlen(enc); - p = s + beg * char_sz; - if (p > e) { - return 0; - } + p = s + beg * char_sz; + if (p > e) { + return 0; + } else if (len * char_sz > e - p) len = e - p; else - len *= char_sz; + len *= char_sz; } else if ((p = str_nth_len(s, e, &beg, enc)) == e) { - if (beg > 0) return 0; - len = 0; + if (beg > 0) return 0; + len = 0; } else { - len = str_offset(p, e, len, enc, 0); + len = str_offset(p, e, len, enc, 0); } end: *lenp = len; @@ -2875,38 +3017,31 @@ rb_str_substr(VALUE str, long beg, long len) static VALUE str_substr(VALUE str, long beg, long len, int empty) { - VALUE str2; char *p = rb_str_subpos(str, beg, &len); if (!p) return Qnil; - if (!STR_EMBEDDABLE_P(len, TERM_LEN(str)) && - SHARABLE_SUBSTRING_P(p, len, RSTRING_END(str))) { - long ofs = p - RSTRING_PTR(str); - str2 = rb_str_new_frozen(str); - str2 = str_new_shared(rb_cString, str2); - RSTRING(str2)->as.heap.ptr += ofs; - RSTRING(str2)->as.heap.len = len; - ENC_CODERANGE_CLEAR(str2); - } - else { - if (!len && !empty) return Qnil; - str2 = rb_str_new(p, len); - RB_GC_GUARD(str); - } - rb_enc_cr_str_copy_for_substr(str2, str); + if (!len && !empty) return Qnil; + beg = p - RSTRING_PTR(str); + + VALUE str2 = str_subseq(str, beg, len); + rb_enc_cr_str_copy_for_substr(str2, str); return str2; } +/* :nodoc: */ VALUE rb_str_freeze(VALUE str) { + if (CHILLED_STRING_P(str)) { + FL_UNSET_RAW(str, STR_CHILLED); + } + if (OBJ_FROZEN(str)) return str; rb_str_resize(str, RSTRING_LEN(str)); return rb_obj_freeze(str); } - /* * call-seq: * +string -> new_string or self @@ -2919,23 +3054,33 @@ static VALUE str_uplus(VALUE str) { if (OBJ_FROZEN(str)) { - return rb_str_dup(str); + return rb_str_dup(str); } else { - return str; + return str; } } /* * call-seq: * -string -> frozen_string + * dedup -> frozen_string * * Returns a frozen, possibly pre-existing copy of the string. * - * The returned \String will be deduplicated as long as it does not have + * The returned +String+ will be deduplicated as long as it does not have * any instance variables set on it and is not a String subclass. * - * String#dedup is an alias for String#-@. + * Note that <tt>-string</tt> variant is more convenient for defining + * constants: + * + * FILENAME = -'config/database.yml' + * + * while +dedup+ is better suitable for using the method in chains + * of calculations: + * + * @url_list.concat(urls.map(&:dedup)) + * */ static VALUE str_uminus(VALUE str) @@ -2953,7 +3098,7 @@ VALUE rb_str_locktmp(VALUE str) { if (FL_TEST(str, STR_TMPLOCK)) { - rb_raise(rb_eRuntimeError, "temporal locking already locked string"); + rb_raise(rb_eRuntimeError, "temporal locking already locked string"); } FL_SET(str, STR_TMPLOCK); return str; @@ -2963,13 +3108,13 @@ VALUE rb_str_unlocktmp(VALUE str) { if (!FL_TEST(str, STR_TMPLOCK)) { - rb_raise(rb_eRuntimeError, "temporal unlocking already unlocked string"); + rb_raise(rb_eRuntimeError, "temporal unlocking already unlocked string"); } FL_UNSET(str, STR_TMPLOCK); return str; } -RUBY_FUNC_EXPORTED VALUE +VALUE rb_str_locktmp_ensure(VALUE str, VALUE (*func)(VALUE), VALUE arg) { rb_str_locktmp(str); @@ -2984,11 +3129,38 @@ rb_str_set_len(VALUE str, long len) str_modifiable(str); if (STR_SHARED_P(str)) { - rb_raise(rb_eRuntimeError, "can't set length of shared string"); + rb_raise(rb_eRuntimeError, "can't set length of shared string"); } if (len > (capa = (long)str_capacity(str, termlen)) || len < 0) { - rb_bug("probable buffer overflow: %ld for %ld", len, capa); + rb_bug("probable buffer overflow: %ld for %ld", len, capa); } + + int cr = ENC_CODERANGE(str); + if (cr == ENC_CODERANGE_UNKNOWN) { + /* Leave unknown. */ + } + else if (len > RSTRING_LEN(str)) { + if (ENC_CODERANGE_CLEAN_P(cr)) { + /* Update the coderange regarding the extended part. */ + const char *const prev_end = RSTRING_END(str); + const char *const new_end = RSTRING_PTR(str) + len; + rb_encoding *enc = rb_enc_get(str); + rb_str_coderange_scan_restartable(prev_end, new_end, enc, &cr); + ENC_CODERANGE_SET(str, cr); + } + else if (cr == ENC_CODERANGE_BROKEN) { + /* May be valid now, by appended part. */ + ENC_CODERANGE_SET(str, ENC_CODERANGE_UNKNOWN); + } + } + else if (len < RSTRING_LEN(str)) { + if (cr != ENC_CODERANGE_7BIT) { + /* ASCII-only string is keeping after truncated. Valid + * and broken may be invalid or valid, leave unknown. */ + ENC_CODERANGE_SET(str, ENC_CODERANGE_UNKNOWN); + } + } + STR_SET_LEN(str, len); TERM_FILL(&RSTRING_PTR(str)[len], termlen); } @@ -2996,95 +3168,91 @@ rb_str_set_len(VALUE str, long len) VALUE rb_str_resize(VALUE str, long len) { - long slen; - int independent; - if (len < 0) { - rb_raise(rb_eArgError, "negative string size (or size too big)"); + rb_raise(rb_eArgError, "negative string size (or size too big)"); } - independent = str_independent(str); - ENC_CODERANGE_CLEAR(str); - slen = RSTRING_LEN(str); + int independent = str_independent(str); + long slen = RSTRING_LEN(str); + + if (slen > len && ENC_CODERANGE(str) != ENC_CODERANGE_7BIT) { + ENC_CODERANGE_CLEAR(str); + } { - long capa; - const int termlen = TERM_LEN(str); - if (STR_EMBED_P(str)) { - if (len == slen) return str; + long capa; + const int termlen = TERM_LEN(str); + if (STR_EMBED_P(str)) { + if (len == slen) return str; if (str_embed_capa(str) >= len + termlen) { - STR_SET_EMBED_LEN(str, len); + STR_SET_LEN(str, len); TERM_FILL(RSTRING(str)->as.embed.ary + len, termlen); - return str; - } - str_make_independent_expand(str, slen, len - slen, termlen); - } + return str; + } + str_make_independent_expand(str, slen, len - slen, termlen); + } else if (str_embed_capa(str) >= len + termlen) { - char *ptr = STR_HEAP_PTR(str); - STR_SET_EMBED(str); - if (slen > len) slen = len; + char *ptr = STR_HEAP_PTR(str); + STR_SET_EMBED(str); + if (slen > len) slen = len; if (slen > 0) MEMCPY(RSTRING(str)->as.embed.ary, ptr, char, slen); TERM_FILL(RSTRING(str)->as.embed.ary + len, termlen); - STR_SET_EMBED_LEN(str, len); - if (independent) ruby_xfree(ptr); - return str; - } - else if (!independent) { - if (len == slen) return str; - str_make_independent_expand(str, slen, len - slen, termlen); - } - else if ((capa = RSTRING(str)->as.heap.aux.capa) < len || - (capa - len) > (len < 1024 ? len : 1024)) { - SIZED_REALLOC_N(RSTRING(str)->as.heap.ptr, char, - (size_t)len + termlen, STR_HEAP_SIZE(str)); - RSTRING(str)->as.heap.aux.capa = len; - } - else if (len == slen) return str; - RSTRING(str)->as.heap.len = len; - TERM_FILL(RSTRING(str)->as.heap.ptr + len, termlen); /* sentinel */ + STR_SET_LEN(str, len); + if (independent) ruby_xfree(ptr); + return str; + } + else if (!independent) { + if (len == slen) return str; + str_make_independent_expand(str, slen, len - slen, termlen); + } + else if ((capa = RSTRING(str)->as.heap.aux.capa) < len || + (capa - len) > (len < 1024 ? len : 1024)) { + SIZED_REALLOC_N(RSTRING(str)->as.heap.ptr, char, + (size_t)len + termlen, STR_HEAP_SIZE(str)); + RSTRING(str)->as.heap.aux.capa = len; + } + else if (len == slen) return str; + STR_SET_LEN(str, len); + TERM_FILL(RSTRING(str)->as.heap.ptr + len, termlen); /* sentinel */ } return str; } static VALUE -str_buf_cat(VALUE str, const char *ptr, long len) +str_buf_cat4(VALUE str, const char *ptr, long len, bool keep_cr) { - long capa, total, olen, off = -1; + if (keep_cr) { + str_modify_keep_cr(str); + } + else { + rb_str_modify(str); + } + if (len == 0) return 0; + + long total, olen, off = -1; char *sptr; const int termlen = TERM_LEN(str); -#if !USE_RVARGC - assert(termlen < RSTRING_EMBED_LEN_MAX + 1); /* < (LONG_MAX/2) */ -#endif RSTRING_GETMEM(str, sptr, olen); if (ptr >= sptr && ptr <= sptr + olen) { off = ptr - sptr; } - rb_str_modify(str); - if (len == 0) return 0; - if (STR_EMBED_P(str)) { - capa = str_embed_capa(str) - termlen; - sptr = RSTRING(str)->as.embed.ary; - olen = RSTRING_EMBED_LEN(str); - } - else { - capa = RSTRING(str)->as.heap.aux.capa; - sptr = RSTRING(str)->as.heap.ptr; - olen = RSTRING(str)->as.heap.len; - } + + long capa = str_capacity(str, termlen); + if (olen > LONG_MAX - len) { - rb_raise(rb_eArgError, "string sizes too big"); + rb_raise(rb_eArgError, "string sizes too big"); } total = olen + len; if (capa < total) { - if (total >= LONG_MAX / 2) { - capa = total; - } - while (total > capa) { - capa = 2 * capa + termlen; /* == 2*(capa+termlen)-termlen */ - } - RESIZE_CAPA_TERM(str, capa, termlen); - sptr = RSTRING_PTR(str); + if (total >= LONG_MAX / 2) { + capa = total; + } + while (total > capa) { + capa = 2 * capa + termlen; /* == 2*(capa+termlen)-termlen */ + } + RESIZE_CAPA_TERM(str, capa, termlen); + sptr = RSTRING_PTR(str); } if (off != -1) { ptr = sptr + off; @@ -3096,14 +3264,15 @@ str_buf_cat(VALUE str, const char *ptr, long len) return str; } -#define str_buf_cat2(str, ptr) str_buf_cat((str), (ptr), rb_strlen_lit(ptr)) +#define str_buf_cat(str, ptr, len) str_buf_cat4((str), (ptr), len, false) +#define str_buf_cat2(str, ptr) str_buf_cat4((str), (ptr), rb_strlen_lit(ptr), false) VALUE rb_str_cat(VALUE str, const char *ptr, long len) { if (len == 0) return str; if (len < 0) { - rb_raise(rb_eArgError, "negative string size (or size too big)"); + rb_raise(rb_eArgError, "negative string size (or size too big)"); } return str_buf_cat(str, ptr, len); } @@ -3131,13 +3300,13 @@ rb_enc_cr_str_buf_cat(VALUE str, const char *ptr, long len, str_cr = RSTRING_LEN(str) ? ENC_CODERANGE(str) : ENC_CODERANGE_7BIT; if (str_encindex == ptr_encindex) { - if (str_cr != ENC_CODERANGE_UNKNOWN && ptr_cr == ENC_CODERANGE_UNKNOWN) { + if (str_cr != ENC_CODERANGE_UNKNOWN && ptr_cr == ENC_CODERANGE_UNKNOWN) { ptr_cr = coderange_scan(ptr, len, rb_enc_from_index(ptr_encindex)); } } else { - str_enc = rb_enc_from_index(str_encindex); - ptr_enc = rb_enc_from_index(ptr_encindex); + str_enc = rb_enc_from_index(str_encindex); + ptr_enc = rb_enc_from_index(ptr_encindex); if (!rb_enc_asciicompat(str_enc) || !rb_enc_asciicompat(ptr_enc)) { if (len == 0) return str; @@ -3149,9 +3318,9 @@ rb_enc_cr_str_buf_cat(VALUE str, const char *ptr, long len, } goto incompatible; } - if (ptr_cr == ENC_CODERANGE_UNKNOWN) { - ptr_cr = coderange_scan(ptr, len, ptr_enc); - } + if (ptr_cr == ENC_CODERANGE_UNKNOWN) { + ptr_cr = coderange_scan(ptr, len, ptr_enc); + } if (str_cr == ENC_CODERANGE_UNKNOWN) { if (ENCODING_IS_ASCII8BIT(str) || ptr_cr != ENC_CODERANGE_7BIT) { str_cr = rb_enc_str_coderange(str); @@ -3164,8 +3333,8 @@ rb_enc_cr_str_buf_cat(VALUE str, const char *ptr, long len, if (str_encindex != ptr_encindex && str_cr != ENC_CODERANGE_7BIT && ptr_cr != ENC_CODERANGE_7BIT) { - str_enc = rb_enc_from_index(str_encindex); - ptr_enc = rb_enc_from_index(ptr_encindex); + str_enc = rb_enc_from_index(str_encindex); + ptr_enc = rb_enc_from_index(ptr_encindex); goto incompatible; } @@ -3185,10 +3354,10 @@ rb_enc_cr_str_buf_cat(VALUE str, const char *ptr, long len, } else if (str_cr == ENC_CODERANGE_VALID) { res_encindex = str_encindex; - if (ENC_CODERANGE_CLEAN_P(ptr_cr)) - res_cr = str_cr; - else - res_cr = ptr_cr; + if (ENC_CODERANGE_CLEAN_P(ptr_cr)) + res_cr = str_cr; + else + res_cr = ptr_cr; } else { /* str_cr == ENC_CODERANGE_BROKEN */ res_encindex = str_encindex; @@ -3197,7 +3366,7 @@ rb_enc_cr_str_buf_cat(VALUE str, const char *ptr, long len, } if (len < 0) { - rb_raise(rb_eArgError, "negative string size (or size too big)"); + rb_raise(rb_eArgError, "negative string size (or size too big)"); } str_buf_cat(str, ptr, len); ENCODING_CODERANGE_SET(str, res_encindex, res_cr); @@ -3205,7 +3374,7 @@ rb_enc_cr_str_buf_cat(VALUE str, const char *ptr, long len, incompatible: rb_raise(rb_eEncCompatError, "incompatible character encodings: %s and %s", - rb_enc_name(str_enc), rb_enc_name(ptr_enc)); + rb_enc_inspect_name(str_enc), rb_enc_inspect_name(ptr_enc)); UNREACHABLE_RETURN(Qundef); } @@ -3243,9 +3412,28 @@ rb_str_buf_cat_ascii(VALUE str, const char *ptr) VALUE rb_str_buf_append(VALUE str, VALUE str2) { - int str2_cr; + int str2_cr = rb_enc_str_coderange(str2); - str2_cr = ENC_CODERANGE(str2); + if (str_enc_fastpath(str)) { + switch (str2_cr) { + case ENC_CODERANGE_7BIT: + // If RHS is 7bit we can do simple concatenation + str_buf_cat4(str, RSTRING_PTR(str2), RSTRING_LEN(str2), true); + RB_GC_GUARD(str2); + return str; + case ENC_CODERANGE_VALID: + // If RHS is valid, we can do simple concatenation if encodings are the same + if (ENCODING_GET_INLINED(str) == ENCODING_GET_INLINED(str2)) { + str_buf_cat4(str, RSTRING_PTR(str2), RSTRING_LEN(str2), true); + int str_cr = ENC_CODERANGE(str); + if (UNLIKELY(str_cr != ENC_CODERANGE_VALID)) { + ENC_CODERANGE_SET(str, RB_ENC_CODERANGE_AND(str_cr, str2_cr)); + } + RB_GC_GUARD(str2); + return str; + } + } + } rb_enc_cr_str_buf_cat(str, RSTRING_PTR(str2), RSTRING_LEN(str2), ENCODING_GET(str2), str2_cr, &str2_cr); @@ -3262,39 +3450,29 @@ rb_str_append(VALUE str, VALUE str2) return rb_str_buf_append(str, str2); } -#define MIN_PRE_ALLOC_SIZE 48 - -MJIT_FUNC_EXPORTED VALUE +VALUE rb_str_concat_literals(size_t num, const VALUE *strary) { VALUE str; - size_t i, s; - long len = 1; + size_t i, s = 0; + unsigned long len = 1; if (UNLIKELY(!num)) return rb_str_new(0, 0); if (UNLIKELY(num == 1)) return rb_str_resurrect(strary[0]); for (i = 0; i < num; ++i) { len += RSTRING_LEN(strary[i]); } - if (LIKELY(len < MIN_PRE_ALLOC_SIZE)) { - str = rb_str_resurrect(strary[0]); - s = 1; - } - else { - str = rb_str_buf_new(len); - rb_enc_copy(str, strary[0]); - s = 0; - } + str = rb_str_buf_new(len); + str_enc_copy_direct(str, strary[0]); for (i = s; i < num; ++i) { - const VALUE v = strary[i]; - int encidx = ENCODING_GET(v); + const VALUE v = strary[i]; + int encidx = ENCODING_GET(v); - rb_enc_cr_str_buf_cat(str, RSTRING_PTR(v), RSTRING_LEN(v), - encidx, ENC_CODERANGE(v), NULL); - if (encidx != ENCINDEX_US_ASCII) { - if (ENCODING_GET_INLINED(str) == ENCINDEX_US_ASCII) - rb_enc_set_index(str, encidx); - } + rb_str_buf_append(str, v); + if (encidx != ENCINDEX_US_ASCII) { + if (ENCODING_GET_INLINED(str) == ENCINDEX_US_ASCII) + rb_enc_set_index(str, encidx); + } } return str; } @@ -3309,7 +3487,7 @@ rb_str_concat_literals(size_t num, const VALUE *strary) * s.concat('bar', 'baz') # => "foobarbaz" * s # => "foobarbaz" * - * For each given object +object+ that is an \Integer, + * For each given object +object+ that is an Integer, * the value is considered a codepoint and converted to a character before concatenation: * * s = 'foo' @@ -3323,16 +3501,16 @@ rb_str_concat_multi(int argc, VALUE *argv, VALUE str) str_modifiable(str); if (argc == 1) { - return rb_str_concat(str, argv[0]); + return rb_str_concat(str, argv[0]); } else if (argc > 1) { - int i; - VALUE arg_str = rb_str_tmp_new(0); - rb_enc_copy(arg_str, str); - for (i = 0; i < argc; i++) { - rb_str_concat(arg_str, argv[i]); - } - rb_str_buf_append(str, arg_str); + int i; + VALUE arg_str = rb_str_tmp_new(0); + rb_enc_copy(arg_str, str); + for (i = 0; i < argc; i++) { + rb_str_concat(arg_str, argv[i]); + } + rb_str_buf_append(str, arg_str); } return str; @@ -3348,7 +3526,7 @@ rb_str_concat_multi(int argc, VALUE *argv, VALUE str) * s << 'bar' # => "foobar" * s # => "foobar" * - * If +object+ is an \Integer, + * If +object+ is an Integer, * the value is considered a codepoint and converted to a character before concatenation: * * s = 'foo' @@ -3364,62 +3542,82 @@ rb_str_concat(VALUE str1, VALUE str2) int encidx; if (RB_INTEGER_TYPE_P(str2)) { - if (rb_num_to_uint(str2, &code) == 0) { - } - else if (FIXNUM_P(str2)) { - rb_raise(rb_eRangeError, "%ld out of char range", FIX2LONG(str2)); - } - else { - rb_raise(rb_eRangeError, "bignum out of char range"); - } + if (rb_num_to_uint(str2, &code) == 0) { + } + else if (FIXNUM_P(str2)) { + rb_raise(rb_eRangeError, "%ld out of char range", FIX2LONG(str2)); + } + else { + rb_raise(rb_eRangeError, "bignum out of char range"); + } } else { - return rb_str_append(str1, str2); + return rb_str_append(str1, str2); } - encidx = rb_enc_to_index(enc); - if (encidx == ENCINDEX_ASCII || encidx == ENCINDEX_US_ASCII) { - /* US-ASCII automatically extended to ASCII-8BIT */ - char buf[1]; - buf[0] = (char)code; - if (code > 0xFF) { - rb_raise(rb_eRangeError, "%u out of char range", code); - } - rb_str_cat(str1, buf, 1); - if (encidx == ENCINDEX_US_ASCII && code > 127) { - rb_enc_associate_index(str1, ENCINDEX_ASCII); - ENC_CODERANGE_SET(str1, ENC_CODERANGE_VALID); - } + encidx = rb_ascii8bit_appendable_encoding_index(enc, code); + if (encidx >= 0) { + char buf[1]; + buf[0] = (char)code; + rb_str_cat(str1, buf, 1); + if (encidx != rb_enc_to_index(enc)) { + rb_enc_associate_index(str1, encidx); + ENC_CODERANGE_SET(str1, ENC_CODERANGE_VALID); + } } else { - long pos = RSTRING_LEN(str1); - int cr = ENC_CODERANGE(str1); - int len; - char *buf; - - switch (len = rb_enc_codelen(code, enc)) { - case ONIGERR_INVALID_CODE_POINT_VALUE: - rb_raise(rb_eRangeError, "invalid codepoint 0x%X in %s", code, rb_enc_name(enc)); - break; - case ONIGERR_TOO_BIG_WIDE_CHAR_VALUE: - case 0: - rb_raise(rb_eRangeError, "%u out of char range", code); - break; - } - buf = ALLOCA_N(char, len + 1); - rb_enc_mbcput(code, buf, enc); - if (rb_enc_precise_mbclen(buf, buf + len + 1, enc) != len) { - rb_raise(rb_eRangeError, "invalid codepoint 0x%X in %s", code, rb_enc_name(enc)); - } - rb_str_resize(str1, pos+len); - memcpy(RSTRING_PTR(str1) + pos, buf, len); - if (cr == ENC_CODERANGE_7BIT && code > 127) - cr = ENC_CODERANGE_VALID; - ENC_CODERANGE_SET(str1, cr); + long pos = RSTRING_LEN(str1); + int cr = ENC_CODERANGE(str1); + int len; + char *buf; + + switch (len = rb_enc_codelen(code, enc)) { + case ONIGERR_INVALID_CODE_POINT_VALUE: + rb_raise(rb_eRangeError, "invalid codepoint 0x%X in %s", code, rb_enc_name(enc)); + break; + case ONIGERR_TOO_BIG_WIDE_CHAR_VALUE: + case 0: + rb_raise(rb_eRangeError, "%u out of char range", code); + break; + } + buf = ALLOCA_N(char, len + 1); + rb_enc_mbcput(code, buf, enc); + if (rb_enc_precise_mbclen(buf, buf + len + 1, enc) != len) { + rb_raise(rb_eRangeError, "invalid codepoint 0x%X in %s", code, rb_enc_name(enc)); + } + rb_str_resize(str1, pos+len); + memcpy(RSTRING_PTR(str1) + pos, buf, len); + if (cr == ENC_CODERANGE_7BIT && code > 127) { + cr = ENC_CODERANGE_VALID; + } + else if (cr == ENC_CODERANGE_BROKEN) { + cr = ENC_CODERANGE_UNKNOWN; + } + ENC_CODERANGE_SET(str1, cr); } return str1; } +int +rb_ascii8bit_appendable_encoding_index(rb_encoding *enc, unsigned int code) +{ + int encidx = rb_enc_to_index(enc); + + if (encidx == ENCINDEX_ASCII_8BIT || encidx == ENCINDEX_US_ASCII) { + /* US-ASCII automatically extended to ASCII-8BIT */ + if (code > 0xFF) { + rb_raise(rb_eRangeError, "%u out of char range", code); + } + if (encidx == ENCINDEX_US_ASCII && code > 127) { + return ENCINDEX_ASCII_8BIT; + } + return encidx; + } + else { + return -1; + } +} + /* * call-seq: * prepend(*other_strings) -> string @@ -3439,16 +3637,16 @@ rb_str_prepend_multi(int argc, VALUE *argv, VALUE str) str_modifiable(str); if (argc == 1) { - rb_str_update(str, 0L, 0L, argv[0]); + rb_str_update(str, 0L, 0L, argv[0]); } else if (argc > 1) { - int i; - VALUE arg_str = rb_str_tmp_new(0); - rb_enc_copy(arg_str, str); - for (i = 0; i < argc; i++) { - rb_str_append(arg_str, argv[i]); - } - rb_str_update(str, 0L, 0L, arg_str); + int i; + VALUE arg_str = rb_str_tmp_new(0); + rb_enc_copy(arg_str, str); + for (i = 0; i < argc; i++) { + rb_str_append(arg_str, argv[i]); + } + rb_str_update(str, 0L, 0L, arg_str); } return str; @@ -3457,11 +3655,12 @@ rb_str_prepend_multi(int argc, VALUE *argv, VALUE str) st_index_t rb_str_hash(VALUE str) { - int e = ENCODING_GET(str); - if (e && rb_enc_str_coderange(str) == ENC_CODERANGE_7BIT) { - e = 0; + st_index_t h = rb_memhash((const void *)RSTRING_PTR(str), RSTRING_LEN(str)); + int e = RSTRING_LEN(str) ? ENCODING_GET(str) : 0; + if (e && !is_ascii_string(str)) { + h = rb_hash_end(rb_hash_uint32(h, (uint32_t)e)); } - return rb_memhash((const void *)RSTRING_PTR(str), RSTRING_LEN(str)) ^ e; + return h; } int @@ -3472,8 +3671,8 @@ rb_str_hash_cmp(VALUE str1, VALUE str2) RSTRING_GETMEM(str1, ptr1, len1); RSTRING_GETMEM(str2, ptr2, len2); return (len1 != len2 || - !rb_str_comparable(str1, str2) || - memcmp(ptr1, ptr2, len1) != 0); + !rb_str_comparable(str1, str2) || + memcmp(ptr1, ptr2, len1) != 0); } /* @@ -3509,13 +3708,13 @@ rb_str_comparable(VALUE str1, VALUE str2) rc1 = rb_enc_str_coderange(str1); rc2 = rb_enc_str_coderange(str2); if (rc1 == ENC_CODERANGE_7BIT) { - if (rc2 == ENC_CODERANGE_7BIT) return TRUE; - if (rb_enc_asciicompat(rb_enc_from_index(idx2))) - return TRUE; + if (rc2 == ENC_CODERANGE_7BIT) return TRUE; + if (rb_enc_asciicompat(rb_enc_from_index(idx2))) + return TRUE; } if (rc2 == ENC_CODERANGE_7BIT) { - if (rb_enc_asciicompat(rb_enc_from_index(idx1))) - return TRUE; + if (rb_enc_asciicompat(rb_enc_from_index(idx1))) + return TRUE; } return FALSE; } @@ -3531,16 +3730,16 @@ rb_str_cmp(VALUE str1, VALUE str2) RSTRING_GETMEM(str1, ptr1, len1); RSTRING_GETMEM(str2, ptr2, len2); if (ptr1 == ptr2 || (retval = memcmp(ptr1, ptr2, lesser(len1, len2))) == 0) { - if (len1 == len2) { - if (!rb_str_comparable(str1, str2)) { - if (ENCODING_GET(str1) > ENCODING_GET(str2)) - return 1; - return -1; - } - return 0; - } - if (len1 > len2) return 1; - return -1; + if (len1 == len2) { + if (!rb_str_comparable(str1, str2)) { + if (ENCODING_GET(str1) > ENCODING_GET(str2)) + return 1; + return -1; + } + return 0; + } + if (len1 > len2) return 1; + return -1; } if (retval > 0) return 1; return -1; @@ -3562,7 +3761,7 @@ rb_str_cmp(VALUE str1, VALUE str2) * Returns +false+ if the two strings' encodings are not compatible: * "\u{e4 f6 fc}".encode("ISO-8859-1") == ("\u{c4 d6 dc}") # => false * - * If +object+ is not an instance of \String but responds to +to_str+, then the + * If +object+ is not an instance of +String+ but responds to +to_str+, then the * two strings are compared using <code>object.==</code>. */ @@ -3571,10 +3770,10 @@ rb_str_equal(VALUE str1, VALUE str2) { if (str1 == str2) return Qtrue; if (!RB_TYPE_P(str2, T_STRING)) { - if (!rb_respond_to(str2, idTo_str)) { - return Qfalse; - } - return rb_equal(str2, str1); + if (!rb_respond_to(str2, idTo_str)) { + return Qfalse; + } + return rb_equal(str2, str1); } return rb_str_eql_internal(str1, str2); } @@ -3597,7 +3796,7 @@ rb_str_equal(VALUE str1, VALUE str2) * */ -MJIT_FUNC_EXPORTED VALUE +VALUE rb_str_eql(VALUE str1, VALUE str2) { if (str1 == str2) return Qtrue; @@ -3633,7 +3832,7 @@ rb_str_cmp_m(VALUE str1, VALUE str2) int result; VALUE s = rb_check_string_type(str2); if (NIL_P(s)) { - return rb_invcmp(str1, str2); + return rb_invcmp(str1, str2); } result = rb_str_cmp(str1, s); return INT2FIX(result); @@ -3673,7 +3872,7 @@ rb_str_casecmp(VALUE str1, VALUE str2) { VALUE s = rb_check_string_type(str2); if (NIL_P(s)) { - return Qnil; + return Qnil; } return str_casecmp(str1, s); } @@ -3687,25 +3886,25 @@ str_casecmp(VALUE str1, VALUE str2) enc = rb_enc_compatible(str1, str2); if (!enc) { - return Qnil; + return Qnil; } p1 = RSTRING_PTR(str1); p1end = RSTRING_END(str1); p2 = RSTRING_PTR(str2); p2end = RSTRING_END(str2); if (single_byte_optimizable(str1) && single_byte_optimizable(str2)) { - while (p1 < p1end && p2 < p2end) { - if (*p1 != *p2) { + while (p1 < p1end && p2 < p2end) { + if (*p1 != *p2) { unsigned int c1 = TOLOWER(*p1 & 0xff); unsigned int c2 = TOLOWER(*p2 & 0xff); if (c1 != c2) return INT2FIX(c1 < c2 ? -1 : 1); - } - p1++; - p2++; - } + } + p1++; + p2++; + } } else { - while (p1 < p1end && p2 < p2end) { + while (p1 < p1end && p2 < p2end) { int l1, c1 = rb_enc_ascget(p1, p1end, &l1, enc); int l2, c2 = rb_enc_ascget(p2, p2end, &l2, enc); @@ -3726,9 +3925,9 @@ str_casecmp(VALUE str1, VALUE str2) if (l1 != l2) return INT2FIX(l1 < l2 ? -1 : 1); } - p1 += l1; - p2 += l2; - } + p1 += l1; + p2 += l2; + } } if (RSTRING_LEN(str1) == RSTRING_LEN(str2)) return INT2FIX(0); if (RSTRING_LEN(str1) > RSTRING_LEN(str2)) return INT2FIX(1); @@ -3763,7 +3962,7 @@ rb_str_casecmp_p(VALUE str1, VALUE str2) { VALUE s = rb_check_string_type(str2); if (NIL_P(s)) { - return Qnil; + return Qnil; } return str_casecmp_p(str1, s); } @@ -3777,7 +3976,7 @@ str_casecmp_p(VALUE str1, VALUE str2) enc = rb_enc_compatible(str1, str2); if (!enc) { - return Qnil; + return Qnil; } folded_str1 = rb_str_downcase(1, &fold_opt, str1); @@ -3788,26 +3987,28 @@ str_casecmp_p(VALUE str1, VALUE str2) static long strseq_core(const char *str_ptr, const char *str_ptr_end, long str_len, - const char *sub_ptr, long sub_len, long offset, rb_encoding *enc) + const char *sub_ptr, long sub_len, long offset, rb_encoding *enc) { const char *search_start = str_ptr; long pos, search_len = str_len - offset; for (;;) { - const char *t; - pos = rb_memsearch(sub_ptr, sub_len, search_start, search_len, enc); - if (pos < 0) return pos; - t = rb_enc_right_char_head(search_start, search_start+pos, str_ptr_end, enc); - if (t == search_start + pos) break; - search_len -= t - search_start; - if (search_len <= 0) return -1; - offset += t - search_start; - search_start = t; + const char *t; + pos = rb_memsearch(sub_ptr, sub_len, search_start, search_len, enc); + if (pos < 0) return pos; + t = rb_enc_right_char_head(search_start, search_start+pos, str_ptr_end, enc); + if (t == search_start + pos) break; + search_len -= t - search_start; + if (search_len <= 0) return -1; + offset += t - search_start; + search_start = t; } return pos + offset; } +/* found index in byte */ #define rb_str_index(str, sub, offset) rb_strseq_index(str, sub, offset, 0) +#define rb_str_byteindex(str, sub, offset) rb_strseq_index(str, sub, offset, 1) static long rb_strseq_index(VALUE str, VALUE sub, long offset, int in_byte) @@ -3828,17 +4029,17 @@ rb_strseq_index(VALUE str, VALUE sub, long offset, int in_byte) if (str_len < sub_len) return -1; if (offset != 0) { - long str_len_char, sub_len_char; + long str_len_char, sub_len_char; int single_byte = single_byte_optimizable(str); - str_len_char = (in_byte || single_byte) ? str_len : str_strlen(str, enc); - sub_len_char = in_byte ? sub_len : str_strlen(sub, enc); - if (offset < 0) { - offset += str_len_char; - if (offset < 0) return -1; - } - if (str_len_char - offset < sub_len_char) return -1; - if (!in_byte) offset = str_offset(str_ptr, str_ptr_end, offset, enc, single_byte); - str_ptr += offset; + str_len_char = (in_byte || single_byte) ? str_len : str_strlen(str, enc); + sub_len_char = in_byte ? sub_len : str_strlen(sub, enc); + if (offset < 0) { + offset += str_len_char; + if (offset < 0) return -1; + } + if (str_len_char - offset < sub_len_char) return -1; + if (!in_byte) offset = str_offset(str_ptr, str_ptr_end, offset, enc, single_byte); + str_ptr += offset; } if (sub_len == 0) return offset; @@ -3861,34 +4062,28 @@ rb_str_index_m(int argc, VALUE *argv, VALUE str) { VALUE sub; VALUE initpos; + rb_encoding *enc = STR_ENC_GET(str); long pos; if (rb_scan_args(argc, argv, "11", &sub, &initpos) == 2) { - pos = NUM2LONG(initpos); + long slen = str_strlen(str, enc); /* str's enc */ + pos = NUM2LONG(initpos); + if (pos < 0 ? (pos += slen) < 0 : pos > slen) { + if (RB_TYPE_P(sub, T_REGEXP)) { + rb_backref_set(Qnil); + } + return Qnil; + } } else { - pos = 0; - } - if (pos < 0) { - pos += str_strlen(str, NULL); - if (pos < 0) { - if (RB_TYPE_P(sub, T_REGEXP)) { - rb_backref_set(Qnil); - } - return Qnil; - } + pos = 0; } if (RB_TYPE_P(sub, T_REGEXP)) { - if (pos > str_strlen(str, NULL)) - return Qnil; - pos = str_offset(RSTRING_PTR(str), RSTRING_END(str), pos, - rb_enc_check(str, sub), single_byte_optimizable(str)); + pos = str_offset(RSTRING_PTR(str), RSTRING_END(str), pos, + enc, single_byte_optimizable(str)); - if (rb_reg_search(sub, str, pos, 0) < 0) { - return Qnil; - } - else { + if (rb_reg_search(sub, str, pos, 0) >= 0) { VALUE match = rb_backref_get(); struct re_registers *regs = RMATCH_REGS(match); pos = rb_str_sublen(str, BEG(0)); @@ -3897,26 +4092,29 @@ rb_str_index_m(int argc, VALUE *argv, VALUE str) } else { StringValue(sub); - pos = rb_str_index(str, sub, pos); - pos = rb_str_sublen(str, pos); + pos = rb_str_index(str, sub, pos); + if (pos >= 0) { + pos = rb_str_sublen(str, pos); + return LONG2NUM(pos); + } } - - if (pos == -1) return Qnil; - return LONG2NUM(pos); + return Qnil; } -/* whether given pos is valid character boundary or not +/* Ensure that the given pos is a valid character boundary. * Note that in this function, "character" means a code point * (Unicode scalar value), not a grapheme cluster. */ -static bool -str_check_byte_pos(VALUE str, long pos) +static void +str_ensure_byte_pos(VALUE str, long pos) { const char *s = RSTRING_PTR(str); const char *e = RSTRING_END(str); const char *p = s + pos; - const char *pp = rb_enc_left_char_head(s, p, e, rb_enc_get(str)); - return p == pp; + if (!at_char_boundary(s, p, e, rb_enc_get(str))) { + rb_raise(rb_eIndexError, + "offset %ld does not land on character boundary", pos); + } } /* @@ -3924,7 +4122,7 @@ str_check_byte_pos(VALUE str, long pos) * byteindex(substring, offset = 0) -> integer or nil * byteindex(regexp, offset = 0) -> integer or nil * - * Returns the \Integer byte-based index of the first occurrence of the given +substring+, + * Returns the Integer byte-based index of the first occurrence of the given +substring+, * or +nil+ if none found: * * 'foo'.byteindex('f') # => 0 @@ -3932,7 +4130,7 @@ str_check_byte_pos(VALUE str, long pos) * 'foo'.byteindex('oo') # => 1 * 'foo'.byteindex('ooo') # => nil * - * Returns the \Integer byte-based index of the first match for the given \Regexp +regexp+, + * Returns the Integer byte-based index of the first match for the given Regexp +regexp+, * or +nil+ if none found: * * 'foo'.byteindex(/f/) # => 0 @@ -3940,7 +4138,7 @@ str_check_byte_pos(VALUE str, long pos) * 'foo'.byteindex(/oo/) # => 1 * 'foo'.byteindex(/ooo/) # => nil * - * \Integer argument +offset+, if given, specifies the byte-based position in the + * Integer argument +offset+, if given, specifies the byte-based position in the * string to begin the search: * * 'foo'.byteindex('o', 1) # => 1 @@ -3968,33 +4166,23 @@ rb_str_byteindex_m(int argc, VALUE *argv, VALUE str) long pos; if (rb_scan_args(argc, argv, "11", &sub, &initpos) == 2) { + long slen = RSTRING_LEN(str); pos = NUM2LONG(initpos); - } - else { - pos = 0; - } - if (pos < 0) { - pos += RSTRING_LEN(str); - if (pos < 0) { + if (pos < 0 ? (pos += slen) < 0 : pos > slen) { if (RB_TYPE_P(sub, T_REGEXP)) { rb_backref_set(Qnil); } return Qnil; } } - - if (!str_check_byte_pos(str, pos)) { - rb_raise(rb_eIndexError, - "offset %ld does not land on character boundary", pos); + else { + pos = 0; } + str_ensure_byte_pos(str, pos); + if (RB_TYPE_P(sub, T_REGEXP)) { - if (pos > RSTRING_LEN(str)) - return Qnil; - if (rb_reg_search(sub, str, pos, 0) < 0) { - return Qnil; - } - else { + if (rb_reg_search(sub, str, pos, 0) >= 0) { VALUE match = rb_backref_get(); struct re_registers *regs = RMATCH_REGS(match); pos = BEG(0); @@ -4003,11 +4191,10 @@ rb_str_byteindex_m(int argc, VALUE *argv, VALUE str) } else { StringValue(sub); - pos = rb_strseq_index(str, sub, pos, 1); + pos = rb_str_byteindex(str, sub, pos); + if (pos >= 0) return LONG2NUM(pos); } - - if (pos == -1) return Qnil; - return LONG2NUM(pos); + return Qnil; } #ifdef HAVE_MEMRCHR @@ -4028,16 +4215,16 @@ str_rindex(VALUE str, VALUE sub, const char *s, rb_encoding *enc) searchlen = s - sbeg + 1; do { - hit = memrchr(sbeg, c, searchlen); - if (!hit) break; - adjusted = rb_enc_left_char_head(sbeg, hit, e, enc); - if (hit != adjusted) { - searchlen = adjusted - sbeg; - continue; - } - if (memcmp(hit, t, slen) == 0) - return hit - sbeg; - searchlen = adjusted - sbeg; + hit = memrchr(sbeg, c, searchlen); + if (!hit) break; + adjusted = rb_enc_left_char_head(sbeg, hit, e, enc); + if (hit != adjusted) { + searchlen = adjusted - sbeg; + continue; + } + if (memcmp(hit, t, slen) == 0) + return hit - sbeg; + searchlen = adjusted - sbeg; } while (searchlen > 0); return -1; @@ -4055,17 +4242,18 @@ str_rindex(VALUE str, VALUE sub, const char *s, rb_encoding *enc) slen = RSTRING_LEN(sub); while (s) { - if (memcmp(s, t, slen) == 0) { - return s - sbeg; - } + if (memcmp(s, t, slen) == 0) { + return s - sbeg; + } if (s <= sbeg) break; - s = rb_enc_prev_char(sbeg, s, e, enc); + s = rb_enc_prev_char(sbeg, s, e, enc); } return -1; } #endif +/* found index in byte */ static long rb_str_rindex(VALUE str, VALUE sub, long pos) { @@ -4088,14 +4276,14 @@ rb_str_rindex(VALUE str, VALUE sub, long pos) sbeg = RSTRING_PTR(str); if (pos == 0) { - if (memcmp(sbeg, RSTRING_PTR(sub), RSTRING_LEN(sub)) == 0) - return 0; - else - return -1; + if (memcmp(sbeg, RSTRING_PTR(sub), RSTRING_LEN(sub)) == 0) + return 0; + else + return -1; } s = str_nth(sbeg, RSTRING_END(str), pos, enc, singlebyte); - return rb_str_sublen(str, str_rindex(str, sub, s, enc)); + return str_rindex(str, sub, s, enc); } /* @@ -4103,7 +4291,7 @@ rb_str_rindex(VALUE str, VALUE sub, long pos) * rindex(substring, offset = self.length) -> integer or nil * rindex(regexp, offset = self.length) -> integer or nil * - * Returns the \Integer index of the _last_ occurrence of the given +substring+, + * Returns the Integer index of the _last_ occurrence of the given +substring+, * or +nil+ if none found: * * 'foo'.rindex('f') # => 0 @@ -4111,7 +4299,7 @@ rb_str_rindex(VALUE str, VALUE sub, long pos) * 'foo'.rindex('oo') # => 1 * 'foo'.rindex('ooo') # => nil * - * Returns the \Integer index of the _last_ match for the given \Regexp +regexp+, + * Returns the Integer index of the _last_ match for the given Regexp +regexp+, * or +nil+ if none found: * * 'foo'.rindex(/f/) # => 0 @@ -4136,7 +4324,7 @@ rb_str_rindex(VALUE str, VALUE sub, long pos) * 'foo'.index(/o+(?!.*o)/) # => 1 * $~ #=> #<MatchData "oo"> * - * \Integer argument +offset+, if given and non-negative, specifies the maximum starting position in the + * Integer argument +offset+, if given and non-negative, specifies the maximum starting position in the * string to _end_ the search: * * 'foo'.rindex('o', 0) # => nil @@ -4144,7 +4332,7 @@ rb_str_rindex(VALUE str, VALUE sub, long pos) * 'foo'.rindex('o', 2) # => 2 * 'foo'.rindex('o', 3) # => 2 * - * If +offset+ is a negative \Integer, the maximum starting position in the + * If +offset+ is a negative Integer, the maximum starting position in the * string to _end_ the search is the sum of the string's length and +offset+: * * 'foo'.rindex('o', -1) # => 2 @@ -4159,33 +4347,30 @@ static VALUE rb_str_rindex_m(int argc, VALUE *argv, VALUE str) { VALUE sub; - VALUE vpos; + VALUE initpos; rb_encoding *enc = STR_ENC_GET(str); long pos, len = str_strlen(str, enc); /* str's enc */ - if (rb_scan_args(argc, argv, "11", &sub, &vpos) == 2) { - pos = NUM2LONG(vpos); - if (pos < 0) { - pos += len; - if (pos < 0) { - if (RB_TYPE_P(sub, T_REGEXP)) { - rb_backref_set(Qnil); - } - return Qnil; - } - } - if (pos > len) pos = len; + if (rb_scan_args(argc, argv, "11", &sub, &initpos) == 2) { + pos = NUM2LONG(initpos); + if (pos < 0 && (pos += len) < 0) { + if (RB_TYPE_P(sub, T_REGEXP)) { + rb_backref_set(Qnil); + } + return Qnil; + } + if (pos > len) pos = len; } else { - pos = len; + pos = len; } if (RB_TYPE_P(sub, T_REGEXP)) { - /* enc = rb_get_check(str, sub); */ - pos = str_offset(RSTRING_PTR(str), RSTRING_END(str), pos, - enc, single_byte_optimizable(str)); + /* enc = rb_enc_check(str, sub); */ + pos = str_offset(RSTRING_PTR(str), RSTRING_END(str), pos, + enc, single_byte_optimizable(str)); - if (rb_reg_search(sub, str, pos, 1) >= 0) { + if (rb_reg_search(sub, str, pos, 1) >= 0) { VALUE match = rb_backref_get(); struct re_registers *regs = RMATCH_REGS(match); pos = rb_str_sublen(str, BEG(0)); @@ -4194,8 +4379,11 @@ rb_str_rindex_m(int argc, VALUE *argv, VALUE str) } else { StringValue(sub); - pos = rb_str_rindex(str, sub, pos); - if (pos >= 0) return LONG2NUM(pos); + pos = rb_str_rindex(str, sub, pos); + if (pos >= 0) { + pos = rb_str_sublen(str, pos); + return LONG2NUM(pos); + } } return Qnil; } @@ -4236,7 +4424,7 @@ rb_str_byterindex(VALUE str, VALUE sub, long pos) * byterindex(substring, offset = self.bytesize) -> integer or nil * byterindex(regexp, offset = self.bytesize) -> integer or nil * - * Returns the \Integer byte-based index of the _last_ occurrence of the given +substring+, + * Returns the Integer byte-based index of the _last_ occurrence of the given +substring+, * or +nil+ if none found: * * 'foo'.byterindex('f') # => 0 @@ -4244,7 +4432,7 @@ rb_str_byterindex(VALUE str, VALUE sub, long pos) * 'foo'.byterindex('oo') # => 1 * 'foo'.byterindex('ooo') # => nil * - * Returns the \Integer byte-based index of the _last_ match for the given \Regexp +regexp+, + * Returns the Integer byte-based index of the _last_ match for the given Regexp +regexp+, * or +nil+ if none found: * * 'foo'.byterindex(/f/) # => 0 @@ -4269,7 +4457,7 @@ rb_str_byterindex(VALUE str, VALUE sub, long pos) * 'foo'.byteindex(/o+(?!.*o)/) # => 1 * $~ #=> #<MatchData "oo"> * - * \Integer argument +offset+, if given and non-negative, specifies the maximum starting byte-based position in the + * Integer argument +offset+, if given and non-negative, specifies the maximum starting byte-based position in the * string to _end_ the search: * * 'foo'.byterindex('o', 0) # => nil @@ -4277,7 +4465,7 @@ rb_str_byterindex(VALUE str, VALUE sub, long pos) * 'foo'.byterindex('o', 2) # => 2 * 'foo'.byterindex('o', 3) # => 2 * - * If +offset+ is a negative \Integer, the maximum starting position in the + * If +offset+ is a negative Integer, the maximum starting position in the * string to _end_ the search is the sum of the string's length and +offset+: * * 'foo'.byterindex('o', -1) # => 2 @@ -4295,19 +4483,16 @@ static VALUE rb_str_byterindex_m(int argc, VALUE *argv, VALUE str) { VALUE sub; - VALUE vpos; + VALUE initpos; long pos, len = RSTRING_LEN(str); - if (rb_scan_args(argc, argv, "11", &sub, &vpos) == 2) { - pos = NUM2LONG(vpos); - if (pos < 0) { - pos += len; - if (pos < 0) { - if (RB_TYPE_P(sub, T_REGEXP)) { - rb_backref_set(Qnil); - } - return Qnil; + if (rb_scan_args(argc, argv, "11", &sub, &initpos) == 2) { + pos = NUM2LONG(initpos); + if (pos < 0 && (pos += len) < 0) { + if (RB_TYPE_P(sub, T_REGEXP)) { + rb_backref_set(Qnil); } + return Qnil; } if (pos > len) pos = len; } @@ -4315,10 +4500,7 @@ rb_str_byterindex_m(int argc, VALUE *argv, VALUE str) pos = len; } - if (!str_check_byte_pos(str, pos)) { - rb_raise(rb_eIndexError, - "offset %ld does not land on character boundary", pos); - } + str_ensure_byte_pos(str, pos); if (RB_TYPE_P(sub, T_REGEXP)) { if (rb_reg_search(sub, str, pos, 1) >= 0) { @@ -4341,16 +4523,16 @@ rb_str_byterindex_m(int argc, VALUE *argv, VALUE str) * string =~ regexp -> integer or nil * string =~ object -> integer or nil * - * Returns the \Integer index of the first substring that matches + * Returns the Integer index of the first substring that matches * the given +regexp+, or +nil+ if no match found: * * 'foo' =~ /f/ # => 0 * 'foo' =~ /o/ # => 1 * 'foo' =~ /x/ # => nil * - * Note: also updates Regexp@Special+global+variables. + * Note: also updates Regexp@Global+Variables. * - * If the given +object+ is not a \Regexp, returns the value + * If the given +object+ is not a Regexp, returns the value * returned by <tt>object =~ self</tt>. * * Note that <tt>string =~ regexp</tt> is different from <tt>regexp =~ string</tt> @@ -4369,13 +4551,13 @@ rb_str_match(VALUE x, VALUE y) { switch (OBJ_BUILTIN_TYPE(y)) { case T_STRING: - rb_raise(rb_eTypeError, "type mismatch: String given"); + rb_raise(rb_eTypeError, "type mismatch: String given"); case T_REGEXP: - return rb_reg_match(y, x); + return rb_reg_match(y, x); default: - return rb_funcall(y, idEqTilde, 1, x); + return rb_funcall(y, idEqTilde, 1, x); } } @@ -4388,13 +4570,13 @@ static VALUE get_pat(VALUE); * match(pattern, offset = 0) -> matchdata or nil * match(pattern, offset = 0) {|matchdata| ... } -> object * - * Returns a \Matchdata object (or +nil+) based on +self+ and the given +pattern+. + * Returns a MatchData object (or +nil+) based on +self+ and the given +pattern+. * - * Note: also updates Regexp@Special+global+variables. + * Note: also updates Regexp@Global+Variables. * - * - Computes +regexp+ by converting +pattern+ (if not already a \Regexp). + * - Computes +regexp+ by converting +pattern+ (if not already a Regexp). * regexp = Regexp.new(pattern) - * - Computes +matchdata+, which will be either a \MatchData object or +nil+ + * - Computes +matchdata+, which will be either a MatchData object or +nil+ * (see Regexp#match): * matchdata = <tt>regexp.match(self) * @@ -4404,7 +4586,7 @@ static VALUE get_pat(VALUE); * 'foo'.match('o') # => #<MatchData "o"> * 'foo'.match('x') # => nil * - * If \Integer argument +offset+ is given, the search begins at index +offset+: + * If Integer argument +offset+ is given, the search begins at index +offset+: * * 'foo'.match('f', 1) # => nil * 'foo'.match('o', 1) # => #<MatchData "o"> @@ -4423,12 +4605,12 @@ rb_str_match_m(int argc, VALUE *argv, VALUE str) { VALUE re, result; if (argc < 1) - rb_check_arity(argc, 1, 2); + rb_check_arity(argc, 1, 2); re = argv[0]; argv[0] = str; result = rb_funcallv(get_pat(re), rb_intern("match"), argc, argv); if (!NIL_P(result) && rb_block_given_p()) { - return rb_yield(result); + return rb_yield(result); } return result; } @@ -4439,19 +4621,19 @@ rb_str_match_m(int argc, VALUE *argv, VALUE str) * * Returns +true+ or +false+ based on whether a match is found for +self+ and +pattern+. * - * Note: does not update Regexp@Special+global+variables. + * Note: does not update Regexp@Global+Variables. * - * Computes +regexp+ by converting +pattern+ (if not already a \Regexp). + * Computes +regexp+ by converting +pattern+ (if not already a Regexp). * regexp = Regexp.new(pattern) * - * Returns +true+ if <tt>self+.match(regexp)</tt> returns a \Matchdata object, + * Returns +true+ if <tt>self+.match(regexp)</tt> returns a MatchData object, * +false+ otherwise: * * 'foo'.match?(/o/) # => true * 'foo'.match?('o') # => true * 'foo'.match?(/x/) # => false * - * If \Integer argument +offset+ is given, the search begins at index +offset+: + * If Integer argument +offset+ is given, the search begins at index +offset+: * 'foo'.match?('f', 1) # => false * 'foo'.match?('o', 1) # => true * @@ -4479,21 +4661,21 @@ enc_succ_char(char *p, long len, rb_encoding *enc) int l; if (rb_enc_mbminlen(enc) > 1) { - /* wchar, trivial case */ - int r = rb_enc_precise_mbclen(p, p + len, enc), c; - if (!MBCLEN_CHARFOUND_P(r)) { - return NEIGHBOR_NOT_CHAR; - } - c = rb_enc_mbc_to_codepoint(p, p + len, enc) + 1; - l = rb_enc_code_to_mbclen(c, enc); - if (!l) return NEIGHBOR_NOT_CHAR; - if (l != len) return NEIGHBOR_WRAPPED; - rb_enc_mbcput(c, p, enc); - r = rb_enc_precise_mbclen(p, p + len, enc); - if (!MBCLEN_CHARFOUND_P(r)) { - return NEIGHBOR_NOT_CHAR; - } - return NEIGHBOR_FOUND; + /* wchar, trivial case */ + int r = rb_enc_precise_mbclen(p, p + len, enc), c; + if (!MBCLEN_CHARFOUND_P(r)) { + return NEIGHBOR_NOT_CHAR; + } + c = rb_enc_mbc_to_codepoint(p, p + len, enc) + 1; + l = rb_enc_code_to_mbclen(c, enc); + if (!l) return NEIGHBOR_NOT_CHAR; + if (l != len) return NEIGHBOR_WRAPPED; + rb_enc_mbcput(c, p, enc); + r = rb_enc_precise_mbclen(p, p + len, enc); + if (!MBCLEN_CHARFOUND_P(r)) { + return NEIGHBOR_NOT_CHAR; + } + return NEIGHBOR_FOUND; } while (1) { for (i = len-1; 0 <= i && (unsigned char)p[i] == 0xff; i--) @@ -4530,23 +4712,23 @@ enc_pred_char(char *p, long len, rb_encoding *enc) long i; int l; if (rb_enc_mbminlen(enc) > 1) { - /* wchar, trivial case */ - int r = rb_enc_precise_mbclen(p, p + len, enc), c; - if (!MBCLEN_CHARFOUND_P(r)) { - return NEIGHBOR_NOT_CHAR; - } - c = rb_enc_mbc_to_codepoint(p, p + len, enc); - if (!c) return NEIGHBOR_NOT_CHAR; - --c; - l = rb_enc_code_to_mbclen(c, enc); - if (!l) return NEIGHBOR_NOT_CHAR; - if (l != len) return NEIGHBOR_WRAPPED; - rb_enc_mbcput(c, p, enc); - r = rb_enc_precise_mbclen(p, p + len, enc); - if (!MBCLEN_CHARFOUND_P(r)) { - return NEIGHBOR_NOT_CHAR; - } - return NEIGHBOR_FOUND; + /* wchar, trivial case */ + int r = rb_enc_precise_mbclen(p, p + len, enc), c; + if (!MBCLEN_CHARFOUND_P(r)) { + return NEIGHBOR_NOT_CHAR; + } + c = rb_enc_mbc_to_codepoint(p, p + len, enc); + if (!c) return NEIGHBOR_NOT_CHAR; + --c; + l = rb_enc_code_to_mbclen(c, enc); + if (!l) return NEIGHBOR_NOT_CHAR; + if (l != len) return NEIGHBOR_WRAPPED; + rb_enc_mbcput(c, p, enc); + r = rb_enc_precise_mbclen(p, p + len, enc); + if (!MBCLEN_CHARFOUND_P(r)) { + return NEIGHBOR_NOT_CHAR; + } + return NEIGHBOR_FOUND; } while (1) { for (i = len-1; 0 <= i && (unsigned char)p[i] == 0; i--) @@ -4609,12 +4791,12 @@ enc_succ_alnum_char(char *p, long len, rb_encoding *enc, char *carry) MEMCPY(save, p, char, len); for (try = 0; try <= max_gaps; ++try) { - ret = enc_succ_char(p, len, enc); - if (ret == NEIGHBOR_FOUND) { - c = rb_enc_mbc_to_codepoint(p, p+len, enc); - if (rb_enc_isctype(c, ctype, enc)) - return NEIGHBOR_FOUND; - } + ret = enc_succ_char(p, len, enc); + if (ret == NEIGHBOR_FOUND) { + c = rb_enc_mbc_to_codepoint(p, p+len, enc); + if (rb_enc_isctype(c, ctype, enc)) + return NEIGHBOR_FOUND; + } } MEMCPY(p, save, char, len); range = 1; @@ -4703,11 +4885,10 @@ static VALUE str_succ(VALUE str); * s = '99zz99zz' * s.succ # => "100aa00aa" * - * The successor to an empty \String is a new empty \String: + * The successor to an empty +String+ is a new empty +String+: * * ''.succ # => "" * - * String#next is an alias for String#succ. */ VALUE @@ -4738,50 +4919,50 @@ str_succ(VALUE str) s = e = sbeg + slen; while ((s = rb_enc_prev_char(sbeg, s, e, enc)) != 0) { - if (neighbor == NEIGHBOR_NOT_CHAR && last_alnum) { - if (ISALPHA(*last_alnum) ? ISDIGIT(*s) : - ISDIGIT(*last_alnum) ? ISALPHA(*s) : 0) { - break; - } - } - l = rb_enc_precise_mbclen(s, e, enc); - if (!ONIGENC_MBCLEN_CHARFOUND_P(l)) continue; - l = ONIGENC_MBCLEN_CHARFOUND_LEN(l); + if (neighbor == NEIGHBOR_NOT_CHAR && last_alnum) { + if (ISALPHA(*last_alnum) ? ISDIGIT(*s) : + ISDIGIT(*last_alnum) ? ISALPHA(*s) : 0) { + break; + } + } + l = rb_enc_precise_mbclen(s, e, enc); + if (!ONIGENC_MBCLEN_CHARFOUND_P(l)) continue; + l = ONIGENC_MBCLEN_CHARFOUND_LEN(l); neighbor = enc_succ_alnum_char(s, l, enc, carry); switch (neighbor) { - case NEIGHBOR_NOT_CHAR: - continue; - case NEIGHBOR_FOUND: - return str; - case NEIGHBOR_WRAPPED: - last_alnum = s; - break; - } + case NEIGHBOR_NOT_CHAR: + continue; + case NEIGHBOR_FOUND: + return str; + case NEIGHBOR_WRAPPED: + last_alnum = s; + break; + } found_alnum = 1; carry_pos = s - sbeg; carry_len = l; } if (!found_alnum) { /* str contains no alnum */ - s = e; - while ((s = rb_enc_prev_char(sbeg, s, e, enc)) != 0) { + s = e; + while ((s = rb_enc_prev_char(sbeg, s, e, enc)) != 0) { enum neighbor_char neighbor; - char tmp[ONIGENC_CODE_TO_MBC_MAXLEN]; - l = rb_enc_precise_mbclen(s, e, enc); - if (!ONIGENC_MBCLEN_CHARFOUND_P(l)) continue; - l = ONIGENC_MBCLEN_CHARFOUND_LEN(l); - MEMCPY(tmp, s, char, l); - neighbor = enc_succ_char(tmp, l, enc); - switch (neighbor) { - case NEIGHBOR_FOUND: - MEMCPY(s, tmp, char, l); + char tmp[ONIGENC_CODE_TO_MBC_MAXLEN]; + l = rb_enc_precise_mbclen(s, e, enc); + if (!ONIGENC_MBCLEN_CHARFOUND_P(l)) continue; + l = ONIGENC_MBCLEN_CHARFOUND_LEN(l); + MEMCPY(tmp, s, char, l); + neighbor = enc_succ_char(tmp, l, enc); + switch (neighbor) { + case NEIGHBOR_FOUND: + MEMCPY(s, tmp, char, l); return str; - break; - case NEIGHBOR_WRAPPED: - MEMCPY(s, tmp, char, l); - break; - case NEIGHBOR_NOT_CHAR: - break; - } + break; + case NEIGHBOR_WRAPPED: + MEMCPY(s, tmp, char, l); + break; + case NEIGHBOR_NOT_CHAR: + break; + } if (rb_enc_precise_mbclen(s, s+l, enc) != l) { /* wrapped to \0...\0. search next valid char. */ enc_succ_char(s, l, enc); @@ -4791,8 +4972,8 @@ str_succ(VALUE str) carry_len = l; } carry_pos = s - sbeg; - } - ENC_CODERANGE_SET(str, ENC_CODERANGE_UNKNOWN); + } + ENC_CODERANGE_SET(str, ENC_CODERANGE_UNKNOWN); } RESIZE_CAPA(str, slen + carry_len); sbeg = RSTRING_PTR(str); @@ -4812,8 +4993,6 @@ str_succ(VALUE str) * succ! -> self * * Equivalent to String#succ, but modifies +self+ in place; returns +self+. - * - * String#next! is an alias for String#succ!. */ static VALUE @@ -4828,8 +5007,8 @@ static int all_digits_p(const char *s, long len) { while (len-- > 0) { - if (!ISDIGIT(*s)) return 0; - s++; + if (!ISDIGIT(*s)) return 0; + s++; } return 1; } @@ -4846,7 +5025,7 @@ str_upto_i(VALUE str, VALUE arg) * upto(other_string, exclusive = false) {|string| ... } -> self * upto(other_string, exclusive = false) -> new_enumerator * - * With a block given, calls the block with each \String value + * With a block given, calls the block with each +String+ value * returned by successive calls to String#succ; * the first value is +self+, the next is <tt>self.succ</tt>, and so on; * the sequence terminates when value +other_string+ is reached; @@ -4870,7 +5049,7 @@ str_upto_i(VALUE str, VALUE arg) * '25'.upto('5') {|s| fail s } * 'aa'.upto('a') {|s| fail s } * - * With no block given, returns a new \Enumerator: + * With no block given, returns a new Enumerator: * * 'a8'.upto('b6') # => #<Enumerator: "a8":upto("b6")> * @@ -4900,51 +5079,51 @@ rb_str_upto_each(VALUE beg, VALUE end, int excl, int (*each)(VALUE, VALUE), VALU ascii = (is_ascii_string(beg) && is_ascii_string(end)); /* single character */ if (RSTRING_LEN(beg) == 1 && RSTRING_LEN(end) == 1 && ascii) { - char c = RSTRING_PTR(beg)[0]; - char e = RSTRING_PTR(end)[0]; - - if (c > e || (excl && c == e)) return beg; - for (;;) { - if ((*each)(rb_enc_str_new(&c, 1, enc), arg)) break; - if (!excl && c == e) break; - c++; - if (excl && c == e) break; - } - return beg; + char c = RSTRING_PTR(beg)[0]; + char e = RSTRING_PTR(end)[0]; + + if (c > e || (excl && c == e)) return beg; + for (;;) { + if ((*each)(rb_enc_str_new(&c, 1, enc), arg)) break; + if (!excl && c == e) break; + c++; + if (excl && c == e) break; + } + return beg; } /* both edges are all digits */ if (ascii && ISDIGIT(RSTRING_PTR(beg)[0]) && ISDIGIT(RSTRING_PTR(end)[0]) && - all_digits_p(RSTRING_PTR(beg), RSTRING_LEN(beg)) && - all_digits_p(RSTRING_PTR(end), RSTRING_LEN(end))) { - VALUE b, e; - int width; - - width = RSTRING_LENINT(beg); - b = rb_str_to_inum(beg, 10, FALSE); - e = rb_str_to_inum(end, 10, FALSE); - if (FIXNUM_P(b) && FIXNUM_P(e)) { - long bi = FIX2LONG(b); - long ei = FIX2LONG(e); - rb_encoding *usascii = rb_usascii_encoding(); - - while (bi <= ei) { - if (excl && bi == ei) break; - if ((*each)(rb_enc_sprintf(usascii, "%.*ld", width, bi), arg)) break; - bi++; - } - } - else { - ID op = excl ? '<' : idLE; - VALUE args[2], fmt = rb_fstring_lit("%.*d"); - - args[0] = INT2FIX(width); - while (rb_funcall(b, op, 1, e)) { - args[1] = b; - if ((*each)(rb_str_format(numberof(args), args, fmt), arg)) break; - b = rb_funcallv(b, succ, 0, 0); - } - } - return beg; + all_digits_p(RSTRING_PTR(beg), RSTRING_LEN(beg)) && + all_digits_p(RSTRING_PTR(end), RSTRING_LEN(end))) { + VALUE b, e; + int width; + + width = RSTRING_LENINT(beg); + b = rb_str_to_inum(beg, 10, FALSE); + e = rb_str_to_inum(end, 10, FALSE); + if (FIXNUM_P(b) && FIXNUM_P(e)) { + long bi = FIX2LONG(b); + long ei = FIX2LONG(e); + rb_encoding *usascii = rb_usascii_encoding(); + + while (bi <= ei) { + if (excl && bi == ei) break; + if ((*each)(rb_enc_sprintf(usascii, "%.*ld", width, bi), arg)) break; + bi++; + } + } + else { + ID op = excl ? '<' : idLE; + VALUE args[2], fmt = rb_fstring_lit("%.*d"); + + args[0] = INT2FIX(width); + while (rb_funcall(b, op, 1, e)) { + args[1] = b; + if ((*each)(rb_str_format(numberof(args), args, fmt), arg)) break; + b = rb_funcallv(b, succ, 0, 0); + } + } + return beg; } /* normal case */ n = rb_str_cmp(beg, end); @@ -4953,16 +5132,16 @@ rb_str_upto_each(VALUE beg, VALUE end, int excl, int (*each)(VALUE, VALUE), VALU after_end = rb_funcallv(end, succ, 0, 0); current = str_duplicate(rb_cString, beg); while (!rb_str_equal(current, after_end)) { - VALUE next = Qnil; - if (excl || !rb_str_equal(current, end)) - next = rb_funcallv(current, succ, 0, 0); - if ((*each)(current, arg)) break; - if (NIL_P(next)) break; - current = next; - StringValue(current); - if (excl && rb_str_equal(current, end)) break; - if (RSTRING_LEN(current) > RSTRING_LEN(end) || RSTRING_LEN(current) == 0) - break; + VALUE next = Qnil; + if (excl || !rb_str_equal(current, end)) + next = rb_funcallv(current, succ, 0, 0); + if ((*each)(current, arg)) break; + if (NIL_P(next)) break; + current = next; + StringValue(current); + if (excl && rb_str_equal(current, end)) break; + if (RSTRING_LEN(current) > RSTRING_LEN(end) || RSTRING_LEN(current) == 0) + break; } return beg; @@ -4977,36 +5156,36 @@ rb_str_upto_endless_each(VALUE beg, int (*each)(VALUE, VALUE), VALUE arg) CONST_ID(succ, "succ"); /* both edges are all digits */ if (is_ascii_string(beg) && ISDIGIT(RSTRING_PTR(beg)[0]) && - all_digits_p(RSTRING_PTR(beg), RSTRING_LEN(beg))) { - VALUE b, args[2], fmt = rb_fstring_lit("%.*d"); - int width = RSTRING_LENINT(beg); - b = rb_str_to_inum(beg, 10, FALSE); - if (FIXNUM_P(b)) { - long bi = FIX2LONG(b); - rb_encoding *usascii = rb_usascii_encoding(); - - while (FIXABLE(bi)) { - if ((*each)(rb_enc_sprintf(usascii, "%.*ld", width, bi), arg)) break; - bi++; - } - b = LONG2NUM(bi); - } - args[0] = INT2FIX(width); - while (1) { - args[1] = b; - if ((*each)(rb_str_format(numberof(args), args, fmt), arg)) break; - b = rb_funcallv(b, succ, 0, 0); - } + all_digits_p(RSTRING_PTR(beg), RSTRING_LEN(beg))) { + VALUE b, args[2], fmt = rb_fstring_lit("%.*d"); + int width = RSTRING_LENINT(beg); + b = rb_str_to_inum(beg, 10, FALSE); + if (FIXNUM_P(b)) { + long bi = FIX2LONG(b); + rb_encoding *usascii = rb_usascii_encoding(); + + while (FIXABLE(bi)) { + if ((*each)(rb_enc_sprintf(usascii, "%.*ld", width, bi), arg)) break; + bi++; + } + b = LONG2NUM(bi); + } + args[0] = INT2FIX(width); + while (1) { + args[1] = b; + if ((*each)(rb_str_format(numberof(args), args, fmt), arg)) break; + b = rb_funcallv(b, succ, 0, 0); + } } /* normal case */ current = str_duplicate(rb_cString, beg); while (1) { - VALUE next = rb_funcallv(current, succ, 0, 0); - if ((*each)(current, arg)) break; - current = next; - StringValue(current); - if (RSTRING_LEN(current) == 0) - break; + VALUE next = rb_funcallv(current, succ, 0, 0); + if ((*each)(current, arg)) break; + current = next; + StringValue(current); + if (RSTRING_LEN(current) == 0) + break; } return beg; @@ -5031,32 +5210,32 @@ rb_str_include_range_p(VALUE beg, VALUE end, VALUE val, VALUE exclusive) val = rb_check_string_type(val); if (NIL_P(val)) return Qfalse; if (rb_enc_asciicompat(STR_ENC_GET(beg)) && - rb_enc_asciicompat(STR_ENC_GET(end)) && - rb_enc_asciicompat(STR_ENC_GET(val))) { - const char *bp = RSTRING_PTR(beg); - const char *ep = RSTRING_PTR(end); - const char *vp = RSTRING_PTR(val); - if (RSTRING_LEN(beg) == 1 && RSTRING_LEN(end) == 1) { - if (RSTRING_LEN(val) == 0 || RSTRING_LEN(val) > 1) - return Qfalse; - else { - char b = *bp; - char e = *ep; - char v = *vp; - - if (ISASCII(b) && ISASCII(e) && ISASCII(v)) { - if (b <= v && v < e) return Qtrue; - return RBOOL(!RTEST(exclusive) && v == e); - } - } - } + rb_enc_asciicompat(STR_ENC_GET(end)) && + rb_enc_asciicompat(STR_ENC_GET(val))) { + const char *bp = RSTRING_PTR(beg); + const char *ep = RSTRING_PTR(end); + const char *vp = RSTRING_PTR(val); + if (RSTRING_LEN(beg) == 1 && RSTRING_LEN(end) == 1) { + if (RSTRING_LEN(val) == 0 || RSTRING_LEN(val) > 1) + return Qfalse; + else { + char b = *bp; + char e = *ep; + char v = *vp; + + if (ISASCII(b) && ISASCII(e) && ISASCII(v)) { + if (b <= v && v < e) return Qtrue; + return RBOOL(!RTEST(exclusive) && v == e); + } + } + } #if 0 - /* both edges are all digits */ - if (ISDIGIT(*bp) && ISDIGIT(*ep) && - all_digits_p(bp, RSTRING_LEN(beg)) && - all_digits_p(ep, RSTRING_LEN(end))) { - /* TODO */ - } + /* both edges are all digits */ + if (ISDIGIT(*bp) && ISDIGIT(*ep) && + all_digits_p(bp, RSTRING_LEN(beg)) && + all_digits_p(ep, RSTRING_LEN(end))) { + /* TODO */ + } #endif } rb_str_upto_each(beg, end, RTEST(exclusive), include_range_i, (VALUE)&val); @@ -5070,7 +5249,7 @@ rb_str_subpat(VALUE str, VALUE re, VALUE backref) if (rb_reg_search(re, str, 0, 0) >= 0) { VALUE match = rb_backref_get(); int nth = rb_reg_backref_number(match, backref); - return rb_reg_nth_match(nth, match); + return rb_reg_nth_match(nth, match); } return Qnil; } @@ -5081,28 +5260,28 @@ rb_str_aref(VALUE str, VALUE indx) long idx; if (FIXNUM_P(indx)) { - idx = FIX2LONG(indx); + idx = FIX2LONG(indx); } else if (RB_TYPE_P(indx, T_REGEXP)) { - return rb_str_subpat(str, indx, INT2FIX(0)); + return rb_str_subpat(str, indx, INT2FIX(0)); } else if (RB_TYPE_P(indx, T_STRING)) { - if (rb_str_index(str, indx, 0) != -1) + if (rb_str_index(str, indx, 0) != -1) return str_duplicate(rb_cString, indx); - return Qnil; + return Qnil; } else { - /* check if indx is Range */ - long beg, len = str_strlen(str, NULL); - switch (rb_range_beg_len(indx, &beg, &len, len, 0)) { - case Qfalse: - break; - case Qnil: - return Qnil; - default: - return rb_str_substr(str, beg, len); - } - idx = NUM2LONG(indx); + /* check if indx is Range */ + long beg, len = str_strlen(str, NULL); + switch (rb_range_beg_len(indx, &beg, &len, len, 0)) { + case Qfalse: + break; + case Qnil: + return Qnil; + default: + return rb_str_substr(str, beg, len); + } + idx = NUM2LONG(indx); } return str_substr(str, idx, 1, FALSE); @@ -5127,14 +5306,14 @@ static VALUE rb_str_aref_m(int argc, VALUE *argv, VALUE str) { if (argc == 2) { - if (RB_TYPE_P(argv[0], T_REGEXP)) { - return rb_str_subpat(str, argv[0], argv[1]); - } - else { - long beg = NUM2LONG(argv[0]); - long len = NUM2LONG(argv[1]); - return rb_str_substr(str, beg, len); - } + if (RB_TYPE_P(argv[0], T_REGEXP)) { + return rb_str_subpat(str, argv[0], argv[1]); + } + else { + long beg = NUM2LONG(argv[0]); + long len = NUM2LONG(argv[1]); + return rb_str_substr(str, beg, len); + } } rb_check_arity(argc, 1, 2); return rb_str_aref(str, argv[0]); @@ -5150,13 +5329,12 @@ rb_str_drop_bytes(VALUE str, long len) if (len > olen) len = olen; nlen = olen - len; if (str_embed_capa(str) >= nlen + TERM_LEN(str)) { - char *oldptr = ptr; - int fl = (int)(RBASIC(str)->flags & (STR_NOEMBED|STR_SHARED|STR_NOFREE)); - STR_SET_EMBED(str); - STR_SET_EMBED_LEN(str, nlen); + char *oldptr = ptr; + int fl = (int)(RBASIC(str)->flags & (STR_NOEMBED|STR_SHARED|STR_NOFREE)); + STR_SET_EMBED(str); ptr = RSTRING(str)->as.embed.ary; - memmove(ptr, oldptr + len, nlen); - if (fl == STR_NOEMBED) xfree(oldptr); + memmove(ptr, oldptr + len, nlen); + if (fl == STR_NOEMBED) xfree(oldptr); } else { if (!STR_SHARED_P(str)) { @@ -5164,49 +5342,52 @@ rb_str_drop_bytes(VALUE str, long len) rb_enc_cr_str_exact_copy(shared, str); OBJ_FREEZE(shared); } - ptr = RSTRING(str)->as.heap.ptr += len; - RSTRING(str)->as.heap.len = nlen; + ptr = RSTRING(str)->as.heap.ptr += len; + } + STR_SET_LEN(str, nlen); + + if (!SHARABLE_MIDDLE_SUBSTRING) { + TERM_FILL(ptr + nlen, TERM_LEN(str)); } - ptr[nlen] = 0; ENC_CODERANGE_CLEAR(str); return str; } static void -rb_str_splice_0(VALUE str, long beg, long len, VALUE val) +rb_str_update_1(VALUE str, long beg, long len, VALUE val, long vbeg, long vlen) { char *sptr; - long slen, vlen = RSTRING_LEN(val); + long slen; int cr; if (beg == 0 && vlen == 0) { - rb_str_drop_bytes(str, len); - return; + rb_str_drop_bytes(str, len); + return; } str_modify_keep_cr(str); RSTRING_GETMEM(str, sptr, slen); if (len < vlen) { - /* expand string */ - RESIZE_CAPA(str, slen + vlen - len); - sptr = RSTRING_PTR(str); + /* expand string */ + RESIZE_CAPA(str, slen + vlen - len); + sptr = RSTRING_PTR(str); } if (ENC_CODERANGE(str) == ENC_CODERANGE_7BIT) - cr = rb_enc_str_coderange(val); + cr = rb_enc_str_coderange(val); else - cr = ENC_CODERANGE_UNKNOWN; + cr = ENC_CODERANGE_UNKNOWN; if (vlen != len) { - memmove(sptr + beg + vlen, - sptr + beg + len, - slen - (beg + len)); + memmove(sptr + beg + vlen, + sptr + beg + len, + slen - (beg + len)); } if (vlen < beg && len < 0) { - MEMZERO(sptr + slen, char, -len); + MEMZERO(sptr + slen, char, -len); } if (vlen > 0) { - memmove(sptr + beg, RSTRING_PTR(val), vlen); + memmove(sptr + beg, RSTRING_PTR(val) + vbeg, vlen); } slen += vlen - len; STR_SET_LEN(str, slen); @@ -5214,6 +5395,12 @@ rb_str_splice_0(VALUE str, long beg, long len, VALUE val) ENC_CODERANGE_SET(str, cr); } +static inline void +rb_str_update_0(VALUE str, long beg, long len, VALUE val) +{ + rb_str_update_1(str, beg, len, val, 0, RSTRING_LEN(val)); +} + void rb_str_update(VALUE str, long beg, long len, VALUE val) { @@ -5230,17 +5417,17 @@ rb_str_update(VALUE str, long beg, long len, VALUE val) slen = str_strlen(str, enc); /* rb_enc_check */ if ((slen < beg) || ((beg < 0) && (beg + slen < 0))) { - rb_raise(rb_eIndexError, "index %ld out of string", beg); + rb_raise(rb_eIndexError, "index %ld out of string", beg); } if (beg < 0) { - beg += slen; + beg += slen; } - assert(beg >= 0); - assert(beg <= slen); + RUBY_ASSERT(beg >= 0); + RUBY_ASSERT(beg <= slen); + if (len > slen - beg) { - len = slen - beg; + len = slen - beg; } - str_modify_keep_cr(str); p = str_nth(RSTRING_PTR(str), RSTRING_END(str), beg, enc, singlebyte); if (!p) p = RSTRING_END(str); e = str_nth(p, RSTRING_END(str), len, enc, singlebyte); @@ -5248,15 +5435,13 @@ rb_str_update(VALUE str, long beg, long len, VALUE val) /* error check */ beg = p - RSTRING_PTR(str); /* physical position */ len = e - p; /* physical length */ - rb_str_splice_0(str, beg, len, val); + rb_str_update_0(str, beg, len, val); rb_enc_associate(str, enc); cr = ENC_CODERANGE_AND(ENC_CODERANGE(str), ENC_CODERANGE(val)); if (cr != ENC_CODERANGE_BROKEN) - ENC_CODERANGE_SET(str, cr); + ENC_CODERANGE_SET(str, cr); } -#define rb_str_splice(str, beg, len, val) rb_str_update(str, beg, len, val) - static void rb_str_subpat_set(VALUE str, VALUE re, VALUE backref, VALUE val) { @@ -5267,27 +5452,27 @@ rb_str_subpat_set(VALUE str, VALUE re, VALUE backref, VALUE val) struct re_registers *regs; if (rb_reg_search(re, str, 0, 0) < 0) { - rb_raise(rb_eIndexError, "regexp not matched"); + rb_raise(rb_eIndexError, "regexp not matched"); } match = rb_backref_get(); nth = rb_reg_backref_number(match, backref); regs = RMATCH_REGS(match); if ((nth >= regs->num_regs) || ((nth < 0) && (-nth >= regs->num_regs))) { - rb_raise(rb_eIndexError, "index %d out of regexp", nth); + rb_raise(rb_eIndexError, "index %d out of regexp", nth); } if (nth < 0) { - nth += regs->num_regs; + nth += regs->num_regs; } start = BEG(nth); if (start == -1) { - rb_raise(rb_eIndexError, "regexp group %d not matched", nth); + rb_raise(rb_eIndexError, "regexp group %d not matched", nth); } end = END(nth); len = end - start; StringValue(val); enc = rb_enc_check_str(str, val); - rb_str_splice_0(str, start, len, val); + rb_str_update_0(str, start, len, val); rb_enc_associate(str, enc); } @@ -5298,32 +5483,32 @@ rb_str_aset(VALUE str, VALUE indx, VALUE val) switch (TYPE(indx)) { case T_REGEXP: - rb_str_subpat_set(str, indx, INT2FIX(0), val); - return val; + rb_str_subpat_set(str, indx, INT2FIX(0), val); + return val; case T_STRING: - beg = rb_str_index(str, indx, 0); - if (beg < 0) { - rb_raise(rb_eIndexError, "string not matched"); - } - beg = rb_str_sublen(str, beg); - rb_str_splice(str, beg, str_strlen(indx, NULL), val); - return val; + beg = rb_str_index(str, indx, 0); + if (beg < 0) { + rb_raise(rb_eIndexError, "string not matched"); + } + beg = rb_str_sublen(str, beg); + rb_str_update(str, beg, str_strlen(indx, NULL), val); + return val; default: - /* check if indx is Range */ - { - long beg, len; - if (rb_range_beg_len(indx, &beg, &len, str_strlen(str, NULL), 2)) { - rb_str_splice(str, beg, len, val); - return val; - } - } + /* check if indx is Range */ + { + long beg, len; + if (rb_range_beg_len(indx, &beg, &len, str_strlen(str, NULL), 2)) { + rb_str_update(str, beg, len, val); + return val; + } + } /* FALLTHROUGH */ case T_FIXNUM: - idx = NUM2LONG(indx); - rb_str_splice(str, idx, 1, val); + idx = NUM2LONG(indx); + rb_str_update(str, idx, 1, val); return val; } } @@ -5333,7 +5518,7 @@ rb_str_aset(VALUE str, VALUE indx, VALUE val) * string[index] = new_string * string[start, length] = new_string * string[range] = new_string - * string[regexp, capture = 0) = new_string + * string[regexp, capture = 0] = new_string * string[substring] = new_string * * Replaces all, some, or none of the contents of +self+; returns +new_string+. @@ -5353,21 +5538,19 @@ rb_str_aset(VALUE str, VALUE indx, VALUE val) * s['lly'] = 'ncial' # => "ncial" * s # => "financial" * - * String#slice is an alias for String#[]. - * */ static VALUE rb_str_aset_m(int argc, VALUE *argv, VALUE str) { if (argc == 3) { - if (RB_TYPE_P(argv[0], T_REGEXP)) { - rb_str_subpat_set(str, argv[0], argv[1], argv[2]); - } - else { - rb_str_splice(str, NUM2LONG(argv[0]), NUM2LONG(argv[1]), argv[2]); - } - return argv[2]; + if (RB_TYPE_P(argv[0], T_REGEXP)) { + rb_str_subpat_set(str, argv[0], argv[1], argv[2]); + } + else { + rb_str_update(str, NUM2LONG(argv[0]), NUM2LONG(argv[1]), argv[2]); + } + return argv[2]; } rb_check_arity(argc, 2, 3); return rb_str_aset(str, argv[0], argv[1]); @@ -5379,11 +5562,11 @@ rb_str_aset_m(int argc, VALUE *argv, VALUE str) * * Inserts the given +other_string+ into +self+; returns +self+. * - * If the \Integer +index+ is positive, inserts +other_string+ at offset +index+: + * If the Integer +index+ is positive, inserts +other_string+ at offset +index+: * * 'foo'.insert(1, 'bar') # => "fbaroo" * - * If the \Integer +index+ is negative, counts backward from the end of +self+ + * If the Integer +index+ is negative, counts backward from the end of +self+ * and inserts +other_string+ at offset <tt>index+1</tt> * (that is, _after_ <tt>self[index]</tt>): * @@ -5397,12 +5580,12 @@ rb_str_insert(VALUE str, VALUE idx, VALUE str2) long pos = NUM2LONG(idx); if (pos == -1) { - return rb_str_append(str, str2); + return rb_str_append(str, str2); } else if (pos < 0) { - pos++; + pos++; } - rb_str_splice(str, pos, 0, str2); + rb_str_update(str, pos, 0, str2); return str; } @@ -5441,50 +5624,50 @@ rb_str_slice_bang(int argc, VALUE *argv, VALUE str) str_modify_keep_cr(str); indx = argv[0]; if (RB_TYPE_P(indx, T_REGEXP)) { - if (rb_reg_search(indx, str, 0, 0) < 0) return Qnil; - VALUE match = rb_backref_get(); - struct re_registers *regs = RMATCH_REGS(match); - int nth = 0; - if (argc > 1 && (nth = rb_reg_backref_number(match, argv[1])) < 0) { - if ((nth += regs->num_regs) <= 0) return Qnil; - } - else if (nth >= regs->num_regs) return Qnil; - beg = BEG(nth); - len = END(nth) - beg; + if (rb_reg_search(indx, str, 0, 0) < 0) return Qnil; + VALUE match = rb_backref_get(); + struct re_registers *regs = RMATCH_REGS(match); + int nth = 0; + if (argc > 1 && (nth = rb_reg_backref_number(match, argv[1])) < 0) { + if ((nth += regs->num_regs) <= 0) return Qnil; + } + else if (nth >= regs->num_regs) return Qnil; + beg = BEG(nth); + len = END(nth) - beg; goto subseq; } else if (argc == 2) { - beg = NUM2LONG(indx); - len = NUM2LONG(argv[1]); + beg = NUM2LONG(indx); + len = NUM2LONG(argv[1]); goto num_index; } else if (FIXNUM_P(indx)) { - beg = FIX2LONG(indx); - if (!(p = rb_str_subpos(str, beg, &len))) return Qnil; - if (!len) return Qnil; - beg = p - RSTRING_PTR(str); - goto subseq; + beg = FIX2LONG(indx); + if (!(p = rb_str_subpos(str, beg, &len))) return Qnil; + if (!len) return Qnil; + beg = p - RSTRING_PTR(str); + goto subseq; } else if (RB_TYPE_P(indx, T_STRING)) { - beg = rb_str_index(str, indx, 0); - if (beg == -1) return Qnil; - len = RSTRING_LEN(indx); + beg = rb_str_index(str, indx, 0); + if (beg == -1) return Qnil; + len = RSTRING_LEN(indx); result = str_duplicate(rb_cString, indx); goto squash; } else { - switch (rb_range_beg_len(indx, &beg, &len, str_strlen(str, NULL), 0)) { - case Qnil: - return Qnil; - case Qfalse: - beg = NUM2LONG(indx); - if (!(p = rb_str_subpos(str, beg, &len))) return Qnil; - if (!len) return Qnil; - beg = p - RSTRING_PTR(str); - goto subseq; - default: - goto num_index; - } + switch (rb_range_beg_len(indx, &beg, &len, str_strlen(str, NULL), 0)) { + case Qnil: + return Qnil; + case Qfalse: + beg = NUM2LONG(indx); + if (!(p = rb_str_subpos(str, beg, &len))) return Qnil; + if (!len) return Qnil; + beg = p - RSTRING_PTR(str); + goto subseq; + default: + goto num_index; + } } num_index: @@ -5497,21 +5680,21 @@ rb_str_slice_bang(int argc, VALUE *argv, VALUE str) squash: if (len > 0) { - if (beg == 0) { - rb_str_drop_bytes(str, len); - } - else { - char *sptr = RSTRING_PTR(str); - long slen = RSTRING_LEN(str); - if (beg + len > slen) /* pathological check */ - len = slen - beg; - memmove(sptr + beg, - sptr + beg + len, - slen - (beg + len)); - slen -= len; - STR_SET_LEN(str, slen); - TERM_FILL(&sptr[slen], TERM_LEN(str)); - } + if (beg == 0) { + rb_str_drop_bytes(str, len); + } + else { + char *sptr = RSTRING_PTR(str); + long slen = RSTRING_LEN(str); + if (beg + len > slen) /* pathological check */ + len = slen - beg; + memmove(sptr + beg, + sptr + beg + len, + slen - (beg + len)); + slen -= len; + STR_SET_LEN(str, slen); + TERM_FILL(&sptr[slen], TERM_LEN(str)); + } } return result; } @@ -5523,17 +5706,17 @@ get_pat(VALUE pat) switch (OBJ_BUILTIN_TYPE(pat)) { case T_REGEXP: - return pat; + return pat; case T_STRING: - break; + break; default: - val = rb_check_string_type(pat); - if (NIL_P(val)) { - Check_Type(pat, T_REGEXP); - } - pat = val; + val = rb_check_string_type(pat); + if (NIL_P(val)) { + Check_Type(pat, T_REGEXP); + } + pat = val; } return rb_reg_regcomp(pat); @@ -5546,20 +5729,20 @@ get_pat_quoted(VALUE pat, int check) switch (OBJ_BUILTIN_TYPE(pat)) { case T_REGEXP: - return pat; + return pat; case T_STRING: - break; + break; default: - val = rb_check_string_type(pat); - if (NIL_P(val)) { - Check_Type(pat, T_REGEXP); - } - pat = val; + val = rb_check_string_type(pat); + if (NIL_P(val)) { + Check_Type(pat, T_REGEXP); + } + pat = val; } if (check && is_broken_string(pat)) { - rb_exc_raise(rb_reg_check_preprocess(pat)); + rb_exc_raise(rb_reg_check_preprocess(pat)); } return pat; } @@ -5568,20 +5751,20 @@ static long rb_pat_search(VALUE pat, VALUE str, long pos, int set_backref_str) { if (BUILTIN_TYPE(pat) == T_STRING) { - pos = rb_strseq_index(str, pat, pos, 1); - if (set_backref_str) { - if (pos >= 0) { + pos = rb_str_byteindex(str, pat, pos); + if (set_backref_str) { + if (pos >= 0) { str = rb_str_new_frozen_String(str); - rb_backref_set_string(str, pos, RSTRING_LEN(pat)); - } - else { - rb_backref_set(Qnil); - } - } - return pos; + rb_backref_set_string(str, pos, RSTRING_LEN(pat)); + } + else { + rb_backref_set(Qnil); + } + } + return pos; } else { - return rb_reg_search0(pat, str, pos, 0, set_backref_str); + return rb_reg_search0(pat, str, pos, 0, set_backref_str); } } @@ -5611,14 +5794,14 @@ rb_str_sub_bang(int argc, VALUE *argv, VALUE str) rb_check_arity(argc, min_arity, 2); if (argc == 1) { - iter = 1; + iter = 1; } else { - repl = argv[1]; - hash = rb_check_hash_type(argv[1]); - if (NIL_P(hash)) { - StringValue(repl); - } + repl = argv[1]; + hash = rb_check_hash_type(argv[1]); + if (NIL_P(hash)) { + StringValue(repl); + } } pat = get_pat_quoted(argv[0], 1); @@ -5626,29 +5809,29 @@ rb_str_sub_bang(int argc, VALUE *argv, VALUE str) str_modifiable(str); beg = rb_pat_search(pat, str, 0, 1); if (beg >= 0) { - rb_encoding *enc; - int cr = ENC_CODERANGE(str); - long beg0, end0; - VALUE match, match0 = Qnil; - struct re_registers *regs; - char *p, *rp; - long len, rlen; - - match = rb_backref_get(); - regs = RMATCH_REGS(match); - if (RB_TYPE_P(pat, T_STRING)) { - beg0 = beg; - end0 = beg0 + RSTRING_LEN(pat); - match0 = pat; - } - else { - beg0 = BEG(0); - end0 = END(0); - if (iter) match0 = rb_reg_nth_match(0, match); - } - - if (iter || !NIL_P(hash)) { - p = RSTRING_PTR(str); len = RSTRING_LEN(str); + rb_encoding *enc; + int cr = ENC_CODERANGE(str); + long beg0, end0; + VALUE match, match0 = Qnil; + struct re_registers *regs; + char *p, *rp; + long len, rlen; + + match = rb_backref_get(); + regs = RMATCH_REGS(match); + if (RB_TYPE_P(pat, T_STRING)) { + beg0 = beg; + end0 = beg0 + RSTRING_LEN(pat); + match0 = pat; + } + else { + beg0 = BEG(0); + end0 = END(0); + if (iter) match0 = rb_reg_nth_match(0, match); + } + + if (iter || !NIL_P(hash)) { + p = RSTRING_PTR(str); len = RSTRING_LEN(str); if (iter) { repl = rb_obj_as_string(rb_yield(match0)); @@ -5657,53 +5840,55 @@ rb_str_sub_bang(int argc, VALUE *argv, VALUE str) repl = rb_hash_aref(hash, rb_str_subseq(str, beg0, end0 - beg0)); repl = rb_obj_as_string(repl); } - str_mod_check(str, p, len); - rb_check_frozen(str); - } - else { - repl = rb_reg_regsub(repl, str, regs, RB_TYPE_P(pat, T_STRING) ? Qnil : pat); - } + str_mod_check(str, p, len); + rb_check_frozen(str); + } + else { + repl = rb_reg_regsub(repl, str, regs, RB_TYPE_P(pat, T_STRING) ? Qnil : pat); + } enc = rb_enc_compatible(str, repl); if (!enc) { rb_encoding *str_enc = STR_ENC_GET(str); - p = RSTRING_PTR(str); len = RSTRING_LEN(str); - if (coderange_scan(p, beg0, str_enc) != ENC_CODERANGE_7BIT || - coderange_scan(p+end0, len-end0, str_enc) != ENC_CODERANGE_7BIT) { + p = RSTRING_PTR(str); len = RSTRING_LEN(str); + if (coderange_scan(p, beg0, str_enc) != ENC_CODERANGE_7BIT || + coderange_scan(p+end0, len-end0, str_enc) != ENC_CODERANGE_7BIT) { rb_raise(rb_eEncCompatError, "incompatible character encodings: %s and %s", - rb_enc_name(str_enc), - rb_enc_name(STR_ENC_GET(repl))); + rb_enc_inspect_name(str_enc), + rb_enc_inspect_name(STR_ENC_GET(repl))); } enc = STR_ENC_GET(repl); } - rb_str_modify(str); - rb_enc_associate(str, enc); - if (ENC_CODERANGE_UNKNOWN < cr && cr < ENC_CODERANGE_BROKEN) { - int cr2 = ENC_CODERANGE(repl); + rb_str_modify(str); + rb_enc_associate(str, enc); + if (ENC_CODERANGE_UNKNOWN < cr && cr < ENC_CODERANGE_BROKEN) { + int cr2 = ENC_CODERANGE(repl); if (cr2 == ENC_CODERANGE_BROKEN || (cr == ENC_CODERANGE_VALID && cr2 == ENC_CODERANGE_7BIT)) cr = ENC_CODERANGE_UNKNOWN; else cr = cr2; - } - plen = end0 - beg0; + } + plen = end0 - beg0; rlen = RSTRING_LEN(repl); - len = RSTRING_LEN(str); - if (rlen > plen) { - RESIZE_CAPA(str, len + rlen - plen); - } - p = RSTRING_PTR(str); - if (rlen != plen) { - memmove(p + beg0 + rlen, p + beg0 + plen, len - beg0 - plen); - } + len = RSTRING_LEN(str); + if (rlen > plen) { + RESIZE_CAPA(str, len + rlen - plen); + } + p = RSTRING_PTR(str); + if (rlen != plen) { + memmove(p + beg0 + rlen, p + beg0 + plen, len - beg0 - plen); + } rp = RSTRING_PTR(repl); memmove(p + beg0, rp, rlen); - len += rlen - plen; - STR_SET_LEN(str, len); - TERM_FILL(&RSTRING_PTR(str)[len], TERM_LEN(str)); - ENC_CODERANGE_SET(str, cr); + len += rlen - plen; + STR_SET_LEN(str, len); + TERM_FILL(&RSTRING_PTR(str)[len], TERM_LEN(str)); + ENC_CODERANGE_SET(str, cr); - return str; + RB_GC_GUARD(match); + + return str; } return Qnil; } @@ -5734,8 +5919,7 @@ rb_str_sub(int argc, VALUE *argv, VALUE str) static VALUE str_gsub(int argc, VALUE *argv, VALUE str, int bang) { - VALUE pat, val = Qnil, repl, match, match0 = Qnil, dest, hash = Qnil; - struct re_registers *regs; + VALUE pat, val = Qnil, repl, match0 = Qnil, dest, hash = Qnil; long beg, beg0, end0; long offset, blen, slen, len, last; enum {STR, ITER, MAP} mode = STR; @@ -5745,19 +5929,19 @@ str_gsub(int argc, VALUE *argv, VALUE str, int bang) switch (argc) { case 1: - RETURN_ENUMERATOR(str, argc, argv); - mode = ITER; - break; + RETURN_ENUMERATOR(str, argc, argv); + mode = ITER; + break; case 2: - repl = argv[1]; - hash = rb_check_hash_type(argv[1]); - if (NIL_P(hash)) { - StringValue(repl); - } - else { - mode = MAP; - } - break; + repl = argv[1]; + hash = rb_check_hash_type(argv[1]); + if (NIL_P(hash)) { + StringValue(repl); + } + else { + mode = MAP; + } + break; default: rb_error_arity(argc, 1, 2); } @@ -5765,7 +5949,7 @@ str_gsub(int argc, VALUE *argv, VALUE str, int bang) pat = get_pat_quoted(argv[0], 1); beg = rb_pat_search(pat, str, 0, need_backref); if (beg < 0) { - if (bang) return Qnil; /* no match, no substitution */ + if (bang) return Qnil; /* no match, no substitution */ return str_duplicate(rb_cString, str); } @@ -5780,20 +5964,20 @@ str_gsub(int argc, VALUE *argv, VALUE str, int bang) ENC_CODERANGE_SET(dest, rb_enc_asciicompat(str_enc) ? ENC_CODERANGE_7BIT : ENC_CODERANGE_VALID); do { - match = rb_backref_get(); - regs = RMATCH_REGS(match); - if (RB_TYPE_P(pat, T_STRING)) { - beg0 = beg; - end0 = beg0 + RSTRING_LEN(pat); - match0 = pat; - } - else { - beg0 = BEG(0); - end0 = END(0); - if (mode == ITER) match0 = rb_reg_nth_match(0, match); - } - - if (mode) { + VALUE match = rb_backref_get(); + struct re_registers *regs = RMATCH_REGS(match); + if (RB_TYPE_P(pat, T_STRING)) { + beg0 = beg; + end0 = beg0 + RSTRING_LEN(pat); + match0 = pat; + } + else { + beg0 = BEG(0); + end0 = END(0); + if (mode == ITER) match0 = rb_reg_nth_match(0, match); + } + + if (mode) { if (mode == ITER) { val = rb_obj_as_string(rb_yield(match0)); } @@ -5801,43 +5985,45 @@ str_gsub(int argc, VALUE *argv, VALUE str, int bang) val = rb_hash_aref(hash, rb_str_subseq(str, beg0, end0 - beg0)); val = rb_obj_as_string(val); } - str_mod_check(str, sp, slen); - if (val == dest) { /* paranoid check [ruby-dev:24827] */ - rb_raise(rb_eRuntimeError, "block should not cheat"); - } - } - else if (need_backref) { - val = rb_reg_regsub(repl, str, regs, RB_TYPE_P(pat, T_STRING) ? Qnil : pat); - if (need_backref < 0) { - need_backref = val != repl; - } - } - else { - val = repl; - } - - len = beg0 - offset; /* copy pre-match substr */ + str_mod_check(str, sp, slen); + if (val == dest) { /* paranoid check [ruby-dev:24827] */ + rb_raise(rb_eRuntimeError, "block should not cheat"); + } + } + else if (need_backref) { + val = rb_reg_regsub(repl, str, regs, RB_TYPE_P(pat, T_STRING) ? Qnil : pat); + if (need_backref < 0) { + need_backref = val != repl; + } + } + else { + val = repl; + } + + len = beg0 - offset; /* copy pre-match substr */ if (len) { rb_enc_str_buf_cat(dest, cp, len, str_enc); } rb_str_buf_append(dest, val); - last = offset; - offset = end0; - if (beg0 == end0) { - /* - * Always consume at least one character of the input string - * in order to prevent infinite loops. - */ - if (RSTRING_LEN(str) <= end0) break; - len = rb_enc_fast_mbclen(RSTRING_PTR(str)+end0, RSTRING_END(str), str_enc); + last = offset; + offset = end0; + if (beg0 == end0) { + /* + * Always consume at least one character of the input string + * in order to prevent infinite loops. + */ + if (RSTRING_LEN(str) <= end0) break; + len = rb_enc_fast_mbclen(RSTRING_PTR(str)+end0, RSTRING_END(str), str_enc); rb_enc_str_buf_cat(dest, RSTRING_PTR(str)+end0, len, str_enc); - offset = end0 + len; - } - cp = RSTRING_PTR(str) + offset; - if (offset > RSTRING_LEN(str)) break; - beg = rb_pat_search(pat, str, offset, need_backref); + offset = end0 + len; + } + cp = RSTRING_PTR(str) + offset; + if (offset > RSTRING_LEN(str)) break; + beg = rb_pat_search(pat, str, offset, need_backref); + + RB_GC_GUARD(match); } while (beg >= 0); if (RSTRING_LEN(str) > offset) { rb_enc_str_buf_cat(dest, cp, RSTRING_LEN(str) - offset, str_enc); @@ -5847,7 +6033,7 @@ str_gsub(int argc, VALUE *argv, VALUE str, int bang) str_shared_replace(str, dest); } else { - str = dest; + str = dest; } return str; @@ -5940,12 +6126,12 @@ rb_str_clear(VALUE str) { str_discard(str); STR_SET_EMBED(str); - STR_SET_EMBED_LEN(str, 0); + STR_SET_LEN(str, 0); RSTRING_PTR(str)[0] = 0; if (rb_enc_asciicompat(STR_ENC_GET(str))) - ENC_CODERANGE_SET(str, ENC_CODERANGE_7BIT); + ENC_CODERANGE_SET(str, ENC_CODERANGE_7BIT); else - ENC_CODERANGE_SET(str, ENC_CODERANGE_VALID); + ENC_CODERANGE_SET(str, ENC_CODERANGE_VALID); return str; } @@ -5979,7 +6165,7 @@ rb_str_chr(VALUE str) * * Related: String#setbyte. */ -static VALUE +VALUE rb_str_getbyte(VALUE str, VALUE index) { long pos = NUM2LONG(index); @@ -6004,7 +6190,7 @@ rb_str_getbyte(VALUE str, VALUE index) * * Related: String#getbyte. */ -static VALUE +VALUE rb_str_setbyte(VALUE str, VALUE index, VALUE value) { long pos = NUM2LONG(index); @@ -6023,34 +6209,34 @@ rb_str_setbyte(VALUE str, VALUE index, VALUE value) char byte = (char)(NUM2INT(w) & 0xFF); if (!str_independent(str)) - str_make_independent(str); + str_make_independent(str); enc = STR_ENC_GET(str); head = RSTRING_PTR(str); ptr = &head[pos]; if (!STR_EMBED_P(str)) { - cr = ENC_CODERANGE(str); - switch (cr) { - case ENC_CODERANGE_7BIT: + cr = ENC_CODERANGE(str); + switch (cr) { + case ENC_CODERANGE_7BIT: left = ptr; - *ptr = byte; - if (ISASCII(byte)) goto end; - nlen = rb_enc_precise_mbclen(left, head+len, enc); - if (!MBCLEN_CHARFOUND_P(nlen)) - ENC_CODERANGE_SET(str, ENC_CODERANGE_BROKEN); - else - ENC_CODERANGE_SET(str, ENC_CODERANGE_VALID); - goto end; - case ENC_CODERANGE_VALID: - left = rb_enc_left_char_head(head, ptr, head+len, enc); - width = rb_enc_precise_mbclen(left, head+len, enc); - *ptr = byte; - nlen = rb_enc_precise_mbclen(left, head+len, enc); - if (!MBCLEN_CHARFOUND_P(nlen)) - ENC_CODERANGE_SET(str, ENC_CODERANGE_BROKEN); - else if (MBCLEN_CHARFOUND_LEN(nlen) != width || ISASCII(byte)) - ENC_CODERANGE_CLEAR(str); - goto end; - } + *ptr = byte; + if (ISASCII(byte)) goto end; + nlen = rb_enc_precise_mbclen(left, head+len, enc); + if (!MBCLEN_CHARFOUND_P(nlen)) + ENC_CODERANGE_SET(str, ENC_CODERANGE_BROKEN); + else + ENC_CODERANGE_SET(str, ENC_CODERANGE_VALID); + goto end; + case ENC_CODERANGE_VALID: + left = rb_enc_left_char_head(head, ptr, head+len, enc); + width = rb_enc_precise_mbclen(left, head+len, enc); + *ptr = byte; + nlen = rb_enc_precise_mbclen(left, head+len, enc); + if (!MBCLEN_CHARFOUND_P(nlen)) + ENC_CODERANGE_SET(str, ENC_CODERANGE_BROKEN); + else if (MBCLEN_CHARFOUND_LEN(nlen) != width || ISASCII(byte)) + ENC_CODERANGE_CLEAR(str); + goto end; + } } ENC_CODERANGE_CLEAR(str); *ptr = byte; @@ -6062,78 +6248,71 @@ rb_str_setbyte(VALUE str, VALUE index, VALUE value) static VALUE str_byte_substr(VALUE str, long beg, long len, int empty) { - char *p, *s = RSTRING_PTR(str); long n = RSTRING_LEN(str); - VALUE str2; if (beg > n || len < 0) return Qnil; if (beg < 0) { - beg += n; - if (beg < 0) return Qnil; + beg += n; + if (beg < 0) return Qnil; } if (len > n - beg) - len = n - beg; + len = n - beg; if (len <= 0) { - if (!empty) return Qnil; - len = 0; - p = 0; + if (!empty) return Qnil; + len = 0; } - else - p = s + beg; - if (!STR_EMBEDDABLE_P(len, TERM_LEN(str)) && SHARABLE_SUBSTRING_P(beg, len, n)) { - str2 = rb_str_new_frozen(str); - str2 = str_new_shared(rb_cString, str2); - RSTRING(str2)->as.heap.ptr += beg; - RSTRING(str2)->as.heap.len = len; - } - else { - str2 = rb_str_new(p, len); - } + VALUE str2 = str_subseq(str, beg, len); - str_enc_copy(str2, str); + str_enc_copy_direct(str2, str); if (RSTRING_LEN(str2) == 0) { - if (!rb_enc_asciicompat(STR_ENC_GET(str))) - ENC_CODERANGE_SET(str2, ENC_CODERANGE_VALID); - else - ENC_CODERANGE_SET(str2, ENC_CODERANGE_7BIT); + if (!rb_enc_asciicompat(STR_ENC_GET(str))) + ENC_CODERANGE_SET(str2, ENC_CODERANGE_VALID); + else + ENC_CODERANGE_SET(str2, ENC_CODERANGE_7BIT); } else { - switch (ENC_CODERANGE(str)) { - case ENC_CODERANGE_7BIT: - ENC_CODERANGE_SET(str2, ENC_CODERANGE_7BIT); - break; - default: - ENC_CODERANGE_SET(str2, ENC_CODERANGE_UNKNOWN); - break; - } + switch (ENC_CODERANGE(str)) { + case ENC_CODERANGE_7BIT: + ENC_CODERANGE_SET(str2, ENC_CODERANGE_7BIT); + break; + default: + ENC_CODERANGE_SET(str2, ENC_CODERANGE_UNKNOWN); + break; + } } return str2; } +VALUE +rb_str_byte_substr(VALUE str, VALUE beg, VALUE len) +{ + return str_byte_substr(str, NUM2LONG(beg), NUM2LONG(len), TRUE); +} + static VALUE str_byte_aref(VALUE str, VALUE indx) { long idx; if (FIXNUM_P(indx)) { - idx = FIX2LONG(indx); + idx = FIX2LONG(indx); } else { - /* check if indx is Range */ - long beg, len = RSTRING_LEN(str); + /* check if indx is Range */ + long beg, len = RSTRING_LEN(str); - switch (rb_range_beg_len(indx, &beg, &len, len, 0)) { - case Qfalse: - break; - case Qnil: - return Qnil; - default: - return str_byte_substr(str, beg, len, TRUE); - } + switch (rb_range_beg_len(indx, &beg, &len, len, 0)) { + case Qfalse: + break; + case Qnil: + return Qnil; + default: + return str_byte_substr(str, beg, len, TRUE); + } - idx = NUM2LONG(indx); + idx = NUM2LONG(indx); } return str_byte_substr(str, idx, 1, FALSE); } @@ -6193,16 +6372,44 @@ rb_str_byteslice(int argc, VALUE *argv, VALUE str) return str_byte_aref(str, argv[0]); } +static void +str_check_beg_len(VALUE str, long *beg, long *len) +{ + long end, slen = RSTRING_LEN(str); + + if (*len < 0) rb_raise(rb_eIndexError, "negative length %ld", *len); + if ((slen < *beg) || ((*beg < 0) && (*beg + slen < 0))) { + rb_raise(rb_eIndexError, "index %ld out of string", *beg); + } + if (*beg < 0) { + *beg += slen; + } + RUBY_ASSERT(*beg >= 0); + RUBY_ASSERT(*beg <= slen); + + if (*len > slen - *beg) { + *len = slen - *beg; + } + end = *beg + *len; + str_ensure_byte_pos(str, *beg); + str_ensure_byte_pos(str, end); +} + /* * call-seq: * bytesplice(index, length, str) -> string - * bytesplice(range, str) -> string + * bytesplice(index, length, str, str_index, str_length) -> string + * bytesplice(range, str) -> string + * bytesplice(range, str, str_range) -> string * - * Replaces some or all of the content of +self+ with +str+, and returns +str+. + * Replaces some or all of the content of +self+ with +str+, and returns +self+. * The portion of the string affected is determined using * the same criteria as String#byteslice, except that +length+ cannot be omitted. * If the replacement string is not the same length as the text it is replacing, * the string will be adjusted accordingly. + * + * If +str_index+ and +str_length+, or +str_range+ are given, the content of +self+ is replaced by str.byteslice(str_index, str_length) or str.byteslice(str_range); however the substring of +str+ is not allocated as a new string. + * * The form that take an Integer will raise an IndexError if the value is out * of range; the Range form will raise a RangeError. * If the beginning or ending offset does not land on character (codepoint) @@ -6212,55 +6419,61 @@ rb_str_byteslice(int argc, VALUE *argv, VALUE str) static VALUE rb_str_bytesplice(int argc, VALUE *argv, VALUE str) { - long beg, end, len, slen; + long beg, len, vbeg, vlen; VALUE val; rb_encoding *enc; int cr; - rb_check_arity(argc, 2, 3); - if (argc == 2) { + rb_check_arity(argc, 2, 5); + if (!(argc == 2 || argc == 3 || argc == 5)) { + rb_raise(rb_eArgError, "wrong number of arguments (given %d, expected 2, 3, or 5)", argc); + } + if (argc == 2 || (argc == 3 && !RB_INTEGER_TYPE_P(argv[0]))) { if (!rb_range_beg_len(argv[0], &beg, &len, RSTRING_LEN(str), 2)) { rb_raise(rb_eTypeError, "wrong argument type %s (expected Range)", rb_builtin_class_name(argv[0])); } val = argv[1]; + StringValue(val); + if (argc == 2) { + /* bytesplice(range, str) */ + vbeg = 0; + vlen = RSTRING_LEN(val); + } + else { + /* bytesplice(range, str, str_range) */ + if (!rb_range_beg_len(argv[2], &vbeg, &vlen, RSTRING_LEN(val), 2)) { + rb_raise(rb_eTypeError, "wrong argument type %s (expected Range)", + rb_builtin_class_name(argv[2])); + } + } } else { beg = NUM2LONG(argv[0]); len = NUM2LONG(argv[1]); val = argv[2]; + StringValue(val); + if (argc == 3) { + /* bytesplice(index, length, str) */ + vbeg = 0; + vlen = RSTRING_LEN(val); + } + else { + /* bytesplice(index, length, str, str_index, str_length) */ + vbeg = NUM2LONG(argv[3]); + vlen = NUM2LONG(argv[4]); + } } - if (len < 0) rb_raise(rb_eIndexError, "negative length %ld", len); - slen = RSTRING_LEN(str); - if ((slen < beg) || ((beg < 0) && (beg + slen < 0))) { - rb_raise(rb_eIndexError, "index %ld out of string", beg); - } - if (beg < 0) { - beg += slen; - } - assert(beg >= 0); - assert(beg <= slen); - if (len > slen - beg) { - len = slen - beg; - } - end = beg + len; - if (!str_check_byte_pos(str, beg)) { - rb_raise(rb_eIndexError, - "offset %ld does not land on character boundary", beg); - } - if (!str_check_byte_pos(str, end)) { - rb_raise(rb_eIndexError, - "offset %ld does not land on character boundary", end); - } - StringValue(val); + str_check_beg_len(str, &beg, &len); + str_check_beg_len(val, &vbeg, &vlen); enc = rb_enc_check(str, val); str_modify_keep_cr(str); - rb_str_splice_0(str, beg, len, val); + rb_str_update_1(str, beg, len, val, vbeg, vlen); rb_enc_associate(str, enc); cr = ENC_CODERANGE_AND(ENC_CODERANGE(str), ENC_CODERANGE(val)); if (cr != ENC_CODERANGE_BROKEN) ENC_CODERANGE_SET(str, cr); - return val; + return str; } /* @@ -6289,35 +6502,35 @@ rb_str_reverse(VALUE str) cr = ENC_CODERANGE(str); if (RSTRING_LEN(str) > 1) { - if (single_byte_optimizable(str)) { - while (s < e) { - *--p = *s++; - } - } - else if (cr == ENC_CODERANGE_VALID) { - while (s < e) { - int clen = rb_enc_fast_mbclen(s, e, enc); - - p -= clen; - memcpy(p, s, clen); - s += clen; - } - } - else { - cr = rb_enc_asciicompat(enc) ? - ENC_CODERANGE_7BIT : ENC_CODERANGE_VALID; - while (s < e) { - int clen = rb_enc_mbclen(s, e, enc); - - if (clen > 1 || (*s & 0x80)) cr = ENC_CODERANGE_UNKNOWN; - p -= clen; - memcpy(p, s, clen); - s += clen; - } - } + if (single_byte_optimizable(str)) { + while (s < e) { + *--p = *s++; + } + } + else if (cr == ENC_CODERANGE_VALID) { + while (s < e) { + int clen = rb_enc_fast_mbclen(s, e, enc); + + p -= clen; + memcpy(p, s, clen); + s += clen; + } + } + else { + cr = rb_enc_asciicompat(enc) ? + ENC_CODERANGE_7BIT : ENC_CODERANGE_VALID; + while (s < e) { + int clen = rb_enc_mbclen(s, e, enc); + + if (clen > 1 || (*s & 0x80)) cr = ENC_CODERANGE_UNKNOWN; + p -= clen; + memcpy(p, s, clen); + s += clen; + } + } } STR_SET_LEN(rev, RSTRING_LEN(str)); - str_enc_copy(rev, str); + str_enc_copy_direct(rev, str); ENC_CODERANGE_SET(rev, cr); return rev; @@ -6340,24 +6553,24 @@ static VALUE rb_str_reverse_bang(VALUE str) { if (RSTRING_LEN(str) > 1) { - if (single_byte_optimizable(str)) { - char *s, *e, c; - - str_modify_keep_cr(str); - s = RSTRING_PTR(str); - e = RSTRING_END(str) - 1; - while (s < e) { - c = *s; - *s++ = *e; - *e-- = c; - } - } - else { - str_shared_replace(str, rb_str_reverse(str)); - } + if (single_byte_optimizable(str)) { + char *s, *e, c; + + str_modify_keep_cr(str); + s = RSTRING_PTR(str); + e = RSTRING_END(str) - 1; + while (s < e) { + c = *s; + *s++ = *e; + *e-- = c; + } + } + else { + str_shared_replace(str, rb_str_reverse(str)); + } } else { - str_modify_keep_cr(str); + str_modify_keep_cr(str); } return str; } @@ -6365,7 +6578,7 @@ rb_str_reverse_bang(VALUE str) /* * call-seq: - * include? other_string -> true or false + * include?(other_string) -> true or false * * Returns +true+ if +self+ contains +other_string+, +false+ otherwise: * @@ -6376,7 +6589,7 @@ rb_str_reverse_bang(VALUE str) * */ -static VALUE +VALUE rb_str_include(VALUE str, VALUE arg) { long i; @@ -6393,11 +6606,21 @@ rb_str_include(VALUE str, VALUE arg) * to_i(base = 10) -> integer * * Returns the result of interpreting leading characters in +self+ - * as an integer in the given +base+ (which must be in (2..36)): + * as an integer in the given +base+ (which must be in (0, 2..36)): * * '123456'.to_i # => 123456 * '123def'.to_i(16) # => 1195503 * + * With +base+ zero, string +object+ may contain leading characters + * to specify the actual base: + * + * '123def'.to_i(0) # => 123 + * '0123def'.to_i(0) # => 83 + * '0b123def'.to_i(0) # => 1 + * '0o123def'.to_i(0) # => 83 + * '0d123def'.to_i(0) # => 123 + * '0x123def'.to_i(0) # => 1195503 + * * Characters past a leading valid number (in the given +base+) are ignored: * * '12.345'.to_i # => 12 @@ -6416,7 +6639,7 @@ rb_str_to_i(int argc, VALUE *argv, VALUE str) int base = 10; if (rb_check_arity(argc, 0, 1) && (base = NUM2INT(argv[0])) < 0) { - rb_raise(rb_eArgError, "invalid radix %d", base); + rb_raise(rb_eArgError, "invalid radix %d", base); } return rb_str_to_inum(str, base, FALSE); } @@ -6429,7 +6652,7 @@ rb_str_to_i(int argc, VALUE *argv, VALUE str) * Returns the result of interpreting leading characters in +self+ as a Float: * * '3.14159'.to_f # => 3.14159 - '1.234e-2'.to_f # => 0.01234 + * '1.234e-2'.to_f # => 0.01234 * * Characters past a leading valid number (in the given +base+) are ignored: * @@ -6452,18 +6675,15 @@ rb_str_to_f(VALUE str) * call-seq: * to_s -> self or string * - * Returns +self+ if +self+ is a \String, - * or +self+ converted to a \String if +self+ is a subclass of \String. - * - * String#to_str is an alias for String#to_s. - * + * Returns +self+ if +self+ is a +String+, + * or +self+ converted to a +String+ if +self+ is a subclass of +String+. */ static VALUE rb_str_to_s(VALUE str) { if (rb_obj_class(str) != rb_cString) { - return str_duplicate(rb_cString, str); + return str_duplicate(rb_cString, str); } return str; } @@ -6492,23 +6712,23 @@ rb_str_buf_cat_escaped_char(VALUE result, unsigned int c, int unicode_p) c &= 0xffffffff; #endif if (unicode_p) { - if (c < 0x7F && ISPRINT(c)) { - snprintf(buf, CHAR_ESC_LEN, "%c", c); - } - else if (c < 0x10000) { - snprintf(buf, CHAR_ESC_LEN, "\\u%04X", c); - } - else { - snprintf(buf, CHAR_ESC_LEN, "\\u{%X}", c); - } + if (c < 0x7F && ISPRINT(c)) { + snprintf(buf, CHAR_ESC_LEN, "%c", c); + } + else if (c < 0x10000) { + snprintf(buf, CHAR_ESC_LEN, "\\u%04X", c); + } + else { + snprintf(buf, CHAR_ESC_LEN, "\\u{%X}", c); + } } else { - if (c < 0x100) { - snprintf(buf, CHAR_ESC_LEN, "\\x%02X", c); - } - else { - snprintf(buf, CHAR_ESC_LEN, "\\x{%X}", c); - } + if (c < 0x100) { + snprintf(buf, CHAR_ESC_LEN, "\\x%02X", c); + } + else { + snprintf(buf, CHAR_ESC_LEN, "\\x{%X}", c); + } } l = (int)strlen(buf); /* CHAR_ESC_LEN cannot exceed INT_MAX */ rb_str_buf_cat(result, buf, l); @@ -6549,9 +6769,9 @@ rb_str_escape(VALUE str) while (p < pend) { unsigned int c; const char *cc; - int n = rb_enc_precise_mbclen(p, pend, enc); + int n = rb_enc_precise_mbclen(p, pend, enc); if (!MBCLEN_CHARFOUND_P(n)) { - if (p > prev) str_buf_cat(result, prev, p - prev); + if (p > prev) str_buf_cat(result, prev, p - prev); n = rb_enc_mbminlen(enc); if (pend < p + n) n = (int)(pend - p); @@ -6560,24 +6780,24 @@ rb_str_escape(VALUE str) str_buf_cat(result, buf, strlen(buf)); prev = ++p; } - continue; - } + continue; + } n = MBCLEN_CHARFOUND_LEN(n); - c = rb_enc_mbc_to_codepoint(p, pend, enc); - p += n; + c = rb_enc_mbc_to_codepoint(p, pend, enc); + p += n; cc = ruby_escaped_char(c); - if (cc) { - if (p - n > prev) str_buf_cat(result, prev, p - n - prev); + if (cc) { + if (p - n > prev) str_buf_cat(result, prev, p - n - prev); str_buf_cat(result, cc, strlen(cc)); - prev = p; - } - else if (asciicompat && rb_enc_isascii(c, enc) && ISPRINT(c)) { - } - else { - if (p - n > prev) str_buf_cat(result, prev, p - n - prev); - rb_str_buf_cat_escaped_char(result, c, unicode_p); - prev = p; - } + prev = p; + } + else if (asciicompat && rb_enc_isascii(c, enc) && ISPRINT(c)) { + } + else { + if (p - n > prev) str_buf_cat(result, prev, p - n - prev); + rb_str_buf_cat_escaped_char(result, c, unicode_p); + prev = p; + } } if (p > prev) str_buf_cat(result, prev, p - prev); ENCODING_CODERANGE_SET(result, rb_usascii_encindex(), ENC_CODERANGE_7BIT); @@ -6602,7 +6822,7 @@ VALUE rb_str_inspect(VALUE str) { int encidx = ENCODING_GET(str); - rb_encoding *enc = rb_enc_from_index(encidx), *actenc; + rb_encoding *enc = rb_enc_from_index(encidx); const char *p, *pend, *prev; char buf[CHAR_ESC_LEN + 1]; VALUE result = rb_str_buf_new(0); @@ -6617,18 +6837,13 @@ rb_str_inspect(VALUE str) p = RSTRING_PTR(str); pend = RSTRING_END(str); prev = p; - actenc = get_actual_encoding(encidx, str); - if (actenc != enc) { - enc = actenc; - if (unicode_p) unicode_p = rb_enc_unicode_p(enc); - } while (p < pend) { - unsigned int c, cc; - int n; + unsigned int c, cc; + int n; n = rb_enc_precise_mbclen(p, pend, enc); if (!MBCLEN_CHARFOUND_P(n)) { - if (p > prev) str_buf_cat(result, prev, p - prev); + if (p > prev) str_buf_cat(result, prev, p - prev); n = rb_enc_mbminlen(enc); if (pend < p + n) n = (int)(pend - p); @@ -6637,54 +6852,62 @@ rb_str_inspect(VALUE str) str_buf_cat(result, buf, strlen(buf)); prev = ++p; } - continue; - } + continue; + } n = MBCLEN_CHARFOUND_LEN(n); - c = rb_enc_mbc_to_codepoint(p, pend, enc); - p += n; - if ((asciicompat || unicode_p) && - (c == '"'|| c == '\\' || - (c == '#' && + c = rb_enc_mbc_to_codepoint(p, pend, enc); + p += n; + if ((asciicompat || unicode_p) && + (c == '"'|| c == '\\' || + (c == '#' && p < pend && MBCLEN_CHARFOUND_P(rb_enc_precise_mbclen(p,pend,enc)) && (cc = rb_enc_codepoint(p,pend,enc), (cc == '$' || cc == '@' || cc == '{'))))) { - if (p - n > prev) str_buf_cat(result, prev, p - n - prev); - str_buf_cat2(result, "\\"); - if (asciicompat || enc == resenc) { - prev = p - n; - continue; - } - } - switch (c) { - case '\n': cc = 'n'; break; - case '\r': cc = 'r'; break; - case '\t': cc = 't'; break; - case '\f': cc = 'f'; break; - case '\013': cc = 'v'; break; - case '\010': cc = 'b'; break; - case '\007': cc = 'a'; break; - case 033: cc = 'e'; break; - default: cc = 0; break; - } - if (cc) { - if (p - n > prev) str_buf_cat(result, prev, p - n - prev); - buf[0] = '\\'; - buf[1] = (char)cc; - str_buf_cat(result, buf, 2); - prev = p; - continue; - } - if ((enc == resenc && rb_enc_isprint(c, enc)) || - (asciicompat && rb_enc_isascii(c, enc) && ISPRINT(c))) { - continue; - } - else { - if (p - n > prev) str_buf_cat(result, prev, p - n - prev); - rb_str_buf_cat_escaped_char(result, c, unicode_p); - prev = p; - continue; - } + if (p - n > prev) str_buf_cat(result, prev, p - n - prev); + str_buf_cat2(result, "\\"); + if (asciicompat || enc == resenc) { + prev = p - n; + continue; + } + } + switch (c) { + case '\n': cc = 'n'; break; + case '\r': cc = 'r'; break; + case '\t': cc = 't'; break; + case '\f': cc = 'f'; break; + case '\013': cc = 'v'; break; + case '\010': cc = 'b'; break; + case '\007': cc = 'a'; break; + case 033: cc = 'e'; break; + default: cc = 0; break; + } + if (cc) { + if (p - n > prev) str_buf_cat(result, prev, p - n - prev); + buf[0] = '\\'; + buf[1] = (char)cc; + str_buf_cat(result, buf, 2); + prev = p; + continue; + } + /* The special casing of 0x85 (NEXT_LINE) here is because + * Oniguruma historically treats it as printable, but it + * doesn't match the print POSIX bracket class or character + * property in regexps. + * + * See Ruby Bug #16842 for details: + * https://bugs.ruby-lang.org/issues/16842 + */ + if ((enc == resenc && rb_enc_isprint(c, enc) && c != 0x85) || + (asciicompat && rb_enc_isascii(c, enc) && ISPRINT(c))) { + continue; + } + else { + if (p - n > prev) str_buf_cat(result, prev, p - n - prev); + rb_str_buf_cat_escaped_char(result, c, unicode_p); + prev = p; + continue; + } } if (p > prev) str_buf_cat(result, prev, p - prev); str_buf_cat2(result, "\""); @@ -6723,55 +6946,55 @@ rb_str_dump(VALUE str) len = 2; /* "" */ if (!rb_enc_asciicompat(enc)) { - len += strlen(nonascii_suffix) - rb_strlen_lit("%s"); - len += strlen(enc->name); + len += strlen(nonascii_suffix) - rb_strlen_lit("%s"); + len += strlen(enc->name); } p = RSTRING_PTR(str); pend = p + RSTRING_LEN(str); while (p < pend) { - int clen; - unsigned char c = *p++; - - switch (c) { - case '"': case '\\': - case '\n': case '\r': - case '\t': case '\f': - case '\013': case '\010': case '\007': case '\033': - clen = 2; - break; - - case '#': - clen = IS_EVSTR(p, pend) ? 2 : 1; - break; - - default: - if (ISPRINT(c)) { - clen = 1; - } - else { - if (u8 && c > 0x7F) { /* \u notation */ - int n = rb_enc_precise_mbclen(p-1, pend, enc); - if (MBCLEN_CHARFOUND_P(n)) { - unsigned int cc = rb_enc_mbc_to_codepoint(p-1, pend, enc); - if (cc <= 0xFFFF) - clen = 6; /* \uXXXX */ - else if (cc <= 0xFFFFF) - clen = 9; /* \u{XXXXX} */ - else - clen = 10; /* \u{XXXXXX} */ - p += MBCLEN_CHARFOUND_LEN(n)-1; - break; - } - } - clen = 4; /* \xNN */ - } - break; - } - - if (clen > LONG_MAX - len) { - rb_raise(rb_eRuntimeError, "string size too big"); - } - len += clen; + int clen; + unsigned char c = *p++; + + switch (c) { + case '"': case '\\': + case '\n': case '\r': + case '\t': case '\f': + case '\013': case '\010': case '\007': case '\033': + clen = 2; + break; + + case '#': + clen = IS_EVSTR(p, pend) ? 2 : 1; + break; + + default: + if (ISPRINT(c)) { + clen = 1; + } + else { + if (u8 && c > 0x7F) { /* \u notation */ + int n = rb_enc_precise_mbclen(p-1, pend, enc); + if (MBCLEN_CHARFOUND_P(n)) { + unsigned int cc = rb_enc_mbc_to_codepoint(p-1, pend, enc); + if (cc <= 0xFFFF) + clen = 6; /* \uXXXX */ + else if (cc <= 0xFFFFF) + clen = 9; /* \u{XXXXX} */ + else + clen = 10; /* \u{XXXXXX} */ + p += MBCLEN_CHARFOUND_LEN(n)-1; + break; + } + } + clen = 4; /* \xNN */ + } + break; + } + + if (clen > LONG_MAX - len) { + rb_raise(rb_eRuntimeError, "string size too big"); + } + len += clen; } result = rb_str_new(0, len); @@ -6780,75 +7003,75 @@ rb_str_dump(VALUE str) *q++ = '"'; while (p < pend) { - unsigned char c = *p++; - - if (c == '"' || c == '\\') { - *q++ = '\\'; - *q++ = c; - } - else if (c == '#') { - if (IS_EVSTR(p, pend)) *q++ = '\\'; - *q++ = '#'; - } - else if (c == '\n') { - *q++ = '\\'; - *q++ = 'n'; - } - else if (c == '\r') { - *q++ = '\\'; - *q++ = 'r'; - } - else if (c == '\t') { - *q++ = '\\'; - *q++ = 't'; - } - else if (c == '\f') { - *q++ = '\\'; - *q++ = 'f'; - } - else if (c == '\013') { - *q++ = '\\'; - *q++ = 'v'; - } - else if (c == '\010') { - *q++ = '\\'; - *q++ = 'b'; - } - else if (c == '\007') { - *q++ = '\\'; - *q++ = 'a'; - } - else if (c == '\033') { - *q++ = '\\'; - *q++ = 'e'; - } - else if (ISPRINT(c)) { - *q++ = c; - } - else { - *q++ = '\\'; - if (u8) { - int n = rb_enc_precise_mbclen(p-1, pend, enc) - 1; - if (MBCLEN_CHARFOUND_P(n)) { - int cc = rb_enc_mbc_to_codepoint(p-1, pend, enc); - p += n; - if (cc <= 0xFFFF) - snprintf(q, qend-q, "u%04X", cc); /* \uXXXX */ - else - snprintf(q, qend-q, "u{%X}", cc); /* \u{XXXXX} or \u{XXXXXX} */ - q += strlen(q); - continue; - } - } - snprintf(q, qend-q, "x%02X", c); - q += 3; - } + unsigned char c = *p++; + + if (c == '"' || c == '\\') { + *q++ = '\\'; + *q++ = c; + } + else if (c == '#') { + if (IS_EVSTR(p, pend)) *q++ = '\\'; + *q++ = '#'; + } + else if (c == '\n') { + *q++ = '\\'; + *q++ = 'n'; + } + else if (c == '\r') { + *q++ = '\\'; + *q++ = 'r'; + } + else if (c == '\t') { + *q++ = '\\'; + *q++ = 't'; + } + else if (c == '\f') { + *q++ = '\\'; + *q++ = 'f'; + } + else if (c == '\013') { + *q++ = '\\'; + *q++ = 'v'; + } + else if (c == '\010') { + *q++ = '\\'; + *q++ = 'b'; + } + else if (c == '\007') { + *q++ = '\\'; + *q++ = 'a'; + } + else if (c == '\033') { + *q++ = '\\'; + *q++ = 'e'; + } + else if (ISPRINT(c)) { + *q++ = c; + } + else { + *q++ = '\\'; + if (u8) { + int n = rb_enc_precise_mbclen(p-1, pend, enc) - 1; + if (MBCLEN_CHARFOUND_P(n)) { + int cc = rb_enc_mbc_to_codepoint(p-1, pend, enc); + p += n; + if (cc <= 0xFFFF) + snprintf(q, qend-q, "u%04X", cc); /* \uXXXX */ + else + snprintf(q, qend-q, "u{%X}", cc); /* \u{XXXXX} or \u{XXXXXX} */ + q += strlen(q); + continue; + } + } + snprintf(q, qend-q, "x%02X", c); + q += 3; + } } *q++ = '"'; *q = '\0'; if (!rb_enc_asciicompat(enc)) { - snprintf(q, qend-q, nonascii_suffix, enc->name); - encidx = rb_ascii8bit_encindex(); + snprintf(q, qend-q, nonascii_suffix, enc->name); + encidx = rb_ascii8bit_encindex(); } /* result from dump is ASCII */ rb_enc_associate_index(result, encidx); @@ -6861,21 +7084,21 @@ unescape_ascii(unsigned int c) { switch (c) { case 'n': - return '\n'; + return '\n'; case 'r': - return '\r'; + return '\r'; case 't': - return '\t'; + return '\t'; case 'f': - return '\f'; + return '\f'; case 'v': - return '\13'; + return '\13'; case 'b': - return '\010'; + return '\010'; case 'a': - return '\007'; + return '\007'; case 'e': - return 033; + return 033; } UNREACHABLE_RETURN(-1); } @@ -6894,9 +7117,9 @@ undump_after_backslash(VALUE undumped, const char **ss, const char *s_end, rb_en case '\\': case '"': case '#': - rb_str_cat(undumped, s, 1); /* cat itself */ - s++; - break; + rb_str_cat(undumped, s, 1); /* cat itself */ + s++; + break; case 'n': case 'r': case 't': @@ -6907,81 +7130,81 @@ undump_after_backslash(VALUE undumped, const char **ss, const char *s_end, rb_en case 'e': *buf = unescape_ascii(*s); rb_str_cat(undumped, (char *)buf, 1); - s++; - break; + s++; + break; case 'u': - if (*binary) { - rb_raise(rb_eRuntimeError, "hex escape and Unicode escape are mixed"); - } - *utf8 = true; - if (++s >= s_end) { - rb_raise(rb_eRuntimeError, "invalid Unicode escape"); - } - if (enc_utf8 == NULL) enc_utf8 = rb_utf8_encoding(); - if (*penc != enc_utf8) { - *penc = enc_utf8; - rb_enc_associate(undumped, enc_utf8); - } - if (*s == '{') { /* handle \u{...} form */ - s++; - for (;;) { - if (s >= s_end) { - rb_raise(rb_eRuntimeError, "unterminated Unicode escape"); - } - if (*s == '}') { - s++; - break; - } - if (ISSPACE(*s)) { - s++; - continue; - } - c = scan_hex(s, s_end-s, &hexlen); - if (hexlen == 0 || hexlen > 6) { - rb_raise(rb_eRuntimeError, "invalid Unicode escape"); - } - if (c > 0x10ffff) { - rb_raise(rb_eRuntimeError, "invalid Unicode codepoint (too large)"); - } - if (0xd800 <= c && c <= 0xdfff) { - rb_raise(rb_eRuntimeError, "invalid Unicode codepoint"); - } + if (*binary) { + rb_raise(rb_eRuntimeError, "hex escape and Unicode escape are mixed"); + } + *utf8 = true; + if (++s >= s_end) { + rb_raise(rb_eRuntimeError, "invalid Unicode escape"); + } + if (enc_utf8 == NULL) enc_utf8 = rb_utf8_encoding(); + if (*penc != enc_utf8) { + *penc = enc_utf8; + rb_enc_associate(undumped, enc_utf8); + } + if (*s == '{') { /* handle \u{...} form */ + s++; + for (;;) { + if (s >= s_end) { + rb_raise(rb_eRuntimeError, "unterminated Unicode escape"); + } + if (*s == '}') { + s++; + break; + } + if (ISSPACE(*s)) { + s++; + continue; + } + c = scan_hex(s, s_end-s, &hexlen); + if (hexlen == 0 || hexlen > 6) { + rb_raise(rb_eRuntimeError, "invalid Unicode escape"); + } + if (c > 0x10ffff) { + rb_raise(rb_eRuntimeError, "invalid Unicode codepoint (too large)"); + } + if (0xd800 <= c && c <= 0xdfff) { + rb_raise(rb_eRuntimeError, "invalid Unicode codepoint"); + } codelen = rb_enc_mbcput(c, (char *)buf, *penc); rb_str_cat(undumped, (char *)buf, codelen); - s += hexlen; - } - } - else { /* handle \uXXXX form */ - c = scan_hex(s, 4, &hexlen); - if (hexlen != 4) { - rb_raise(rb_eRuntimeError, "invalid Unicode escape"); - } - if (0xd800 <= c && c <= 0xdfff) { - rb_raise(rb_eRuntimeError, "invalid Unicode codepoint"); - } + s += hexlen; + } + } + else { /* handle \uXXXX form */ + c = scan_hex(s, 4, &hexlen); + if (hexlen != 4) { + rb_raise(rb_eRuntimeError, "invalid Unicode escape"); + } + if (0xd800 <= c && c <= 0xdfff) { + rb_raise(rb_eRuntimeError, "invalid Unicode codepoint"); + } codelen = rb_enc_mbcput(c, (char *)buf, *penc); rb_str_cat(undumped, (char *)buf, codelen); - s += hexlen; - } - break; + s += hexlen; + } + break; case 'x': - if (*utf8) { - rb_raise(rb_eRuntimeError, "hex escape and Unicode escape are mixed"); - } - *binary = true; - if (++s >= s_end) { - rb_raise(rb_eRuntimeError, "invalid hex escape"); - } - *buf = scan_hex(s, 2, &hexlen); - if (hexlen != 2) { - rb_raise(rb_eRuntimeError, "invalid hex escape"); - } + if (*utf8) { + rb_raise(rb_eRuntimeError, "hex escape and Unicode escape are mixed"); + } + *binary = true; + if (++s >= s_end) { + rb_raise(rb_eRuntimeError, "invalid hex escape"); + } + *buf = scan_hex(s, 2, &hexlen); + if (hexlen != 2) { + rb_raise(rb_eRuntimeError, "invalid hex escape"); + } rb_str_cat(undumped, (char *)buf, 1); - s += hexlen; - break; + s += hexlen; + break; default: - rb_str_cat(undumped, s-1, 2); - s++; + rb_str_cat(undumped, s-1, 2); + s++; } *ss = s; @@ -7017,10 +7240,10 @@ str_undump(VALUE str) rb_must_asciicompat(str); if (rb_str_is_ascii_only_p(str) == Qfalse) { - rb_raise(rb_eRuntimeError, "non-ASCII character detected"); + rb_raise(rb_eRuntimeError, "non-ASCII character detected"); } if (!str_null_check(str, &w)) { - rb_raise(rb_eRuntimeError, "string contains null byte"); + rb_raise(rb_eRuntimeError, "string contains null byte"); } if (RSTRING_LEN(str) < 2) goto invalid_format; if (*s != '"') goto invalid_format; @@ -7029,65 +7252,67 @@ str_undump(VALUE str) s++; for (;;) { - if (s >= s_end) { - rb_raise(rb_eRuntimeError, "unterminated dumped string"); - } - - if (*s == '"') { - /* epilogue */ - s++; - if (s == s_end) { - /* ascii compatible dumped string */ - break; - } - else { - static const char force_encoding_suffix[] = ".force_encoding(\""; /* "\")" */ - static const char dup_suffix[] = ".dup"; - const char *encname; - int encidx; - ptrdiff_t size; - - /* check separately for strings dumped by older versions */ - size = sizeof(dup_suffix) - 1; - if (s_end - s > size && memcmp(s, dup_suffix, size) == 0) s += size; - - size = sizeof(force_encoding_suffix) - 1; - if (s_end - s <= size) goto invalid_format; - if (memcmp(s, force_encoding_suffix, size) != 0) goto invalid_format; - s += size; - - if (utf8) { - rb_raise(rb_eRuntimeError, "dumped string contained Unicode escape but used force_encoding"); - } - - encname = s; - s = memchr(s, '"', s_end-s); - size = s - encname; - if (!s) goto invalid_format; - if (s_end - s != 2) goto invalid_format; - if (s[0] != '"' || s[1] != ')') goto invalid_format; - - encidx = rb_enc_find_index2(encname, (long)size); - if (encidx < 0) { - rb_raise(rb_eRuntimeError, "dumped string has unknown encoding name"); - } - rb_enc_associate_index(undumped, encidx); - } - break; - } - - if (*s == '\\') { - s++; - if (s >= s_end) { - rb_raise(rb_eRuntimeError, "invalid escape"); - } - undump_after_backslash(undumped, &s, s_end, &enc, &utf8, &binary); - } - else { - rb_str_cat(undumped, s++, 1); - } + if (s >= s_end) { + rb_raise(rb_eRuntimeError, "unterminated dumped string"); + } + + if (*s == '"') { + /* epilogue */ + s++; + if (s == s_end) { + /* ascii compatible dumped string */ + break; + } + else { + static const char force_encoding_suffix[] = ".force_encoding(\""; /* "\")" */ + static const char dup_suffix[] = ".dup"; + const char *encname; + int encidx; + ptrdiff_t size; + + /* check separately for strings dumped by older versions */ + size = sizeof(dup_suffix) - 1; + if (s_end - s > size && memcmp(s, dup_suffix, size) == 0) s += size; + + size = sizeof(force_encoding_suffix) - 1; + if (s_end - s <= size) goto invalid_format; + if (memcmp(s, force_encoding_suffix, size) != 0) goto invalid_format; + s += size; + + if (utf8) { + rb_raise(rb_eRuntimeError, "dumped string contained Unicode escape but used force_encoding"); + } + + encname = s; + s = memchr(s, '"', s_end-s); + size = s - encname; + if (!s) goto invalid_format; + if (s_end - s != 2) goto invalid_format; + if (s[0] != '"' || s[1] != ')') goto invalid_format; + + encidx = rb_enc_find_index2(encname, (long)size); + if (encidx < 0) { + rb_raise(rb_eRuntimeError, "dumped string has unknown encoding name"); + } + rb_enc_associate_index(undumped, encidx); + } + break; + } + + if (*s == '\\') { + s++; + if (s >= s_end) { + rb_raise(rb_eRuntimeError, "invalid escape"); + } + undump_after_backslash(undumped, &s, s_end, &enc, &utf8, &binary); + } + else { + rb_str_cat(undumped, s++, 1); + } } + RB_GC_GUARD(str); + return undumped; invalid_format: rb_raise(rb_eRuntimeError, "invalid dumped string; not wrapped with '\"' nor '\"...\".force_encoding(\"...\")' form"); @@ -7097,8 +7322,8 @@ static void rb_str_check_dummy_enc(rb_encoding *enc) { if (rb_enc_dummy_p(enc)) { - rb_raise(rb_eEncCompatError, "incompatible encoding with this operation: %s", - rb_enc_name(enc)); + rb_raise(rb_eEncCompatError, "incompatible encoding with this operation: %s", + rb_enc_name(enc)); } } @@ -7118,35 +7343,35 @@ check_case_options(int argc, VALUE *argv, OnigCaseFoldType flags) if (argc>2) rb_raise(rb_eArgError, "too many options"); if (argv[0]==sym_turkic) { - flags |= ONIGENC_CASE_FOLD_TURKISH_AZERI; - if (argc==2) { - if (argv[1]==sym_lithuanian) - flags |= ONIGENC_CASE_FOLD_LITHUANIAN; - else - rb_raise(rb_eArgError, "invalid second option"); - } + flags |= ONIGENC_CASE_FOLD_TURKISH_AZERI; + if (argc==2) { + if (argv[1]==sym_lithuanian) + flags |= ONIGENC_CASE_FOLD_LITHUANIAN; + else + rb_raise(rb_eArgError, "invalid second option"); + } } else if (argv[0]==sym_lithuanian) { - flags |= ONIGENC_CASE_FOLD_LITHUANIAN; - if (argc==2) { - if (argv[1]==sym_turkic) - flags |= ONIGENC_CASE_FOLD_TURKISH_AZERI; - else - rb_raise(rb_eArgError, "invalid second option"); - } + flags |= ONIGENC_CASE_FOLD_LITHUANIAN; + if (argc==2) { + if (argv[1]==sym_turkic) + flags |= ONIGENC_CASE_FOLD_TURKISH_AZERI; + else + rb_raise(rb_eArgError, "invalid second option"); + } } else if (argc>1) - rb_raise(rb_eArgError, "too many options"); + rb_raise(rb_eArgError, "too many options"); else if (argv[0]==sym_ascii) - flags |= ONIGENC_CASE_ASCII_ONLY; + flags |= ONIGENC_CASE_ASCII_ONLY; else if (argv[0]==sym_fold) { - if ((flags & (ONIGENC_CASE_UPCASE|ONIGENC_CASE_DOWNCASE)) == ONIGENC_CASE_DOWNCASE) - flags ^= ONIGENC_CASE_FOLD|ONIGENC_CASE_DOWNCASE; - else - rb_raise(rb_eArgError, "option :fold only allowed for downcasing"); + if ((flags & (ONIGENC_CASE_UPCASE|ONIGENC_CASE_DOWNCASE)) == ONIGENC_CASE_DOWNCASE) + flags ^= ONIGENC_CASE_FOLD|ONIGENC_CASE_DOWNCASE; + else + rb_raise(rb_eArgError, "option :fold only allowed for downcasing"); } else - rb_raise(rb_eArgError, "invalid option"); + rb_raise(rb_eArgError, "invalid option"); return flags; } @@ -7186,7 +7411,8 @@ mapping_buffer_free(void *p) static const rb_data_type_t mapping_buffer_type = { "mapping_buffer", - {0, mapping_buffer_free,} + {0, mapping_buffer_free,}, + 0, 0, RUBY_TYPED_FREE_IMMEDIATELY | RUBY_TYPED_WB_PROTECTED }; static VALUE @@ -7210,54 +7436,56 @@ rb_str_casemap(VALUE source, OnigCaseFoldType *flags, rb_encoding *enc) buffer_anchor = TypedData_Wrap_Struct(0, &mapping_buffer_type, 0); pre_buffer = (mapping_buffer **)&DATA_PTR(buffer_anchor); while (source_current < source_end) { - /* increase multiplier using buffer count to converge quickly */ - size_t capa = (size_t)(source_end-source_current)*++buffer_count + CASE_MAPPING_ADDITIONAL_LENGTH; - if (CASEMAP_DEBUG) { - fprintf(stderr, "Buffer allocation, capa is %"PRIuSIZE"\n", capa); /* for tuning */ - } + /* increase multiplier using buffer count to converge quickly */ + size_t capa = (size_t)(source_end-source_current)*++buffer_count + CASE_MAPPING_ADDITIONAL_LENGTH; + if (CASEMAP_DEBUG) { + fprintf(stderr, "Buffer allocation, capa is %"PRIuSIZE"\n", capa); /* for tuning */ + } current_buffer = xmalloc(offsetof(mapping_buffer, space) + capa); *pre_buffer = current_buffer; pre_buffer = ¤t_buffer->next; - current_buffer->next = NULL; - current_buffer->capa = capa; - buffer_length_or_invalid = enc->case_map(flags, - &source_current, source_end, - current_buffer->space, - current_buffer->space+current_buffer->capa, - enc); - if (buffer_length_or_invalid < 0) { + current_buffer->next = NULL; + current_buffer->capa = capa; + buffer_length_or_invalid = enc->case_map(flags, + &source_current, source_end, + current_buffer->space, + current_buffer->space+current_buffer->capa, + enc); + if (buffer_length_or_invalid < 0) { current_buffer = DATA_PTR(buffer_anchor); DATA_PTR(buffer_anchor) = 0; mapping_buffer_free(current_buffer); - rb_raise(rb_eArgError, "input string invalid"); - } - target_length += current_buffer->used = buffer_length_or_invalid; + rb_raise(rb_eArgError, "input string invalid"); + } + target_length += current_buffer->used = buffer_length_or_invalid; } if (CASEMAP_DEBUG) { - fprintf(stderr, "Buffer count is %"PRIuSIZE"\n", buffer_count); /* for tuning */ + fprintf(stderr, "Buffer count is %"PRIuSIZE"\n", buffer_count); /* for tuning */ } if (buffer_count==1) { target = rb_str_new((const char*)current_buffer->space, target_length); } else { - char *target_current; + char *target_current; target = rb_str_new(0, target_length); - target_current = RSTRING_PTR(target); + target_current = RSTRING_PTR(target); current_buffer = DATA_PTR(buffer_anchor); - while (current_buffer) { - memcpy(target_current, current_buffer->space, current_buffer->used); - target_current += current_buffer->used; - current_buffer = current_buffer->next; - } + while (current_buffer) { + memcpy(target_current, current_buffer->space, current_buffer->used); + target_current += current_buffer->used; + current_buffer = current_buffer->next; + } } current_buffer = DATA_PTR(buffer_anchor); DATA_PTR(buffer_anchor) = 0; mapping_buffer_free(current_buffer); + RB_GC_GUARD(buffer_anchor); + /* TODO: check about string terminator character */ - str_enc_copy(target, source); + str_enc_copy_direct(target, source); /*ENC_CODERANGE_SET(mapped, cr);*/ return target; @@ -7290,10 +7518,10 @@ rb_str_ascii_casemap(VALUE source, VALUE target, OnigCaseFoldType *flags, rb_enc if (length_or_invalid < 0) rb_raise(rb_eArgError, "input string invalid"); if (CASEMAP_DEBUG && length_or_invalid != old_length) { - fprintf(stderr, "problem with rb_str_ascii_casemap" - "; old_length=%ld, new_length=%d\n", old_length, length_or_invalid); - rb_raise(rb_eArgError, "internal problem with rb_str_ascii_casemap" - "; old_length=%ld, new_length=%d\n", old_length, length_or_invalid); + fprintf(stderr, "problem with rb_str_ascii_casemap" + "; old_length=%ld, new_length=%d\n", old_length, length_or_invalid); + rb_raise(rb_eArgError, "internal problem with rb_str_ascii_casemap" + "; old_length=%ld, new_length=%d\n", old_length, length_or_invalid); } str_enc_copy(target, source); @@ -7354,7 +7582,7 @@ rb_str_upcase_bang(int argc, VALUE *argv, VALUE str) else if (flags&ONIGENC_CASE_ASCII_ONLY) rb_str_ascii_casemap(str, str, &flags, enc); else - str_shared_replace(str, rb_str_casemap(str, &flags, enc)); + str_shared_replace(str, rb_str_casemap(str, &flags, enc)); if (ONIGENC_CASE_MODIFIED&flags) return str; return Qnil; @@ -7388,7 +7616,7 @@ rb_str_upcase(int argc, VALUE *argv, VALUE str) enc = str_true_enc(str); if (case_option_single_p(flags, enc, str)) { ret = rb_str_new(RSTRING_PTR(str), RSTRING_LEN(str)); - str_enc_copy(ret, str); + str_enc_copy_direct(ret, str); upcase_single(ret); } else if (flags&ONIGENC_CASE_ASCII_ONLY) { @@ -7456,7 +7684,7 @@ rb_str_downcase_bang(int argc, VALUE *argv, VALUE str) else if (flags&ONIGENC_CASE_ASCII_ONLY) rb_str_ascii_casemap(str, str, &flags, enc); else - str_shared_replace(str, rb_str_casemap(str, &flags, enc)); + str_shared_replace(str, rb_str_casemap(str, &flags, enc)); if (ONIGENC_CASE_MODIFIED&flags) return str; return Qnil; @@ -7490,7 +7718,7 @@ rb_str_downcase(int argc, VALUE *argv, VALUE str) enc = str_true_enc(str); if (case_option_single_p(flags, enc, str)) { ret = rb_str_new(RSTRING_PTR(str), RSTRING_LEN(str)); - str_enc_copy(ret, str); + str_enc_copy_direct(ret, str); downcase_single(ret); } else if (flags&ONIGENC_CASE_ASCII_ONLY) { @@ -7538,7 +7766,7 @@ rb_str_capitalize_bang(int argc, VALUE *argv, VALUE str) if (flags&ONIGENC_CASE_ASCII_ONLY) rb_str_ascii_casemap(str, str, &flags, enc); else - str_shared_replace(str, rb_str_casemap(str, &flags, enc)); + str_shared_replace(str, rb_str_casemap(str, &flags, enc)); if (ONIGENC_CASE_MODIFIED&flags) return str; return Qnil; @@ -7594,7 +7822,7 @@ rb_str_capitalize(int argc, VALUE *argv, VALUE str) * * s = 'Hello World!' # => "Hello World!" * s.swapcase! # => "hELLO wORLD!" - * s # => "Hello World!" + * s # => "hELLO wORLD!" * ''.swapcase! # => nil * * The casing may be affected by the given +options+; @@ -7616,7 +7844,7 @@ rb_str_swapcase_bang(int argc, VALUE *argv, VALUE str) if (flags&ONIGENC_CASE_ASCII_ONLY) rb_str_ascii_casemap(str, str, &flags, enc); else - str_shared_replace(str, rb_str_casemap(str, &flags, enc)); + str_shared_replace(str, rb_str_casemap(str, &flags, enc)); if (ONIGENC_CASE_MODIFIED&flags) return str; return Qnil; @@ -7676,50 +7904,52 @@ trnext(struct tr *t, rb_encoding *enc) for (;;) { nextpart: - if (!t->gen) { - if (t->p == t->pend) return -1; - if (rb_enc_ascget(t->p, t->pend, &n, enc) == '\\' && t->p + n < t->pend) { - t->p += n; - } - t->now = rb_enc_codepoint_len(t->p, t->pend, &n, enc); - t->p += n; - if (rb_enc_ascget(t->p, t->pend, &n, enc) == '-' && t->p + n < t->pend) { - t->p += n; - if (t->p < t->pend) { - unsigned int c = rb_enc_codepoint_len(t->p, t->pend, &n, enc); - t->p += n; - if (t->now > c) { - if (t->now < 0x80 && c < 0x80) { - rb_raise(rb_eArgError, - "invalid range \"%c-%c\" in string transliteration", - t->now, c); - } - else { - rb_raise(rb_eArgError, "invalid range in string transliteration"); - } - continue; /* not reached */ - } - t->gen = 1; - t->max = c; - } - } - return t->now; - } - else { - while (ONIGENC_CODE_TO_MBCLEN(enc, ++t->now) <= 0) { - if (t->now == t->max) { - t->gen = 0; - goto nextpart; - } - } - if (t->now < t->max) { - return t->now; - } - else { - t->gen = 0; - return t->max; - } - } + if (!t->gen) { + if (t->p == t->pend) return -1; + if (rb_enc_ascget(t->p, t->pend, &n, enc) == '\\' && t->p + n < t->pend) { + t->p += n; + } + t->now = rb_enc_codepoint_len(t->p, t->pend, &n, enc); + t->p += n; + if (rb_enc_ascget(t->p, t->pend, &n, enc) == '-' && t->p + n < t->pend) { + t->p += n; + if (t->p < t->pend) { + unsigned int c = rb_enc_codepoint_len(t->p, t->pend, &n, enc); + t->p += n; + if (t->now > c) { + if (t->now < 0x80 && c < 0x80) { + rb_raise(rb_eArgError, + "invalid range \"%c-%c\" in string transliteration", + t->now, c); + } + else { + rb_raise(rb_eArgError, "invalid range in string transliteration"); + } + continue; /* not reached */ + } + else if (t->now < c) { + t->gen = 1; + t->max = c; + } + } + } + return t->now; + } + else { + while (ONIGENC_CODE_TO_MBCLEN(enc, ++t->now) <= 0) { + if (t->now == t->max) { + t->gen = 0; + goto nextpart; + } + } + if (t->now < t->max) { + return t->now; + } + else { + t->gen = 0; + return t->max; + } + } } } @@ -7743,30 +7973,30 @@ tr_trans(VALUE str, VALUE src, VALUE repl, int sflag) #define CHECK_IF_ASCII(c) \ (void)((cr == ENC_CODERANGE_7BIT && !rb_isascii(c)) ? \ - (cr = ENC_CODERANGE_VALID) : 0) + (cr = ENC_CODERANGE_VALID) : 0) StringValue(src); StringValue(repl); if (RSTRING_LEN(str) == 0 || !RSTRING_PTR(str)) return Qnil; if (RSTRING_LEN(repl) == 0) { - return rb_str_delete_bang(1, &src, str); + return rb_str_delete_bang(1, &src, str); } cr = ENC_CODERANGE(str); e1 = rb_enc_check(str, src); e2 = rb_enc_check(str, repl); if (e1 == e2) { - enc = e1; + enc = e1; } else { - enc = rb_enc_check(src, repl); + enc = rb_enc_check(src, repl); } trsrc.p = RSTRING_PTR(src); trsrc.pend = trsrc.p + RSTRING_LEN(src); if (RSTRING_LEN(src) > 1 && - rb_enc_ascget(trsrc.p, trsrc.pend, &l, enc) == '^' && - trsrc.p + l < trsrc.pend) { - cflag = 1; - trsrc.p += l; + rb_enc_ascget(trsrc.p, trsrc.pend, &l, enc) == '^' && + trsrc.p + l < trsrc.pend) { + cflag = 1; + trsrc.p += l; } trrepl.p = RSTRING_PTR(repl); trrepl.pend = trrepl.p + RSTRING_LEN(repl); @@ -7775,198 +8005,213 @@ tr_trans(VALUE str, VALUE src, VALUE repl, int sflag) trsrc.max = trrepl.max = 0; if (cflag) { - for (i=0; i<256; i++) { - trans[i] = 1; - } - while ((c = trnext(&trsrc, enc)) != errc) { - if (c < 256) { - trans[c] = errc; - } - else { - if (!hash) hash = rb_hash_new(); - rb_hash_aset(hash, UINT2NUM(c), Qtrue); - } - } - while ((c = trnext(&trrepl, enc)) != errc) - /* retrieve last replacer */; - last = trrepl.now; - for (i=0; i<256; i++) { - if (trans[i] != errc) { - trans[i] = last; - } - } + for (i=0; i<256; i++) { + trans[i] = 1; + } + while ((c = trnext(&trsrc, enc)) != errc) { + if (c < 256) { + trans[c] = errc; + } + else { + if (!hash) hash = rb_hash_new(); + rb_hash_aset(hash, UINT2NUM(c), Qtrue); + } + } + while ((c = trnext(&trrepl, enc)) != errc) + /* retrieve last replacer */; + last = trrepl.now; + for (i=0; i<256; i++) { + if (trans[i] != errc) { + trans[i] = last; + } + } } else { - unsigned int r; - - for (i=0; i<256; i++) { - trans[i] = errc; - } - while ((c = trnext(&trsrc, enc)) != errc) { - r = trnext(&trrepl, enc); - if (r == errc) r = trrepl.now; - if (c < 256) { - trans[c] = r; - if (rb_enc_codelen(r, enc) != 1) singlebyte = 0; - } - else { - if (!hash) hash = rb_hash_new(); - rb_hash_aset(hash, UINT2NUM(c), UINT2NUM(r)); - } - } + unsigned int r; + + for (i=0; i<256; i++) { + trans[i] = errc; + } + while ((c = trnext(&trsrc, enc)) != errc) { + r = trnext(&trrepl, enc); + if (r == errc) r = trrepl.now; + if (c < 256) { + trans[c] = r; + if (rb_enc_codelen(r, enc) != 1) singlebyte = 0; + } + else { + if (!hash) hash = rb_hash_new(); + rb_hash_aset(hash, UINT2NUM(c), UINT2NUM(r)); + } + } } if (cr == ENC_CODERANGE_VALID && rb_enc_asciicompat(e1)) - cr = ENC_CODERANGE_7BIT; + cr = ENC_CODERANGE_7BIT; str_modify_keep_cr(str); s = (unsigned char *)RSTRING_PTR(str); send = (unsigned char *)RSTRING_END(str); termlen = rb_enc_mbminlen(enc); if (sflag) { - int clen, tlen; - long offset, max = RSTRING_LEN(str); - unsigned int save = -1; + int clen, tlen; + long offset, max = RSTRING_LEN(str); + unsigned int save = -1; unsigned char *buf = ALLOC_N(unsigned char, max + termlen), *t = buf; - while (s < send) { - int may_modify = 0; - - c0 = c = rb_enc_codepoint_len((char *)s, (char *)send, &clen, e1); - tlen = enc == e1 ? clen : rb_enc_codelen(c, enc); - - s += clen; - if (c < 256) { - c = trans[c]; - } - else if (hash) { - VALUE tmp = rb_hash_lookup(hash, UINT2NUM(c)); - if (NIL_P(tmp)) { - if (cflag) c = last; - else c = errc; - } - else if (cflag) c = errc; - else c = NUM2INT(tmp); - } - else { - c = errc; - } - if (c != (unsigned int)-1) { - if (save == c) { - CHECK_IF_ASCII(c); - continue; - } - save = c; - tlen = rb_enc_codelen(c, enc); - modify = 1; - } - else { - save = -1; - c = c0; - if (enc != e1) may_modify = 1; - } - if ((offset = t - buf) + tlen > max) { - size_t MAYBE_UNUSED(old) = max + termlen; - max = offset + tlen + (send - s); + while (s < send) { + int may_modify = 0; + + int r = rb_enc_precise_mbclen((char *)s, (char *)send, e1); + if (!MBCLEN_CHARFOUND_P(r)) { + xfree(buf); + rb_raise(rb_eArgError, "invalid byte sequence in %s", rb_enc_name(e1)); + } + clen = MBCLEN_CHARFOUND_LEN(r); + c0 = c = rb_enc_mbc_to_codepoint((char *)s, (char *)send, e1); + + tlen = enc == e1 ? clen : rb_enc_codelen(c, enc); + + s += clen; + if (c < 256) { + c = trans[c]; + } + else if (hash) { + VALUE tmp = rb_hash_lookup(hash, UINT2NUM(c)); + if (NIL_P(tmp)) { + if (cflag) c = last; + else c = errc; + } + else if (cflag) c = errc; + else c = NUM2INT(tmp); + } + else { + c = errc; + } + if (c != (unsigned int)-1) { + if (save == c) { + CHECK_IF_ASCII(c); + continue; + } + save = c; + tlen = rb_enc_codelen(c, enc); + modify = 1; + } + else { + save = -1; + c = c0; + if (enc != e1) may_modify = 1; + } + if ((offset = t - buf) + tlen > max) { + size_t MAYBE_UNUSED(old) = max + termlen; + max = offset + tlen + (send - s); SIZED_REALLOC_N(buf, unsigned char, max + termlen, old); - t = buf + offset; - } - rb_enc_mbcput(c, t, enc); - if (may_modify && memcmp(s, t, tlen) != 0) { - modify = 1; - } - CHECK_IF_ASCII(c); - t += tlen; - } - if (!STR_EMBED_P(str)) { - ruby_sized_xfree(STR_HEAP_PTR(str), STR_HEAP_SIZE(str)); - } + t = buf + offset; + } + rb_enc_mbcput(c, t, enc); + if (may_modify && memcmp(s, t, tlen) != 0) { + modify = 1; + } + CHECK_IF_ASCII(c); + t += tlen; + } + if (!STR_EMBED_P(str)) { + ruby_sized_xfree(STR_HEAP_PTR(str), STR_HEAP_SIZE(str)); + } TERM_FILL((char *)t, termlen); RSTRING(str)->as.heap.ptr = (char *)buf; - RSTRING(str)->as.heap.len = t - buf; - STR_SET_NOEMBED(str); - RSTRING(str)->as.heap.aux.capa = max; + STR_SET_LEN(str, t - buf); + STR_SET_NOEMBED(str); + RSTRING(str)->as.heap.aux.capa = max; } else if (rb_enc_mbmaxlen(enc) == 1 || (singlebyte && !hash)) { - while (s < send) { - c = (unsigned char)*s; - if (trans[c] != errc) { - if (!cflag) { - c = trans[c]; - *s = c; - modify = 1; - } - else { - *s = last; - modify = 1; - } - } - CHECK_IF_ASCII(c); - s++; - } + while (s < send) { + c = (unsigned char)*s; + if (trans[c] != errc) { + if (!cflag) { + c = trans[c]; + *s = c; + modify = 1; + } + else { + *s = last; + modify = 1; + } + } + CHECK_IF_ASCII(c); + s++; + } } else { - int clen, tlen; - long offset, max = (long)((send - s) * 1.2); + int clen, tlen; + long offset, max = (long)((send - s) * 1.2); unsigned char *buf = ALLOC_N(unsigned char, max + termlen), *t = buf; - while (s < send) { - int may_modify = 0; - c0 = c = rb_enc_codepoint_len((char *)s, (char *)send, &clen, e1); - tlen = enc == e1 ? clen : rb_enc_codelen(c, enc); - - if (c < 256) { - c = trans[c]; - } - else if (hash) { - VALUE tmp = rb_hash_lookup(hash, UINT2NUM(c)); - if (NIL_P(tmp)) { - if (cflag) c = last; - else c = errc; - } - else if (cflag) c = errc; - else c = NUM2INT(tmp); - } - else { - c = cflag ? last : errc; - } - if (c != errc) { - tlen = rb_enc_codelen(c, enc); - modify = 1; - } - else { - c = c0; - if (enc != e1) may_modify = 1; - } - if ((offset = t - buf) + tlen > max) { - size_t MAYBE_UNUSED(old) = max + termlen; - max = offset + tlen + (long)((send - s) * 1.2); + while (s < send) { + int may_modify = 0; + + int r = rb_enc_precise_mbclen((char *)s, (char *)send, e1); + if (!MBCLEN_CHARFOUND_P(r)) { + xfree(buf); + rb_raise(rb_eArgError, "invalid byte sequence in %s", rb_enc_name(e1)); + } + clen = MBCLEN_CHARFOUND_LEN(r); + c0 = c = rb_enc_mbc_to_codepoint((char *)s, (char *)send, e1); + + tlen = enc == e1 ? clen : rb_enc_codelen(c, enc); + + if (c < 256) { + c = trans[c]; + } + else if (hash) { + VALUE tmp = rb_hash_lookup(hash, UINT2NUM(c)); + if (NIL_P(tmp)) { + if (cflag) c = last; + else c = errc; + } + else if (cflag) c = errc; + else c = NUM2INT(tmp); + } + else { + c = cflag ? last : errc; + } + if (c != errc) { + tlen = rb_enc_codelen(c, enc); + modify = 1; + } + else { + c = c0; + if (enc != e1) may_modify = 1; + } + if ((offset = t - buf) + tlen > max) { + size_t MAYBE_UNUSED(old) = max + termlen; + max = offset + tlen + (long)((send - s) * 1.2); SIZED_REALLOC_N(buf, unsigned char, max + termlen, old); - t = buf + offset; - } - if (s != t) { - rb_enc_mbcput(c, t, enc); - if (may_modify && memcmp(s, t, tlen) != 0) { - modify = 1; - } - } - CHECK_IF_ASCII(c); - s += clen; - t += tlen; - } - if (!STR_EMBED_P(str)) { - ruby_sized_xfree(STR_HEAP_PTR(str), STR_HEAP_SIZE(str)); - } + t = buf + offset; + } + if (s != t) { + rb_enc_mbcput(c, t, enc); + if (may_modify && memcmp(s, t, tlen) != 0) { + modify = 1; + } + } + CHECK_IF_ASCII(c); + s += clen; + t += tlen; + } + if (!STR_EMBED_P(str)) { + ruby_sized_xfree(STR_HEAP_PTR(str), STR_HEAP_SIZE(str)); + } TERM_FILL((char *)t, termlen); RSTRING(str)->as.heap.ptr = (char *)buf; - RSTRING(str)->as.heap.len = t - buf; - STR_SET_NOEMBED(str); - RSTRING(str)->as.heap.aux.capa = max; + STR_SET_LEN(str, t - buf); + STR_SET_NOEMBED(str); + RSTRING(str)->as.heap.aux.capa = max; } if (modify) { - if (cr != ENC_CODERANGE_BROKEN) - ENC_CODERANGE_SET(str, cr); - rb_enc_associate(str, enc); - return str; + if (cr != ENC_CODERANGE_BROKEN) + ENC_CODERANGE_SET(str, cr); + rb_enc_associate(str, enc); + return str; } return Qnil; } @@ -7998,7 +8243,7 @@ rb_str_tr_bang(VALUE str, VALUE src, VALUE repl) * * - Each occurrence of the first character specified by +selector+ * is translated to the first character in +replacements+. - * - Each occurrence of the second character specified by selector+ + * - Each occurrence of the second character specified by +selector+ * is translated to the second character in +replacements+. * - And so on. * @@ -8039,7 +8284,7 @@ rb_str_tr(VALUE str, VALUE src, VALUE repl) #define TR_TABLE_SIZE (TR_TABLE_MAX+1) static void tr_setup_table(VALUE str, char stable[TR_TABLE_SIZE], int first, - VALUE *tablep, VALUE *ctablep, rb_encoding *enc) + VALUE *tablep, VALUE *ctablep, rb_encoding *enc) { const unsigned int errc = -1; char buf[TR_TABLE_MAX]; @@ -8052,51 +8297,51 @@ tr_setup_table(VALUE str, char stable[TR_TABLE_SIZE], int first, tr.gen = tr.now = tr.max = 0; if (RSTRING_LEN(str) > 1 && rb_enc_ascget(tr.p, tr.pend, &l, enc) == '^') { - cflag = 1; - tr.p += l; + cflag = 1; + tr.p += l; } if (first) { - for (i=0; i<TR_TABLE_MAX; i++) { - stable[i] = 1; - } - stable[TR_TABLE_MAX] = cflag; + for (i=0; i<TR_TABLE_MAX; i++) { + stable[i] = 1; + } + stable[TR_TABLE_MAX] = cflag; } else if (stable[TR_TABLE_MAX] && !cflag) { - stable[TR_TABLE_MAX] = 0; + stable[TR_TABLE_MAX] = 0; } for (i=0; i<TR_TABLE_MAX; i++) { - buf[i] = cflag; + buf[i] = cflag; } while ((c = trnext(&tr, enc)) != errc) { - if (c < TR_TABLE_MAX) { - buf[(unsigned char)c] = !cflag; - } - else { - VALUE key = UINT2NUM(c); - - if (!table && (first || *tablep || stable[TR_TABLE_MAX])) { - if (cflag) { - ptable = *ctablep; - table = ptable ? ptable : rb_hash_new(); - *ctablep = table; - } - else { - table = rb_hash_new(); - ptable = *tablep; - *tablep = table; - } - } - if (table && (!ptable || (cflag ^ !NIL_P(rb_hash_aref(ptable, key))))) { - rb_hash_aset(table, key, Qtrue); - } - } + if (c < TR_TABLE_MAX) { + buf[(unsigned char)c] = !cflag; + } + else { + VALUE key = UINT2NUM(c); + + if (!table && (first || *tablep || stable[TR_TABLE_MAX])) { + if (cflag) { + ptable = *ctablep; + table = ptable ? ptable : rb_hash_new(); + *ctablep = table; + } + else { + table = rb_hash_new(); + ptable = *tablep; + *tablep = table; + } + } + if (table && (!ptable || (cflag ^ !NIL_P(rb_hash_aref(ptable, key))))) { + rb_hash_aset(table, key, Qtrue); + } + } } for (i=0; i<TR_TABLE_MAX; i++) { - stable[i] = stable[i] && buf[i]; + stable[i] = stable[i] && buf[i]; } if (!table && !cflag) { - *tablep = 0; + *tablep = 0; } } @@ -8105,21 +8350,21 @@ static int tr_find(unsigned int c, const char table[TR_TABLE_SIZE], VALUE del, VALUE nodel) { if (c < TR_TABLE_MAX) { - return table[c] != 0; + return table[c] != 0; } else { - VALUE v = UINT2NUM(c); + VALUE v = UINT2NUM(c); - if (del) { - if (!NIL_P(rb_hash_lookup(del, v)) && - (!nodel || NIL_P(rb_hash_lookup(nodel, v)))) { - return TRUE; - } - } - else if (nodel && !NIL_P(rb_hash_lookup(nodel, v))) { - return FALSE; - } - return table[TR_TABLE_MAX] ? TRUE : FALSE; + if (del) { + if (!NIL_P(rb_hash_lookup(del, v)) && + (!nodel || NIL_P(rb_hash_lookup(nodel, v)))) { + return TRUE; + } + } + else if (nodel && !NIL_P(rb_hash_lookup(nodel, v))) { + return FALSE; + } + return table[TR_TABLE_MAX] ? TRUE : FALSE; } } @@ -8145,11 +8390,11 @@ rb_str_delete_bang(int argc, VALUE *argv, VALUE str) if (RSTRING_LEN(str) == 0 || !RSTRING_PTR(str)) return Qnil; rb_check_arity(argc, 1, UNLIMITED_ARGUMENTS); for (i=0; i<argc; i++) { - VALUE s = argv[i]; + VALUE s = argv[i]; - StringValue(s); - enc = rb_enc_check(str, s); - tr_setup_table(s, squeez, i==0, &del, &nodel, enc); + StringValue(s); + enc = rb_enc_check(str, s); + tr_setup_table(s, squeez, i==0, &del, &nodel, enc); } str_modify_keep_cr(str); @@ -8158,32 +8403,32 @@ rb_str_delete_bang(int argc, VALUE *argv, VALUE str) send = RSTRING_END(str); cr = ascompat ? ENC_CODERANGE_7BIT : ENC_CODERANGE_VALID; while (s < send) { - unsigned int c; - int clen; - - if (ascompat && (c = *(unsigned char*)s) < 0x80) { - if (squeez[c]) { - modify = 1; - } - else { - if (t != s) *t = c; - t++; - } - s++; - } - else { - c = rb_enc_codepoint_len(s, send, &clen, enc); - - if (tr_find(c, squeez, del, nodel)) { - modify = 1; - } - else { - if (t != s) rb_enc_mbcput(c, t, enc); - t += clen; - if (cr == ENC_CODERANGE_7BIT) cr = ENC_CODERANGE_VALID; - } - s += clen; - } + unsigned int c; + int clen; + + if (ascompat && (c = *(unsigned char*)s) < 0x80) { + if (squeez[c]) { + modify = 1; + } + else { + if (t != s) *t = c; + t++; + } + s++; + } + else { + c = rb_enc_codepoint_len(s, send, &clen, enc); + + if (tr_find(c, squeez, del, nodel)) { + modify = 1; + } + else { + if (t != s) rb_enc_mbcput(c, t, enc); + t += clen; + if (cr == ENC_CODERANGE_7BIT) cr = ENC_CODERANGE_VALID; + } + s += clen; + } } TERM_FILL(t, TERM_LEN(str)); STR_SET_LEN(str, t - RSTRING_PTR(str)); @@ -8237,18 +8482,18 @@ rb_str_squeeze_bang(int argc, VALUE *argv, VALUE str) unsigned int save; if (argc == 0) { - enc = STR_ENC_GET(str); + enc = STR_ENC_GET(str); } else { - for (i=0; i<argc; i++) { - VALUE s = argv[i]; - - StringValue(s); - enc = rb_enc_check(str, s); - if (singlebyte && !single_byte_optimizable(s)) - singlebyte = 0; - tr_setup_table(s, squeez, i==0, &del, &nodel, enc); - } + for (i=0; i<argc; i++) { + VALUE s = argv[i]; + + StringValue(s); + enc = rb_enc_check(str, s); + if (singlebyte && !single_byte_optimizable(s)) + singlebyte = 0; + tr_setup_table(s, squeez, i==0, &del, &nodel, enc); + } } str_modify_keep_cr(str); @@ -8261,39 +8506,39 @@ rb_str_squeeze_bang(int argc, VALUE *argv, VALUE str) if (singlebyte) { while (s < send) { unsigned int c = *s++; - if (c != save || (argc > 0 && !squeez[c])) { - *t++ = save = c; - } - } + if (c != save || (argc > 0 && !squeez[c])) { + *t++ = save = c; + } + } } else { - while (s < send) { - unsigned int c; - int clen; + while (s < send) { + unsigned int c; + int clen; if (ascompat && (c = *s) < 0x80) { - if (c != save || (argc > 0 && !squeez[c])) { - *t++ = save = c; - } - s++; - } - else { + if (c != save || (argc > 0 && !squeez[c])) { + *t++ = save = c; + } + s++; + } + else { c = rb_enc_codepoint_len((char *)s, (char *)send, &clen, enc); - if (c != save || (argc > 0 && !tr_find(c, squeez, del, nodel))) { - if (t != s) rb_enc_mbcput(c, t, enc); - save = c; - t += clen; - } - s += clen; - } - } + if (c != save || (argc > 0 && !tr_find(c, squeez, del, nodel))) { + if (t != s) rb_enc_mbcput(c, t, enc); + save = c; + t += clen; + } + s += clen; + } + } } TERM_FILL((char *)t, TERM_LEN(str)); if ((char *)t - RSTRING_PTR(str) != RSTRING_LEN(str)) { STR_SET_LEN(str, (char *)t - RSTRING_PTR(str)); - modify = 1; + modify = 1; } if (modify) return str; @@ -8408,30 +8653,30 @@ rb_str_count(int argc, VALUE *argv, VALUE str) StringValue(tstr); enc = rb_enc_check(str, tstr); if (argc == 1) { - const char *ptstr; - if (RSTRING_LEN(tstr) == 1 && rb_enc_asciicompat(enc) && - (ptstr = RSTRING_PTR(tstr), - ONIGENC_IS_ALLOWED_REVERSE_MATCH(enc, (const unsigned char *)ptstr, (const unsigned char *)ptstr+1)) && - !is_broken_string(str)) { - int clen; - unsigned char c = rb_enc_codepoint_len(ptstr, ptstr+1, &clen, enc); - - s = RSTRING_PTR(str); - if (!s || RSTRING_LEN(str) == 0) return INT2FIX(0); - send = RSTRING_END(str); - while (s < send) { - if (*(unsigned char*)s++ == c) n++; - } - return SIZET2NUM(n); - } + const char *ptstr; + if (RSTRING_LEN(tstr) == 1 && rb_enc_asciicompat(enc) && + (ptstr = RSTRING_PTR(tstr), + ONIGENC_IS_ALLOWED_REVERSE_MATCH(enc, (const unsigned char *)ptstr, (const unsigned char *)ptstr+1)) && + !is_broken_string(str)) { + int clen; + unsigned char c = rb_enc_codepoint_len(ptstr, ptstr+1, &clen, enc); + + s = RSTRING_PTR(str); + if (!s || RSTRING_LEN(str) == 0) return INT2FIX(0); + send = RSTRING_END(str); + while (s < send) { + if (*(unsigned char*)s++ == c) n++; + } + return SIZET2NUM(n); + } } tr_setup_table(tstr, table, TRUE, &del, &nodel, enc); for (i=1; i<argc; i++) { - tstr = argv[i]; - StringValue(tstr); - enc = rb_enc_check(str, tstr); - tr_setup_table(tstr, table, FALSE, &del, &nodel, enc); + tstr = argv[i]; + StringValue(tstr); + enc = rb_enc_check(str, tstr); + tr_setup_table(tstr, table, FALSE, &del, &nodel, enc); } s = RSTRING_PTR(str); @@ -8439,22 +8684,22 @@ rb_str_count(int argc, VALUE *argv, VALUE str) send = RSTRING_END(str); ascompat = rb_enc_asciicompat(enc); while (s < send) { - unsigned int c; - - if (ascompat && (c = *(unsigned char*)s) < 0x80) { - if (table[c]) { - n++; - } - s++; - } - else { - int clen; - c = rb_enc_codepoint_len(s, send, &clen, enc); - if (tr_find(c, table, del, nodel)) { - n++; - } - s += clen; - } + unsigned int c; + + if (ascompat && (c = *(unsigned char*)s) < 0x80) { + if (table[c]) { + n++; + } + s++; + } + else { + int clen; + c = rb_enc_codepoint_len(s, send, &clen, enc); + if (tr_find(c, table, del, nodel)) { + n++; + } + s += clen; + } } return SIZET2NUM(n); @@ -8464,8 +8709,8 @@ static VALUE rb_fs_check(VALUE val) { if (!NIL_P(val) && !RB_TYPE_P(val, T_STRING) && !RB_TYPE_P(val, T_REGEXP)) { - val = rb_check_string_type(val); - if (NIL_P(val)) return 0; + val = rb_check_string_type(val); + if (NIL_P(val)) return 0; } return val; } @@ -8495,27 +8740,27 @@ static long split_string(VALUE result, VALUE str, long beg, long len, long empty_count) { if (empty_count >= 0 && len == 0) { - return empty_count + 1; + return empty_count + 1; } if (empty_count > 0) { - /* make different substrings */ - if (result) { - do { + /* make different substrings */ + if (result) { + do { rb_ary_push(result, str_new_empty_String(str)); - } while (--empty_count > 0); - } - else { - do { + } while (--empty_count > 0); + } + else { + do { rb_yield(str_new_empty_String(str)); - } while (--empty_count > 0); - } + } while (--empty_count > 0); + } } str = rb_str_subseq(str, beg, len); if (result) { - rb_ary_push(result, str); + rb_ary_push(result, str); } else { - rb_yield(str); + rb_yield(str); } return empty_count; } @@ -8550,7 +8795,7 @@ literal_split_pattern(VALUE spat, split_type_t default_type) } /* - * :call-seq: + * call-seq: * split(field_sep = $;, limit = nil) -> array * split(field_sep = $;, limit = nil) {|substring| ... } -> self * @@ -8571,32 +8816,32 @@ rb_str_split_m(int argc, VALUE *argv, VALUE str) result = rb_block_given_p() ? Qfalse : Qnil; if (rb_scan_args(argc, argv, "02", &spat, &limit) == 2) { - lim = NUM2INT(limit); - if (lim <= 0) limit = Qnil; - else if (lim == 1) { - if (RSTRING_LEN(str) == 0) + lim = NUM2INT(limit); + if (lim <= 0) limit = Qnil; + else if (lim == 1) { + if (RSTRING_LEN(str) == 0) return result ? rb_ary_new2(0) : str; tmp = str_duplicate(rb_cString, str); - if (!result) { - rb_yield(tmp); + if (!result) { + rb_yield(tmp); return str; - } - return rb_ary_new3(1, tmp); - } - i = 1; + } + return rb_ary_new3(1, tmp); + } + i = 1; } if (NIL_P(limit) && !lim) empty_count = 0; enc = STR_ENC_GET(str); split_type = SPLIT_TYPE_REGEXP; if (!NIL_P(spat)) { - spat = get_pat_quoted(spat, 0); + spat = get_pat_quoted(spat, 0); } else if (NIL_P(spat = rb_fs)) { - split_type = SPLIT_TYPE_AWK; + split_type = SPLIT_TYPE_AWK; } else if (!(spat = rb_fs_check(spat))) { - rb_raise(rb_eTypeError, "value of $; must be String or Regexp"); + rb_raise(rb_eTypeError, "value of $; must be String or Regexp"); } else { rb_category_warn(RB_WARN_CATEGORY_DEPRECATED, "$; is set to non-nil value"); @@ -8614,106 +8859,108 @@ rb_str_split_m(int argc, VALUE *argv, VALUE str) break; case T_STRING: - mustnot_broken(spat); + mustnot_broken(spat); split_type = literal_split_pattern(spat, SPLIT_TYPE_STRING); break; default: UNREACHABLE_RETURN(Qnil); - } + } } #define SPLIT_STR(beg, len) (empty_count = split_string(result, str, beg, len, empty_count)) - if (result) result = rb_ary_new(); beg = 0; char *ptr = RSTRING_PTR(str); char *eptr = RSTRING_END(str); if (split_type == SPLIT_TYPE_AWK) { - char *bptr = ptr; - int skip = 1; - unsigned int c; - - end = beg; - if (is_ascii_string(str)) { - while (ptr < eptr) { - c = (unsigned char)*ptr++; - if (skip) { - if (ascii_isspace(c)) { - beg = ptr - bptr; - } - else { - end = ptr - bptr; - skip = 0; - if (!NIL_P(limit) && lim <= i) break; - } - } - else if (ascii_isspace(c)) { - SPLIT_STR(beg, end-beg); - skip = 1; - beg = ptr - bptr; - if (!NIL_P(limit)) ++i; - } - else { - end = ptr - bptr; - } - } - } - else { - while (ptr < eptr) { - int n; - - c = rb_enc_codepoint_len(ptr, eptr, &n, enc); - ptr += n; - if (skip) { - if (rb_isspace(c)) { - beg = ptr - bptr; - } - else { - end = ptr - bptr; - skip = 0; - if (!NIL_P(limit) && lim <= i) break; - } - } - else if (rb_isspace(c)) { - SPLIT_STR(beg, end-beg); - skip = 1; - beg = ptr - bptr; - if (!NIL_P(limit)) ++i; - } - else { - end = ptr - bptr; - } - } - } + char *bptr = ptr; + int skip = 1; + unsigned int c; + + if (result) result = rb_ary_new(); + end = beg; + if (is_ascii_string(str)) { + while (ptr < eptr) { + c = (unsigned char)*ptr++; + if (skip) { + if (ascii_isspace(c)) { + beg = ptr - bptr; + } + else { + end = ptr - bptr; + skip = 0; + if (!NIL_P(limit) && lim <= i) break; + } + } + else if (ascii_isspace(c)) { + SPLIT_STR(beg, end-beg); + skip = 1; + beg = ptr - bptr; + if (!NIL_P(limit)) ++i; + } + else { + end = ptr - bptr; + } + } + } + else { + while (ptr < eptr) { + int n; + + c = rb_enc_codepoint_len(ptr, eptr, &n, enc); + ptr += n; + if (skip) { + if (rb_isspace(c)) { + beg = ptr - bptr; + } + else { + end = ptr - bptr; + skip = 0; + if (!NIL_P(limit) && lim <= i) break; + } + } + else if (rb_isspace(c)) { + SPLIT_STR(beg, end-beg); + skip = 1; + beg = ptr - bptr; + if (!NIL_P(limit)) ++i; + } + else { + end = ptr - bptr; + } + } + } } else if (split_type == SPLIT_TYPE_STRING) { - char *str_start = ptr; - char *substr_start = ptr; - char *sptr = RSTRING_PTR(spat); - long slen = RSTRING_LEN(spat); - - mustnot_broken(str); - enc = rb_enc_check(str, spat); - while (ptr < eptr && - (end = rb_memsearch(sptr, slen, ptr, eptr - ptr, enc)) >= 0) { - /* Check we are at the start of a char */ - char *t = rb_enc_right_char_head(ptr, ptr + end, eptr, enc); - if (t != ptr + end) { - ptr = t; - continue; - } - SPLIT_STR(substr_start - str_start, (ptr+end) - substr_start); - ptr += end + slen; - substr_start = ptr; - if (!NIL_P(limit) && lim <= ++i) break; - } - beg = ptr - str_start; + char *str_start = ptr; + char *substr_start = ptr; + char *sptr = RSTRING_PTR(spat); + long slen = RSTRING_LEN(spat); + + if (result) result = rb_ary_new(); + mustnot_broken(str); + enc = rb_enc_check(str, spat); + while (ptr < eptr && + (end = rb_memsearch(sptr, slen, ptr, eptr - ptr, enc)) >= 0) { + /* Check we are at the start of a char */ + char *t = rb_enc_right_char_head(ptr, ptr + end, eptr, enc); + if (t != ptr + end) { + ptr = t; + continue; + } + SPLIT_STR(substr_start - str_start, (ptr+end) - substr_start); + ptr += end + slen; + substr_start = ptr; + if (!NIL_P(limit) && lim <= ++i) break; + } + beg = ptr - str_start; } else if (split_type == SPLIT_TYPE_CHARS) { char *str_start = ptr; int n; + if (result) result = rb_ary_new_capa(RSTRING_LEN(str)); mustnot_broken(str); enc = rb_enc_get(str); while (ptr < eptr && @@ -8725,11 +8972,12 @@ rb_str_split_m(int argc, VALUE *argv, VALUE str) beg = ptr - str_start; } else { - long len = RSTRING_LEN(str); - long start = beg; - long idx; - int last_null = 0; - struct re_registers *regs; + if (result) result = rb_ary_new(); + long len = RSTRING_LEN(str); + long start = beg; + long idx; + int last_null = 0; + struct re_registers *regs; VALUE match = 0; for (; rb_reg_search(spat, str, start, 0) >= 0; @@ -8738,40 +8986,40 @@ rb_str_split_m(int argc, VALUE *argv, VALUE str) if (!result) rb_match_busy(match); regs = RMATCH_REGS(match); end = BEG(0); - if (start == end && BEG(0) == END(0)) { - if (!ptr) { - SPLIT_STR(0, 0); - break; - } - else if (last_null == 1) { + if (start == end && BEG(0) == END(0)) { + if (!ptr) { + SPLIT_STR(0, 0); + break; + } + else if (last_null == 1) { SPLIT_STR(beg, rb_enc_fast_mbclen(ptr+beg, eptr, enc)); - beg = start; - } - else { + beg = start; + } + else { if (start == len) start++; else start += rb_enc_fast_mbclen(ptr+start,eptr,enc); - last_null = 1; - continue; - } - } - else { - SPLIT_STR(beg, end-beg); - beg = start = END(0); - } - last_null = 0; - - for (idx=1; idx < regs->num_regs; idx++) { - if (BEG(idx) == -1) continue; - SPLIT_STR(BEG(idx), END(idx)-BEG(idx)); - } - if (!NIL_P(limit) && lim <= ++i) break; - } + last_null = 1; + continue; + } + } + else { + SPLIT_STR(beg, end-beg); + beg = start = END(0); + } + last_null = 0; + + for (idx=1; idx < regs->num_regs; idx++) { + if (BEG(idx) == -1) continue; + SPLIT_STR(BEG(idx), END(idx)-BEG(idx)); + } + if (!NIL_P(limit) && lim <= ++i) break; + } if (match) rb_match_unbusy(match); } if (RSTRING_LEN(str) > 0 && (!NIL_P(limit) || RSTRING_LEN(str) > beg || lim < 0)) { - SPLIT_STR(beg, RSTRING_LEN(str)-beg); + SPLIT_STR(beg, RSTRING_LEN(str)-beg); } return result ? result : str; @@ -8793,12 +9041,12 @@ static inline int enumerator_element(VALUE ary, VALUE e) { if (ary) { - rb_ary_push(ary, e); - return 0; + rb_ary_push(ary, e); + return 0; } else { - rb_yield(e); - return 1; + rb_yield(e); + return 1; } } @@ -8809,10 +9057,10 @@ chomp_newline(const char *p, const char *e, rb_encoding *enc) { const char *prev = rb_enc_prev_char(p, e, e, enc); if (rb_enc_is_newline(prev, e, enc)) { - e = prev; - prev = rb_enc_prev_char(p, e, e, enc); - if (prev && rb_enc_ascget(prev, e, NULL, enc) == '\r') - e = prev; + e = prev; + prev = rb_enc_prev_char(p, e, e, enc); + if (prev && rb_enc_ascget(prev, e, NULL, enc) == '\r') + e = prev; } return e; } @@ -8822,9 +9070,9 @@ get_rs(void) { VALUE rs = rb_rs; if (!NIL_P(rs) && - (!RB_TYPE_P(rs, T_STRING) || - RSTRING_LEN(rs) != 1 || - RSTRING_PTR(rs)[0] != '\n')) { + (!RB_TYPE_P(rs, T_STRING) || + RSTRING_LEN(rs) != 1 || + RSTRING_PTR(rs)[0] != '\n')) { rb_category_warn(RB_WARN_CATEGORY_DEPRECATED, "$/ is set to non-default value"); } return rs; @@ -8842,23 +9090,23 @@ rb_str_enumerate_lines(int argc, VALUE *argv, VALUE str, VALUE ary) int rsnewline = 0; if (rb_scan_args(argc, argv, "01:", &rs, &opts) == 0) - rs = rb_rs; + rs = rb_rs; if (!NIL_P(opts)) { - static ID keywords[1]; - if (!keywords[0]) { - keywords[0] = rb_intern_const("chomp"); - } - rb_get_kwargs(opts, keywords, 0, 1, &chomp); - chomp = (chomp != Qundef && RTEST(chomp)); + static ID keywords[1]; + if (!keywords[0]) { + keywords[0] = rb_intern_const("chomp"); + } + rb_get_kwargs(opts, keywords, 0, 1, &chomp); + chomp = (!UNDEF_P(chomp) && RTEST(chomp)); } if (NIL_P(rs)) { - if (!ENUM_ELEM(ary, str)) { - return ary; - } - else { - return orig; - } + if (!ENUM_ELEM(ary, str)) { + return ary; + } + else { + return orig; + } } if (!RSTRING_LEN(str)) goto end; @@ -8870,101 +9118,106 @@ rb_str_enumerate_lines(int argc, VALUE *argv, VALUE str, VALUE ary) rslen = RSTRING_LEN(rs); if (rs == rb_default_rs) - enc = rb_enc_get(str); + enc = rb_enc_get(str); else - enc = rb_enc_check(str, rs); + enc = rb_enc_check(str, rs); if (rslen == 0) { - /* paragraph mode */ - int n; - const char *eol = NULL; - subend = subptr; - while (subend < pend) { - do { - if (rb_enc_ascget(subend, pend, &n, enc) != '\r') - n = 0; - rslen = n + rb_enc_mbclen(subend + n, pend, enc); - if (rb_enc_is_newline(subend + n, pend, enc)) { - if (eol == subend) break; - subend += rslen; - if (subptr) eol = subend; - } - else { - if (!subptr) subptr = subend; - subend += rslen; - } - rslen = 0; - } while (subend < pend); - if (!subptr) break; - line = rb_str_subseq(str, subptr - ptr, - subend - subptr + (chomp ? 0 : rslen)); - if (ENUM_ELEM(ary, line)) { - str_mod_check(str, ptr, len); - } - subptr = eol = NULL; - } - goto end; + /* paragraph mode */ + int n; + const char *eol = NULL; + subend = subptr; + while (subend < pend) { + long chomp_rslen = 0; + do { + if (rb_enc_ascget(subend, pend, &n, enc) != '\r') + n = 0; + rslen = n + rb_enc_mbclen(subend + n, pend, enc); + if (rb_enc_is_newline(subend + n, pend, enc)) { + if (eol == subend) break; + subend += rslen; + if (subptr) { + eol = subend; + chomp_rslen = -rslen; + } + } + else { + if (!subptr) subptr = subend; + subend += rslen; + } + rslen = 0; + } while (subend < pend); + if (!subptr) break; + if (rslen == 0) chomp_rslen = 0; + line = rb_str_subseq(str, subptr - ptr, + subend - subptr + (chomp ? chomp_rslen : rslen)); + if (ENUM_ELEM(ary, line)) { + str_mod_check(str, ptr, len); + } + subptr = eol = NULL; + } + goto end; } else { - rsptr = RSTRING_PTR(rs); - if (RSTRING_LEN(rs) == rb_enc_mbminlen(enc) && - rb_enc_is_newline(rsptr, rsptr + RSTRING_LEN(rs), enc)) { - rsnewline = 1; - } + rsptr = RSTRING_PTR(rs); + if (RSTRING_LEN(rs) == rb_enc_mbminlen(enc) && + rb_enc_is_newline(rsptr, rsptr + RSTRING_LEN(rs), enc)) { + rsnewline = 1; + } } if ((rs == rb_default_rs) && !rb_enc_asciicompat(enc)) { - rs = rb_str_new(rsptr, rslen); - rs = rb_str_encode(rs, rb_enc_from_encoding(enc), 0, Qnil); - rsptr = RSTRING_PTR(rs); - rslen = RSTRING_LEN(rs); + rs = rb_str_new(rsptr, rslen); + rs = rb_str_encode(rs, rb_enc_from_encoding(enc), 0, Qnil); + rsptr = RSTRING_PTR(rs); + rslen = RSTRING_LEN(rs); } while (subptr < pend) { - pos = rb_memsearch(rsptr, rslen, subptr, pend - subptr, enc); - if (pos < 0) break; - hit = subptr + pos; - adjusted = rb_enc_right_char_head(subptr, hit, pend, enc); - if (hit != adjusted) { - subptr = adjusted; - continue; - } - subend = hit += rslen; - if (chomp) { - if (rsnewline) { - subend = chomp_newline(subptr, subend, enc); - } - else { - subend -= rslen; - } - } - line = rb_str_subseq(str, subptr - ptr, subend - subptr); - if (ENUM_ELEM(ary, line)) { - str_mod_check(str, ptr, len); - } - subptr = hit; + pos = rb_memsearch(rsptr, rslen, subptr, pend - subptr, enc); + if (pos < 0) break; + hit = subptr + pos; + adjusted = rb_enc_right_char_head(subptr, hit, pend, enc); + if (hit != adjusted) { + subptr = adjusted; + continue; + } + subend = hit += rslen; + if (chomp) { + if (rsnewline) { + subend = chomp_newline(subptr, subend, enc); + } + else { + subend -= rslen; + } + } + line = rb_str_subseq(str, subptr - ptr, subend - subptr); + if (ENUM_ELEM(ary, line)) { + str_mod_check(str, ptr, len); + } + subptr = hit; } if (subptr != pend) { - if (chomp) { - if (rsnewline) { - pend = chomp_newline(subptr, pend, enc); - } - else if (pend - subptr >= rslen && - memcmp(pend - rslen, rsptr, rslen) == 0) { - pend -= rslen; - } - } - line = rb_str_subseq(str, subptr - ptr, pend - subptr); - ENUM_ELEM(ary, line); - RB_GC_GUARD(str); + if (chomp) { + if (rsnewline) { + pend = chomp_newline(subptr, pend, enc); + } + else if (pend - subptr >= rslen && + memcmp(pend - rslen, rsptr, rslen) == 0) { + pend -= rslen; + } + } + line = rb_str_subseq(str, subptr - ptr, pend - subptr); + ENUM_ELEM(ary, line); + RB_GC_GUARD(str); } end: if (ary) - return ary; + return ary; else - return orig; + return orig; } /* @@ -9011,12 +9264,12 @@ rb_str_enumerate_bytes(VALUE str, VALUE ary) long i; for (i=0; i<RSTRING_LEN(str); i++) { - ENUM_ELEM(ary, INT2FIX((unsigned char)RSTRING_PTR(str)[i])); + ENUM_ELEM(ary, INT2FIX((unsigned char)RSTRING_PTR(str)[i])); } if (ary) - return ary; + return ary; else - return str; + return str; } /* @@ -9070,22 +9323,22 @@ rb_str_enumerate_chars(VALUE str, VALUE ary) enc = rb_enc_get(str); if (ENC_CODERANGE_CLEAN_P(ENC_CODERANGE(str))) { - for (i = 0; i < len; i += n) { - n = rb_enc_fast_mbclen(ptr + i, ptr + len, enc); - ENUM_ELEM(ary, rb_str_subseq(str, i, n)); - } + for (i = 0; i < len; i += n) { + n = rb_enc_fast_mbclen(ptr + i, ptr + len, enc); + ENUM_ELEM(ary, rb_str_subseq(str, i, n)); + } } else { - for (i = 0; i < len; i += n) { - n = rb_enc_mbclen(ptr + i, ptr + len, enc); - ENUM_ELEM(ary, rb_str_subseq(str, i, n)); - } + for (i = 0; i < len; i += n) { + n = rb_enc_mbclen(ptr + i, ptr + len, enc); + ENUM_ELEM(ary, rb_str_subseq(str, i, n)); + } } RB_GC_GUARD(str); if (ary) - return ary; + return ary; else - return orig; + return orig; } /* @@ -9129,7 +9382,7 @@ rb_str_enumerate_codepoints(VALUE str, VALUE ary) rb_encoding *enc; if (single_byte_optimizable(str)) - return rb_str_enumerate_bytes(str, ary); + return rb_str_enumerate_bytes(str, ary); str = rb_str_new_frozen(str); ptr = RSTRING_PTR(str); @@ -9137,15 +9390,15 @@ rb_str_enumerate_codepoints(VALUE str, VALUE ary) enc = STR_ENC_GET(str); while (ptr < end) { - c = rb_enc_codepoint_len(ptr, end, &n, enc); - ENUM_ELEM(ary, UINT2NUM(c)); - ptr += n; + c = rb_enc_codepoint_len(ptr, end, &n, enc); + ENUM_ELEM(ary, UINT2NUM(c)); + ptr += n; } RB_GC_GUARD(str); if (ary) - return ary; + return ary; else - return orig; + return orig; } /* @@ -9183,74 +9436,93 @@ static regex_t * get_reg_grapheme_cluster(rb_encoding *enc) { int encidx = rb_enc_to_index(enc); - regex_t *reg_grapheme_cluster = NULL; - static regex_t *reg_grapheme_cluster_utf8 = NULL; - /* synchronize */ - if (encidx == rb_utf8_encindex() && reg_grapheme_cluster_utf8) { - reg_grapheme_cluster = reg_grapheme_cluster_utf8; - } - if (!reg_grapheme_cluster) { - const OnigUChar source_ascii[] = "\\X"; - OnigErrorInfo einfo; - const OnigUChar *source = source_ascii; - size_t source_len = sizeof(source_ascii) - 1; - switch (encidx) { + const OnigUChar source_ascii[] = "\\X"; + const OnigUChar *source = source_ascii; + size_t source_len = sizeof(source_ascii) - 1; + + switch (encidx) { #define CHARS_16BE(x) (OnigUChar)((x)>>8), (OnigUChar)(x) #define CHARS_16LE(x) (OnigUChar)(x), (OnigUChar)((x)>>8) #define CHARS_32BE(x) CHARS_16BE((x)>>16), CHARS_16BE(x) #define CHARS_32LE(x) CHARS_16LE(x), CHARS_16LE((x)>>16) #define CASE_UTF(e) \ - case ENCINDEX_UTF_##e: { \ - static const OnigUChar source_UTF_##e[] = {CHARS_##e('\\'), CHARS_##e('X')}; \ - source = source_UTF_##e; \ - source_len = sizeof(source_UTF_##e); \ - break; \ - } - CASE_UTF(16BE); CASE_UTF(16LE); CASE_UTF(32BE); CASE_UTF(32LE); + case ENCINDEX_UTF_##e: { \ + static const OnigUChar source_UTF_##e[] = {CHARS_##e('\\'), CHARS_##e('X')}; \ + source = source_UTF_##e; \ + source_len = sizeof(source_UTF_##e); \ + break; \ + } + CASE_UTF(16BE); CASE_UTF(16LE); CASE_UTF(32BE); CASE_UTF(32LE); #undef CASE_UTF #undef CHARS_16BE #undef CHARS_16LE #undef CHARS_32BE #undef CHARS_32LE - } - int r = onig_new(®_grapheme_cluster, source, source + source_len, - ONIG_OPTION_DEFAULT, enc, OnigDefaultSyntax, &einfo); - if (r) { - UChar message[ONIG_MAX_ERROR_MESSAGE_LEN]; - onig_error_code_to_str(message, r, &einfo); - rb_fatal("cannot compile grapheme cluster regexp: %s", (char *)message); - } - if (encidx == rb_utf8_encindex()) { - reg_grapheme_cluster_utf8 = reg_grapheme_cluster; - } } + + regex_t *reg_grapheme_cluster; + OnigErrorInfo einfo; + int r = onig_new(®_grapheme_cluster, source, source + source_len, + ONIG_OPTION_DEFAULT, enc, OnigDefaultSyntax, &einfo); + if (r) { + UChar message[ONIG_MAX_ERROR_MESSAGE_LEN]; + onig_error_code_to_str(message, r, &einfo); + rb_fatal("cannot compile grapheme cluster regexp: %s", (char *)message); + } + return reg_grapheme_cluster; } +static regex_t * +get_cached_reg_grapheme_cluster(rb_encoding *enc) +{ + int encidx = rb_enc_to_index(enc); + static regex_t *reg_grapheme_cluster_utf8 = NULL; + + if (encidx == rb_utf8_encindex()) { + if (!reg_grapheme_cluster_utf8) { + reg_grapheme_cluster_utf8 = get_reg_grapheme_cluster(enc); + } + + return reg_grapheme_cluster_utf8; + } + + return NULL; +} + static VALUE rb_str_each_grapheme_cluster_size(VALUE str, VALUE args, VALUE eobj) { size_t grapheme_cluster_count = 0; - regex_t *reg_grapheme_cluster = NULL; - rb_encoding *enc = rb_enc_from_index(ENCODING_GET(str)); + rb_encoding *enc = get_encoding(str); const char *ptr, *end; if (!rb_enc_unicode_p(enc)) { - return rb_str_length(str); + return rb_str_length(str); + } + + bool cached_reg_grapheme_cluster = true; + regex_t *reg_grapheme_cluster = get_cached_reg_grapheme_cluster(enc); + if (!reg_grapheme_cluster) { + reg_grapheme_cluster = get_reg_grapheme_cluster(enc); + cached_reg_grapheme_cluster = false; } - reg_grapheme_cluster = get_reg_grapheme_cluster(enc); ptr = RSTRING_PTR(str); end = RSTRING_END(str); while (ptr < end) { - OnigPosition len = onig_match(reg_grapheme_cluster, - (const OnigUChar *)ptr, (const OnigUChar *)end, - (const OnigUChar *)ptr, NULL, 0); - if (len <= 0) break; - grapheme_cluster_count++; - ptr += len; + OnigPosition len = onig_match(reg_grapheme_cluster, + (const OnigUChar *)ptr, (const OnigUChar *)end, + (const OnigUChar *)ptr, NULL, 0); + if (len <= 0) break; + grapheme_cluster_count++; + ptr += len; + } + + if (!cached_reg_grapheme_cluster) { + onig_free(reg_grapheme_cluster); } return SIZET2NUM(grapheme_cluster_count); @@ -9260,32 +9532,43 @@ static VALUE rb_str_enumerate_grapheme_clusters(VALUE str, VALUE ary) { VALUE orig = str; - regex_t *reg_grapheme_cluster = NULL; - rb_encoding *enc = rb_enc_from_index(ENCODING_GET(str)); + rb_encoding *enc = get_encoding(str); const char *ptr0, *ptr, *end; if (!rb_enc_unicode_p(enc)) { - return rb_str_enumerate_chars(str, ary); + return rb_str_enumerate_chars(str, ary); } if (!ary) str = rb_str_new_frozen(str); - reg_grapheme_cluster = get_reg_grapheme_cluster(enc); + + bool cached_reg_grapheme_cluster = true; + regex_t *reg_grapheme_cluster = get_cached_reg_grapheme_cluster(enc); + if (!reg_grapheme_cluster) { + reg_grapheme_cluster = get_reg_grapheme_cluster(enc); + cached_reg_grapheme_cluster = false; + } + ptr0 = ptr = RSTRING_PTR(str); end = RSTRING_END(str); while (ptr < end) { - OnigPosition len = onig_match(reg_grapheme_cluster, - (const OnigUChar *)ptr, (const OnigUChar *)end, - (const OnigUChar *)ptr, NULL, 0); - if (len <= 0) break; + OnigPosition len = onig_match(reg_grapheme_cluster, + (const OnigUChar *)ptr, (const OnigUChar *)end, + (const OnigUChar *)ptr, NULL, 0); + if (len <= 0) break; ENUM_ELEM(ary, rb_str_subseq(str, ptr-ptr0, len)); - ptr += len; + ptr += len; } + + if (!cached_reg_grapheme_cluster) { + onig_free(reg_grapheme_cluster); + } + RB_GC_GUARD(str); if (ary) - return ary; + return ary; else - return orig; + return orig; } /* @@ -9331,8 +9614,8 @@ chopped_length(VALUE str) p = rb_enc_prev_char(beg, end, end, enc); if (!p) return 0; if (p > beg && rb_enc_ascget(p, end, 0, enc) == '\n') { - p2 = rb_enc_prev_char(beg, p, end, enc); - if (p2 && rb_enc_ascget(p2, end, 0, enc) == '\r') p = p2; + p2 = rb_enc_prev_char(beg, p, end, enc); + if (p2 && rb_enc_ascget(p2, end, 0, enc) == '\r') p = p2; } return p - beg; } @@ -9352,14 +9635,14 @@ rb_str_chop_bang(VALUE str) { str_modify_keep_cr(str); if (RSTRING_LEN(str) > 0) { - long len; - len = chopped_length(str); - STR_SET_LEN(str, len); - TERM_FILL(&RSTRING_PTR(str)[len], TERM_LEN(str)); - if (ENC_CODERANGE(str) != ENC_CODERANGE_7BIT) { - ENC_CODERANGE_CLEAR(str); - } - return str; + long len; + len = chopped_length(str); + STR_SET_LEN(str, len); + TERM_FILL(&RSTRING_PTR(str)[len], TERM_LEN(str)); + if (ENC_CODERANGE(str) != ENC_CODERANGE_7BIT) { + ENC_CODERANGE_CLEAR(str); + } + return str; } return Qnil; } @@ -9430,55 +9713,55 @@ chompped_length(VALUE str, VALUE rs) enc = rb_enc_get(str); RSTRING_GETMEM(rs, rsptr, rslen); if (rslen == 0) { - if (rb_enc_mbminlen(enc) > 1) { - while (e > p) { - pp = rb_enc_left_char_head(p, e-rb_enc_mbminlen(enc), e, enc); - if (!rb_enc_is_newline(pp, e, enc)) break; - e = pp; - pp -= rb_enc_mbminlen(enc); - if (pp >= p) { - pp = rb_enc_left_char_head(p, pp, e, enc); - if (rb_enc_ascget(pp, e, 0, enc) == '\r') { - e = pp; - } - } - } - } - else { - while (e > p && *(e-1) == '\n') { - --e; - if (e > p && *(e-1) == '\r') - --e; - } - } - return e - p; + if (rb_enc_mbminlen(enc) > 1) { + while (e > p) { + pp = rb_enc_left_char_head(p, e-rb_enc_mbminlen(enc), e, enc); + if (!rb_enc_is_newline(pp, e, enc)) break; + e = pp; + pp -= rb_enc_mbminlen(enc); + if (pp >= p) { + pp = rb_enc_left_char_head(p, pp, e, enc); + if (rb_enc_ascget(pp, e, 0, enc) == '\r') { + e = pp; + } + } + } + } + else { + while (e > p && *(e-1) == '\n') { + --e; + if (e > p && *(e-1) == '\r') + --e; + } + } + return e - p; } if (rslen > len) return len; enc = rb_enc_get(rs); newline = rsptr[rslen-1]; if (rslen == rb_enc_mbminlen(enc)) { - if (rslen == 1) { - if (newline == '\n') + if (rslen == 1) { + if (newline == '\n') return smart_chomp(str, e, p); - } - else { - if (rb_enc_is_newline(rsptr, rsptr+rslen, enc)) + } + else { + if (rb_enc_is_newline(rsptr, rsptr+rslen, enc)) return smart_chomp(str, e, p); - } + } } enc = rb_enc_check(str, rs); if (is_broken_string(rs)) { - return len; + return len; } pp = e - rslen; if (p[len-1] == newline && - (rslen <= 1 || - memcmp(rsptr, pp, rslen) == 0)) { - if (rb_enc_left_char_head(p, pp, e, enc) == pp) - return len - rslen; - RB_GC_GUARD(rs); + (rslen <= 1 || + memcmp(rsptr, pp, rslen) == 0)) { + if (at_char_boundary(p, pp, e, enc)) + return len - rslen; + RB_GC_GUARD(rs); } return len; } @@ -9493,12 +9776,12 @@ chomp_rs(int argc, const VALUE *argv) { rb_check_arity(argc, 0, 1); if (argc > 0) { - VALUE rs = argv[0]; - if (!NIL_P(rs)) StringValue(rs); - return rs; + VALUE rs = argv[0]; + if (!NIL_P(rs)) StringValue(rs); + return rs; } else { - return rb_rs; + return rb_rs; } } @@ -9512,7 +9795,7 @@ rb_str_chomp_string(VALUE str, VALUE rs) STR_SET_LEN(str, len); TERM_FILL(&RSTRING_PTR(str)[len], TERM_LEN(str)); if (ENC_CODERANGE(str) != ENC_CODERANGE_7BIT) { - ENC_CODERANGE_CLEAR(str); + ENC_CODERANGE_CLEAR(str); } return str; } @@ -9531,7 +9814,7 @@ rb_str_chomp_bang(int argc, VALUE *argv, VALUE str) { VALUE rs; str_modifiable(str); - if (RSTRING_LEN(str) == 0) return Qnil; + if (RSTRING_LEN(str) == 0 && argc < 2) return Qnil; rs = chomp_rs(argc, argv); if (NIL_P(rs)) return Qnil; return rb_str_chomp_string(str, rs); @@ -9566,13 +9849,13 @@ lstrip_offset(VALUE str, const char *s, const char *e, rb_encoding *enc) while (s < e && (*s == '\0' || ascii_isspace(*s))) s++; } else { - while (s < e) { - int n; - unsigned int cc = rb_enc_codepoint_len(s, e, &n, enc); + while (s < e) { + int n; + unsigned int cc = rb_enc_codepoint_len(s, e, &n, enc); if (cc && !rb_isspace(cc)) break; - s += n; - } + s += n; + } } return s - start; } @@ -9599,12 +9882,12 @@ rb_str_lstrip_bang(VALUE str) RSTRING_GETMEM(str, start, olen); loffset = lstrip_offset(str, start, start+olen, enc); if (loffset > 0) { - long len = olen-loffset; - s = start + loffset; - memmove(start, s, len); - STR_SET_LEN(str, len); - TERM_FILL(start+len, rb_enc_mbminlen(enc)); - return str; + long len = olen-loffset; + s = start + loffset; + memmove(start, s, len); + STR_SET_LEN(str, len); + TERM_FILL(start+len, rb_enc_mbminlen(enc)); + return str; } return Qnil; } @@ -9642,22 +9925,25 @@ rstrip_offset(VALUE str, const char *s, const char *e, rb_encoding *enc) const char *t; rb_str_check_dummy_enc(enc); + if (rb_enc_str_coderange(str) == ENC_CODERANGE_BROKEN) { + rb_raise(rb_eEncCompatError, "invalid byte sequence in %s", rb_enc_name(enc)); + } if (!s || s >= e) return 0; t = e; /* remove trailing spaces or '\0's */ if (single_byte_optimizable(str)) { - unsigned char c; - while (s < t && ((c = *(t-1)) == '\0' || ascii_isspace(c))) t--; + unsigned char c; + while (s < t && ((c = *(t-1)) == '\0' || ascii_isspace(c))) t--; } else { - char *tp; + char *tp; while ((tp = rb_enc_prev_char(s, t, e, enc)) != NULL) { - unsigned int c = rb_enc_codepoint(tp, e, enc); - if (c && !rb_isspace(c)) break; - t = tp; - } + unsigned int c = rb_enc_codepoint(tp, e, enc); + if (c && !rb_isspace(c)) break; + t = tp; + } } return e - t; } @@ -9684,11 +9970,11 @@ rb_str_rstrip_bang(VALUE str) RSTRING_GETMEM(str, start, olen); roffset = rstrip_offset(str, start, start+olen, enc); if (roffset > 0) { - long len = olen - roffset; + long len = olen - roffset; - STR_SET_LEN(str, len); - TERM_FILL(start+len, rb_enc_mbminlen(enc)); - return str; + STR_SET_LEN(str, len); + TERM_FILL(start+len, rb_enc_mbminlen(enc)); + return str; } return Qnil; } @@ -9749,14 +10035,14 @@ rb_str_strip_bang(VALUE str) roffset = rstrip_offset(str, start+loffset, start+olen, enc); if (loffset > 0 || roffset > 0) { - long len = olen-roffset; - if (loffset > 0) { - len -= loffset; - memmove(start, start + loffset, len); - } - STR_SET_LEN(str, len); - TERM_FILL(start+len, rb_enc_mbminlen(enc)); - return str; + long len = olen-roffset; + if (loffset > 0) { + len -= loffset; + memmove(start, start + loffset, len); + } + STR_SET_LEN(str, len); + TERM_FILL(start+len, rb_enc_mbminlen(enc)); + return str; } return Qnil; } @@ -9795,51 +10081,57 @@ rb_str_strip(VALUE str) static VALUE scan_once(VALUE str, VALUE pat, long *start, int set_backref_str) { - VALUE result, match; - struct re_registers *regs; - int i; + VALUE result = Qnil; long end, pos = rb_pat_search(pat, str, *start, set_backref_str); if (pos >= 0) { - if (BUILTIN_TYPE(pat) == T_STRING) { - regs = NULL; - end = pos + RSTRING_LEN(pat); - } - else { - match = rb_backref_get(); - regs = RMATCH_REGS(match); - pos = BEG(0); - end = END(0); - } - if (pos == end) { - rb_encoding *enc = STR_ENC_GET(str); - /* - * Always consume at least one character of the input string - */ - if (RSTRING_LEN(str) > end) - *start = end + rb_enc_fast_mbclen(RSTRING_PTR(str) + end, - RSTRING_END(str), enc); - else - *start = end + 1; - } - else { - *start = end; - } - if (!regs || regs->num_regs == 1) { - result = rb_str_subseq(str, pos, end - pos); - return result; - } - result = rb_ary_new2(regs->num_regs); - for (i=1; i < regs->num_regs; i++) { - VALUE s = Qnil; - if (BEG(i) >= 0) { - s = rb_str_subseq(str, BEG(i), END(i)-BEG(i)); - } - rb_ary_push(result, s); - } - - return result; + VALUE match; + struct re_registers *regs; + if (BUILTIN_TYPE(pat) == T_STRING) { + regs = NULL; + end = pos + RSTRING_LEN(pat); + } + else { + match = rb_backref_get(); + regs = RMATCH_REGS(match); + pos = BEG(0); + end = END(0); + } + + if (pos == end) { + rb_encoding *enc = STR_ENC_GET(str); + /* + * Always consume at least one character of the input string + */ + if (RSTRING_LEN(str) > end) + *start = end + rb_enc_fast_mbclen(RSTRING_PTR(str) + end, + RSTRING_END(str), enc); + else + *start = end + 1; + } + else { + *start = end; + } + + if (!regs || regs->num_regs == 1) { + result = rb_str_subseq(str, pos, end - pos); + return result; + } + else { + result = rb_ary_new2(regs->num_regs); + for (int i = 1; i < regs->num_regs; i++) { + VALUE s = Qnil; + if (BEG(i) >= 0) { + s = rb_str_subseq(str, BEG(i), END(i)-BEG(i)); + } + + rb_ary_push(result, s); + } + } + + RB_GC_GUARD(match); } - return Qnil; + + return result; } @@ -9893,23 +10185,23 @@ rb_str_scan(VALUE str, VALUE pat) pat = get_pat_quoted(pat, 1); mustnot_broken(str); if (!rb_block_given_p()) { - VALUE ary = rb_ary_new(); + VALUE ary = rb_ary_new(); - while (!NIL_P(result = scan_once(str, pat, &start, 0))) { - last = prev; - prev = start; - rb_ary_push(ary, result); - } - if (last >= 0) rb_pat_search(pat, str, last, 1); - else rb_backref_set(Qnil); - return ary; + while (!NIL_P(result = scan_once(str, pat, &start, 0))) { + last = prev; + prev = start; + rb_ary_push(ary, result); + } + if (last >= 0) rb_pat_search(pat, str, last, 1); + else rb_backref_set(Qnil); + return ary; } while (!NIL_P(result = scan_once(str, pat, &start, 1))) { - last = prev; - prev = start; - rb_yield(result); - str_mod_check(str, p, len); + last = prev; + prev = start; + rb_yield(result); + str_mod_check(str, p, len); } if (last >= 0) rb_pat_search(pat, str, last, 1); return str; @@ -9950,9 +10242,9 @@ rb_str_hex(VALUE str) * returns zero if there is no such leading substring: * * '123'.oct # => 83 - '-377'.oct # => -255 - '0377non-numeric'.oct # => 255 - 'non-numeric'.oct # => 0 + * '-377'.oct # => -255 + * '0377non-numeric'.oct # => 255 + * 'non-numeric'.oct # => 0 * * If +self+ starts with <tt>0</tt>, radix indicators are honored; * see Kernel#Integer. @@ -10065,15 +10357,15 @@ rb_str_crypt(VALUE str, VALUE salt) s = StringValueCStr(str); saltp = RSTRING_PTR(salt); if (RSTRING_LEN(salt) < 2 || !saltp[0] || !saltp[1]) { - rb_raise(rb_eArgError, "salt too short (need >=2 bytes)"); + rb_raise(rb_eArgError, "salt too short (need >=2 bytes)"); } #ifdef BROKEN_CRYPT if (!ISASCII((unsigned char)saltp[0]) || !ISASCII((unsigned char)saltp[1])) { - salt_8bit_clean[0] = saltp[0] & 0x7f; - salt_8bit_clean[1] = saltp[1] & 0x7f; - salt_8bit_clean[2] = '\0'; - saltp = salt_8bit_clean; + salt_8bit_clean[0] = saltp[0] & 0x7f; + salt_8bit_clean[1] = saltp[1] & 0x7f; + salt_8bit_clean[2] = '\0'; + saltp = salt_8bit_clean; } #endif #ifdef HAVE_CRYPT_R @@ -10088,9 +10380,9 @@ rb_str_crypt(VALUE str, VALUE salt) res = crypt(s, saltp); #endif if (!res) { - int err = errno; - CRYPT_END(); - rb_syserr_fail(err, "crypt"); + int err = errno; + CRYPT_END(); + rb_syserr_fail(err, "crypt"); } result = rb_str_new_cstr(res); CRYPT_END(); @@ -10194,15 +10486,15 @@ rb_str_justify(int argc, VALUE *argv, VALUE str, char jflag) termlen = rb_enc_mbminlen(enc); width = NUM2LONG(w); if (argc == 2) { - StringValue(pad); - enc = rb_enc_check(str, pad); - f = RSTRING_PTR(pad); - flen = RSTRING_LEN(pad); - fclen = str_strlen(pad, enc); /* rb_enc_check */ - singlebyte = single_byte_optimizable(pad); - if (flen == 0 || fclen == 0) { - rb_raise(rb_eArgError, "zero width padding"); - } + StringValue(pad); + enc = rb_enc_check(str, pad); + f = RSTRING_PTR(pad); + flen = RSTRING_LEN(pad); + fclen = str_strlen(pad, enc); /* rb_enc_check */ + singlebyte = single_byte_optimizable(pad); + if (flen == 0 || fclen == 0) { + rb_raise(rb_eArgError, "zero width padding"); + } } len = str_strlen(str, enc); /* rb_enc_check */ if (width < 0 || len >= width) return str_duplicate(rb_cString, str); @@ -10229,14 +10521,14 @@ rb_str_justify(int argc, VALUE *argv, VALUE str, char jflag) } else { while (llen >= fclen) { - memcpy(p,f,flen); - p += flen; - llen -= fclen; - } + memcpy(p,f,flen); + p += flen; + llen -= fclen; + } if (llen > 0) { memcpy(p, f, llen2); p += llen2; - } + } } memcpy(p, RSTRING_PTR(str), size); p += size; @@ -10246,22 +10538,22 @@ rb_str_justify(int argc, VALUE *argv, VALUE str, char jflag) } else { while (rlen >= fclen) { - memcpy(p,f,flen); - p += flen; - rlen -= fclen; - } + memcpy(p,f,flen); + p += flen; + rlen -= fclen; + } if (rlen > 0) { memcpy(p, f, rlen2); p += rlen2; - } + } } TERM_FILL(p, termlen); STR_SET_LEN(res, p-RSTRING_PTR(res)); rb_enc_associate(res, enc); if (argc == 2) - cr = ENC_CODERANGE_AND(cr, ENC_CODERANGE(pad)); + cr = ENC_CODERANGE_AND(cr, ENC_CODERANGE(pad)); if (cr != ENC_CODERANGE_BROKEN) - ENC_CODERANGE_SET(res, cr); + ENC_CODERANGE_SET(res, cr); RB_GC_GUARD(pad); return res; @@ -10332,23 +10624,23 @@ rb_str_partition(VALUE str, VALUE sep) sep = get_pat_quoted(sep, 0); if (RB_TYPE_P(sep, T_REGEXP)) { - if (rb_reg_search(sep, str, 0, 0) < 0) { + if (rb_reg_search(sep, str, 0, 0) < 0) { goto failed; - } - VALUE match = rb_backref_get(); - struct re_registers *regs = RMATCH_REGS(match); + } + VALUE match = rb_backref_get(); + struct re_registers *regs = RMATCH_REGS(match); pos = BEG(0); - sep = rb_str_subseq(str, pos, END(0) - pos); + sep = rb_str_subseq(str, pos, END(0) - pos); } else { - pos = rb_str_index(str, sep, 0); - if (pos < 0) goto failed; + pos = rb_str_index(str, sep, 0); + if (pos < 0) goto failed; } return rb_ary_new3(3, rb_str_subseq(str, 0, pos), - sep, - rb_str_subseq(str, pos+RSTRING_LEN(sep), - RSTRING_LEN(str)-pos-RSTRING_LEN(sep))); + sep, + rb_str_subseq(str, pos+RSTRING_LEN(sep), + RSTRING_LEN(str)-pos-RSTRING_LEN(sep))); failed: return rb_ary_new3(3, str_duplicate(rb_cString, str), str_new_empty_String(str), str_new_empty_String(str)); @@ -10373,24 +10665,23 @@ rb_str_rpartition(VALUE str, VALUE sep) goto failed; } VALUE match = rb_backref_get(); - struct re_registers *regs = RMATCH_REGS(match); + struct re_registers *regs = RMATCH_REGS(match); pos = BEG(0); sep = rb_str_subseq(str, pos, END(0) - pos); } else { - pos = rb_str_sublen(str, pos); - pos = rb_str_rindex(str, sep, pos); + pos = rb_str_sublen(str, pos); + pos = rb_str_rindex(str, sep, pos); if (pos < 0) { goto failed; } - pos = rb_str_offset(str, pos); } return rb_ary_new3(3, rb_str_subseq(str, 0, pos), - sep, - rb_str_subseq(str, pos+RSTRING_LEN(sep), - RSTRING_LEN(str)-pos-RSTRING_LEN(sep))); + sep, + rb_str_subseq(str, pos+RSTRING_LEN(sep), + RSTRING_LEN(str)-pos-RSTRING_LEN(sep))); failed: return rb_ary_new3(3, str_new_empty_String(str), str_new_empty_String(str), str_duplicate(rb_cString, str)); } @@ -10409,18 +10700,28 @@ rb_str_start_with(int argc, VALUE *argv, VALUE str) int i; for (i=0; i<argc; i++) { - VALUE tmp = argv[i]; - if (RB_TYPE_P(tmp, T_REGEXP)) { - if (rb_reg_start_with_p(tmp, str)) - return Qtrue; - } - else { - StringValue(tmp); - rb_enc_check(str, tmp); - if (RSTRING_LEN(str) < RSTRING_LEN(tmp)) continue; - if (memcmp(RSTRING_PTR(str), RSTRING_PTR(tmp), RSTRING_LEN(tmp)) == 0) - return Qtrue; - } + VALUE tmp = argv[i]; + if (RB_TYPE_P(tmp, T_REGEXP)) { + if (rb_reg_start_with_p(tmp, str)) + return Qtrue; + } + else { + const char *p, *s, *e; + long slen, tlen; + rb_encoding *enc; + + StringValue(tmp); + enc = rb_enc_check(str, tmp); + if ((tlen = RSTRING_LEN(tmp)) == 0) return Qtrue; + if ((slen = RSTRING_LEN(str)) < tlen) continue; + p = RSTRING_PTR(str); + e = p + slen; + s = p + tlen; + if (!at_char_right_boundary(p, s, e, enc)) + continue; + if (memcmp(p, RSTRING_PTR(tmp), tlen) == 0) + return Qtrue; + } } return Qfalse; } @@ -10437,23 +10738,24 @@ static VALUE rb_str_end_with(int argc, VALUE *argv, VALUE str) { int i; - char *p, *s, *e; - rb_encoding *enc; for (i=0; i<argc; i++) { - VALUE tmp = argv[i]; - long slen, tlen; - StringValue(tmp); - enc = rb_enc_check(str, tmp); - if ((tlen = RSTRING_LEN(tmp)) == 0) return Qtrue; - if ((slen = RSTRING_LEN(str)) < tlen) continue; - p = RSTRING_PTR(str); + VALUE tmp = argv[i]; + const char *p, *s, *e; + long slen, tlen; + rb_encoding *enc; + + StringValue(tmp); + enc = rb_enc_check(str, tmp); + if ((tlen = RSTRING_LEN(tmp)) == 0) return Qtrue; + if ((slen = RSTRING_LEN(str)) < tlen) continue; + p = RSTRING_PTR(str); e = p + slen; - s = e - tlen; - if (rb_enc_left_char_head(p, s, e, enc) != s) - continue; - if (memcmp(s, RSTRING_PTR(tmp), RSTRING_LEN(tmp)) == 0) - return Qtrue; + s = e - tlen; + if (!at_char_boundary(p, s, e, enc)) + continue; + if (memcmp(s, RSTRING_PTR(tmp), tlen) == 0) + return Qtrue; } return Qfalse; } @@ -10470,12 +10772,17 @@ rb_str_end_with(int argc, VALUE *argv, VALUE str) static long deleted_prefix_length(VALUE str, VALUE prefix) { - char *strptr, *prefixptr; + const char *strptr, *prefixptr; long olen, prefixlen; + rb_encoding *enc = rb_enc_get(str); StringValue(prefix); - if (is_broken_string(prefix)) return 0; - rb_enc_check(str, prefix); + + if (!is_broken_string(prefix) || + !rb_enc_asciicompat(enc) || + !rb_enc_asciicompat(rb_enc_get(prefix))) { + enc = rb_enc_check(str, prefix); + } /* return 0 if not start with prefix */ prefixlen = RSTRING_LEN(prefix); @@ -10485,6 +10792,19 @@ deleted_prefix_length(VALUE str, VALUE prefix) strptr = RSTRING_PTR(str); prefixptr = RSTRING_PTR(prefix); if (memcmp(strptr, prefixptr, prefixlen) != 0) return 0; + if (is_broken_string(prefix)) { + if (!is_broken_string(str)) { + /* prefix in a valid string cannot be broken */ + return 0; + } + const char *strend = strptr + olen; + const char *after_prefix = strptr + prefixlen; + if (!at_char_right_boundary(strptr, after_prefix, strend, enc)) { + /* prefix does not end at char-boundary */ + return 0; + } + } + /* prefix part in `str` also should be valid. */ return prefixlen; } @@ -10541,7 +10861,7 @@ rb_str_delete_prefix(VALUE str, VALUE prefix) static long deleted_suffix_length(VALUE str, VALUE suffix) { - char *strptr, *suffixptr, *s; + const char *strptr, *suffixptr; long olen, suffixlen; rb_encoding *enc; @@ -10556,9 +10876,10 @@ deleted_suffix_length(VALUE str, VALUE suffix) if (olen < suffixlen) return 0; strptr = RSTRING_PTR(str); suffixptr = RSTRING_PTR(suffix); - s = strptr + olen - suffixlen; - if (memcmp(s, suffixptr, suffixlen) != 0) return 0; - if (rb_enc_left_char_head(strptr, s, strptr + olen, enc) != s) return 0; + const char *strend = strptr + olen; + const char *before_suffix = strend - suffixlen; + if (memcmp(before_suffix, suffixptr, suffixlen) != 0) return 0; + if (!at_char_boundary(strptr, before_suffix, strend, enc)) return 0; return suffixlen; } @@ -10587,7 +10908,7 @@ rb_str_delete_suffix_bang(VALUE str, VALUE suffix) STR_SET_LEN(str, len); TERM_FILL(&RSTRING_PTR(str)[len], TERM_LEN(str)); if (ENC_CODERANGE(str) != ENC_CODERANGE_7BIT) { - ENC_CODERANGE_CLEAR(str); + ENC_CODERANGE_CLEAR(str); } return str; } @@ -10615,7 +10936,7 @@ void rb_str_setter(VALUE val, ID id, VALUE *var) { if (!NIL_P(val) && !RB_TYPE_P(val, T_STRING)) { - rb_raise(rb_eTypeError, "value of %"PRIsVALUE" must be String", rb_id2str(id)); + rb_raise(rb_eTypeError, "value of %"PRIsVALUE" must be String", rb_id2str(id)); } *var = val; } @@ -10625,12 +10946,12 @@ rb_fs_setter(VALUE val, ID id, VALUE *var) { val = rb_fs_check(val); if (!val) { - rb_raise(rb_eTypeError, - "value of %"PRIsVALUE" must be String or Regexp", - rb_id2str(id)); + rb_raise(rb_eTypeError, + "value of %"PRIsVALUE" must be String or Regexp", + rb_id2str(id)); } if (!NIL_P(val)) { - rb_warn_deprecated("`$;'", NULL); + rb_warn_deprecated("'$;'", NULL); } *var = val; } @@ -10648,7 +10969,23 @@ static VALUE rb_str_force_encoding(VALUE str, VALUE enc) { str_modifiable(str); - rb_enc_associate(str, rb_to_encoding(enc)); + + rb_encoding *encoding = rb_to_encoding(enc); + int idx = rb_enc_to_index(encoding); + + // If the encoding is unchanged, we do nothing. + if (ENCODING_GET(str) == idx) { + return str; + } + + rb_enc_associate_index(str, idx); + + // If the coderange was 7bit and the new encoding is ASCII-compatible + // we can keep the coderange. + if (ENC_CODERANGE(str) == ENC_CODERANGE_7BIT && encoding && rb_enc_asciicompat(encoding)) { + return str; + } + ENC_CODERANGE_CLEAR(str); return str; } @@ -10665,14 +11002,32 @@ static VALUE rb_str_b(VALUE str) { VALUE str2; - if (FL_TEST(str, STR_NOEMBED)) { - str2 = str_alloc_heap(rb_cString); + if (STR_EMBED_P(str)) { + str2 = str_alloc_embed(rb_cString, RSTRING_LEN(str) + TERM_LEN(str)); } else { - str2 = str_alloc_embed(rb_cString, RSTRING_EMBED_LEN(str) + TERM_LEN(str)); + str2 = str_alloc_heap(rb_cString); } str_replace_shared_without_enc(str2, str); - ENC_CODERANGE_CLEAR(str2); + + if (rb_enc_asciicompat(STR_ENC_GET(str))) { + // BINARY strings can never be broken; they're either 7-bit ASCII or VALID. + // If we know the receiver's code range then we know the result's code range. + int cr = ENC_CODERANGE(str); + switch (cr) { + case ENC_CODERANGE_7BIT: + ENC_CODERANGE_SET(str2, ENC_CODERANGE_7BIT); + break; + case ENC_CODERANGE_BROKEN: + case ENC_CODERANGE_VALID: + ENC_CODERANGE_SET(str2, ENC_CODERANGE_VALID); + break; + default: + ENC_CODERANGE_CLEAR(str2); + break; + } + } + return str2; } @@ -10727,27 +11082,27 @@ rb_str_ellipsize(VALUE str, long len) if (len < 0) rb_raise(rb_eIndexError, "negative length %ld", len); if (len * rb_enc_mbminlen(enc) >= blen || - (e = rb_enc_nth(p, e, len, enc)) - p == blen) { - ret = str; + (e = rb_enc_nth(p, e, len, enc)) - p == blen) { + ret = str; } else if (len <= ellipsislen || - !(e = rb_enc_step_back(p, e, e, len = ellipsislen, enc))) { - if (rb_enc_asciicompat(enc)) { + !(e = rb_enc_step_back(p, e, e, len = ellipsislen, enc))) { + if (rb_enc_asciicompat(enc)) { ret = rb_str_new(ellipsis, len); - rb_enc_associate(ret, enc); - } - else { - estr = rb_usascii_str_new(ellipsis, len); - ret = rb_str_encode(estr, rb_enc_from_encoding(enc), 0, Qnil); - } + rb_enc_associate(ret, enc); + } + else { + estr = rb_usascii_str_new(ellipsis, len); + ret = rb_str_encode(estr, rb_enc_from_encoding(enc), 0, Qnil); + } } else if (ret = rb_str_subseq(str, 0, e - p), rb_enc_asciicompat(enc)) { - rb_str_cat(ret, ellipsis, ellipsislen); + rb_str_cat(ret, ellipsis, ellipsislen); } else { - estr = rb_str_encode(rb_usascii_str_new(ellipsis, ellipsislen), - rb_enc_from_encoding(enc), 0, Qnil); - rb_str_append(ret, estr); + estr = rb_str_encode(rb_usascii_str_new(ellipsis, ellipsislen), + rb_enc_from_encoding(enc), 0, Qnil); + rb_str_append(ret, estr); } return ret; } @@ -10759,14 +11114,14 @@ str_compat_and_valid(VALUE str, rb_encoding *enc) str = StringValue(str); cr = rb_enc_str_coderange(str); if (cr == ENC_CODERANGE_BROKEN) { - rb_raise(rb_eArgError, "replacement must be valid byte sequence '%+"PRIsVALUE"'", str); + rb_raise(rb_eArgError, "replacement must be valid byte sequence '%+"PRIsVALUE"'", str); } else { - rb_encoding *e = STR_ENC_GET(str); - if (cr == ENC_CODERANGE_7BIT ? rb_enc_mbminlen(enc) != 1 : enc != e) { - rb_raise(rb_eEncCompatError, "incompatible character encodings: %s and %s", - rb_enc_name(enc), rb_enc_name(e)); - } + rb_encoding *e = STR_ENC_GET(str); + if (cr == ENC_CODERANGE_7BIT ? rb_enc_mbminlen(enc) != 1 : enc != e) { + rb_raise(rb_eEncCompatError, "incompatible character encodings: %s and %s", + rb_enc_inspect_name(enc), rb_enc_inspect_name(e)); + } } return str; } @@ -10785,9 +11140,9 @@ rb_enc_str_scrub(rb_encoding *enc, VALUE str, VALUE repl) { int cr = ENC_CODERANGE_UNKNOWN; if (enc == STR_ENC_GET(str)) { - /* cached coderange makes sense only when enc equals the - * actual encoding of str */ - cr = ENC_CODERANGE(str); + /* cached coderange makes sense only when enc equals the + * actual encoding of str */ + cr = ENC_CODERANGE(str); } return enc_str_scrub(enc, str, repl, cr); } @@ -10802,26 +11157,26 @@ enc_str_scrub(rb_encoding *enc, VALUE str, VALUE repl, int cr) long slen; if (rb_block_given_p()) { - if (!NIL_P(repl)) - rb_raise(rb_eArgError, "both of block and replacement given"); - replen = 0; + if (!NIL_P(repl)) + rb_raise(rb_eArgError, "both of block and replacement given"); + replen = 0; } if (ENC_CODERANGE_CLEAN_P(cr)) - return Qnil; + return Qnil; if (!NIL_P(repl)) { - repl = str_compat_and_valid(repl, enc); + repl = str_compat_and_valid(repl, enc); } if (rb_enc_dummy_p(enc)) { - return Qnil; + return Qnil; } encidx = rb_enc_to_index(enc); #define DEFAULT_REPLACE_CHAR(str) do { \ - static const char replace[sizeof(str)-1] = str; \ - rep = replace; replen = (int)sizeof(replace); \ + static const char replace[sizeof(str)-1] = str; \ + rep = replace; replen = (int)sizeof(replace); \ } while (0) slen = RSTRING_LEN(str); @@ -10831,204 +11186,204 @@ enc_str_scrub(rb_encoding *enc, VALUE str, VALUE repl, int cr) sp = p; if (rb_enc_asciicompat(enc)) { - int rep7bit_p; - if (!replen) { - rep = NULL; - rep7bit_p = FALSE; - } - else if (!NIL_P(repl)) { - rep = RSTRING_PTR(repl); - replen = RSTRING_LEN(repl); - rep7bit_p = (ENC_CODERANGE(repl) == ENC_CODERANGE_7BIT); - } - else if (encidx == rb_utf8_encindex()) { - DEFAULT_REPLACE_CHAR("\xEF\xBF\xBD"); - rep7bit_p = FALSE; - } - else { - DEFAULT_REPLACE_CHAR("?"); - rep7bit_p = TRUE; - } - cr = ENC_CODERANGE_7BIT; - - p = search_nonascii(p, e); - if (!p) { - p = e; - } - while (p < e) { - int ret = rb_enc_precise_mbclen(p, e, enc); - if (MBCLEN_NEEDMORE_P(ret)) { - break; - } - else if (MBCLEN_CHARFOUND_P(ret)) { - cr = ENC_CODERANGE_VALID; - p += MBCLEN_CHARFOUND_LEN(ret); - } - else if (MBCLEN_INVALID_P(ret)) { - /* - * p1~p: valid ascii/multibyte chars - * p ~e: invalid bytes + unknown bytes - */ - long clen = rb_enc_mbmaxlen(enc); - if (NIL_P(buf)) buf = rb_str_buf_new(RSTRING_LEN(str)); - if (p > p1) { - rb_str_buf_cat(buf, p1, p - p1); - } - - if (e - p < clen) clen = e - p; - if (clen <= 2) { - clen = 1; - } - else { - const char *q = p; - clen--; - for (; clen > 1; clen--) { - ret = rb_enc_precise_mbclen(q, q + clen, enc); - if (MBCLEN_NEEDMORE_P(ret)) break; - if (MBCLEN_INVALID_P(ret)) continue; - UNREACHABLE; - } - } - if (rep) { - rb_str_buf_cat(buf, rep, replen); - if (!rep7bit_p) cr = ENC_CODERANGE_VALID; - } - else { - repl = rb_yield(rb_enc_str_new(p, clen, enc)); + int rep7bit_p; + if (!replen) { + rep = NULL; + rep7bit_p = FALSE; + } + else if (!NIL_P(repl)) { + rep = RSTRING_PTR(repl); + replen = RSTRING_LEN(repl); + rep7bit_p = (ENC_CODERANGE(repl) == ENC_CODERANGE_7BIT); + } + else if (encidx == rb_utf8_encindex()) { + DEFAULT_REPLACE_CHAR("\xEF\xBF\xBD"); + rep7bit_p = FALSE; + } + else { + DEFAULT_REPLACE_CHAR("?"); + rep7bit_p = TRUE; + } + cr = ENC_CODERANGE_7BIT; + + p = search_nonascii(p, e); + if (!p) { + p = e; + } + while (p < e) { + int ret = rb_enc_precise_mbclen(p, e, enc); + if (MBCLEN_NEEDMORE_P(ret)) { + break; + } + else if (MBCLEN_CHARFOUND_P(ret)) { + cr = ENC_CODERANGE_VALID; + p += MBCLEN_CHARFOUND_LEN(ret); + } + else if (MBCLEN_INVALID_P(ret)) { + /* + * p1~p: valid ascii/multibyte chars + * p ~e: invalid bytes + unknown bytes + */ + long clen = rb_enc_mbmaxlen(enc); + if (NIL_P(buf)) buf = rb_str_buf_new(RSTRING_LEN(str)); + if (p > p1) { + rb_str_buf_cat(buf, p1, p - p1); + } + + if (e - p < clen) clen = e - p; + if (clen <= 2) { + clen = 1; + } + else { + const char *q = p; + clen--; + for (; clen > 1; clen--) { + ret = rb_enc_precise_mbclen(q, q + clen, enc); + if (MBCLEN_NEEDMORE_P(ret)) break; + if (MBCLEN_INVALID_P(ret)) continue; + UNREACHABLE; + } + } + if (rep) { + rb_str_buf_cat(buf, rep, replen); + if (!rep7bit_p) cr = ENC_CODERANGE_VALID; + } + else { + repl = rb_yield(rb_enc_str_new(p, clen, enc)); str_mod_check(str, sp, slen); - repl = str_compat_and_valid(repl, enc); - rb_str_buf_cat(buf, RSTRING_PTR(repl), RSTRING_LEN(repl)); - if (ENC_CODERANGE(repl) == ENC_CODERANGE_VALID) - cr = ENC_CODERANGE_VALID; - } - p += clen; - p1 = p; - p = search_nonascii(p, e); - if (!p) { - p = e; - break; - } - } - else { - UNREACHABLE; - } - } - if (NIL_P(buf)) { - if (p == e) { - ENC_CODERANGE_SET(str, cr); - return Qnil; - } - buf = rb_str_buf_new(RSTRING_LEN(str)); - } - if (p1 < p) { - rb_str_buf_cat(buf, p1, p - p1); - } - if (p < e) { - if (rep) { - rb_str_buf_cat(buf, rep, replen); - if (!rep7bit_p) cr = ENC_CODERANGE_VALID; - } - else { - repl = rb_yield(rb_enc_str_new(p, e-p, enc)); + repl = str_compat_and_valid(repl, enc); + rb_str_buf_cat(buf, RSTRING_PTR(repl), RSTRING_LEN(repl)); + if (ENC_CODERANGE(repl) == ENC_CODERANGE_VALID) + cr = ENC_CODERANGE_VALID; + } + p += clen; + p1 = p; + p = search_nonascii(p, e); + if (!p) { + p = e; + break; + } + } + else { + UNREACHABLE; + } + } + if (NIL_P(buf)) { + if (p == e) { + ENC_CODERANGE_SET(str, cr); + return Qnil; + } + buf = rb_str_buf_new(RSTRING_LEN(str)); + } + if (p1 < p) { + rb_str_buf_cat(buf, p1, p - p1); + } + if (p < e) { + if (rep) { + rb_str_buf_cat(buf, rep, replen); + if (!rep7bit_p) cr = ENC_CODERANGE_VALID; + } + else { + repl = rb_yield(rb_enc_str_new(p, e-p, enc)); str_mod_check(str, sp, slen); - repl = str_compat_and_valid(repl, enc); - rb_str_buf_cat(buf, RSTRING_PTR(repl), RSTRING_LEN(repl)); - if (ENC_CODERANGE(repl) == ENC_CODERANGE_VALID) - cr = ENC_CODERANGE_VALID; - } - } + repl = str_compat_and_valid(repl, enc); + rb_str_buf_cat(buf, RSTRING_PTR(repl), RSTRING_LEN(repl)); + if (ENC_CODERANGE(repl) == ENC_CODERANGE_VALID) + cr = ENC_CODERANGE_VALID; + } + } } else { - /* ASCII incompatible */ - long mbminlen = rb_enc_mbminlen(enc); - if (!replen) { - rep = NULL; - } - else if (!NIL_P(repl)) { - rep = RSTRING_PTR(repl); - replen = RSTRING_LEN(repl); - } - else if (encidx == ENCINDEX_UTF_16BE) { - DEFAULT_REPLACE_CHAR("\xFF\xFD"); - } - else if (encidx == ENCINDEX_UTF_16LE) { - DEFAULT_REPLACE_CHAR("\xFD\xFF"); - } - else if (encidx == ENCINDEX_UTF_32BE) { - DEFAULT_REPLACE_CHAR("\x00\x00\xFF\xFD"); - } - else if (encidx == ENCINDEX_UTF_32LE) { - DEFAULT_REPLACE_CHAR("\xFD\xFF\x00\x00"); - } - else { - DEFAULT_REPLACE_CHAR("?"); - } - - while (p < e) { - int ret = rb_enc_precise_mbclen(p, e, enc); - if (MBCLEN_NEEDMORE_P(ret)) { - break; - } - else if (MBCLEN_CHARFOUND_P(ret)) { - p += MBCLEN_CHARFOUND_LEN(ret); - } - else if (MBCLEN_INVALID_P(ret)) { - const char *q = p; - long clen = rb_enc_mbmaxlen(enc); - if (NIL_P(buf)) buf = rb_str_buf_new(RSTRING_LEN(str)); - if (p > p1) rb_str_buf_cat(buf, p1, p - p1); - - if (e - p < clen) clen = e - p; - if (clen <= mbminlen * 2) { - clen = mbminlen; - } - else { - clen -= mbminlen; - for (; clen > mbminlen; clen-=mbminlen) { - ret = rb_enc_precise_mbclen(q, q + clen, enc); - if (MBCLEN_NEEDMORE_P(ret)) break; - if (MBCLEN_INVALID_P(ret)) continue; - UNREACHABLE; - } - } - if (rep) { - rb_str_buf_cat(buf, rep, replen); - } - else { - repl = rb_yield(rb_enc_str_new(p, clen, enc)); + /* ASCII incompatible */ + long mbminlen = rb_enc_mbminlen(enc); + if (!replen) { + rep = NULL; + } + else if (!NIL_P(repl)) { + rep = RSTRING_PTR(repl); + replen = RSTRING_LEN(repl); + } + else if (encidx == ENCINDEX_UTF_16BE) { + DEFAULT_REPLACE_CHAR("\xFF\xFD"); + } + else if (encidx == ENCINDEX_UTF_16LE) { + DEFAULT_REPLACE_CHAR("\xFD\xFF"); + } + else if (encidx == ENCINDEX_UTF_32BE) { + DEFAULT_REPLACE_CHAR("\x00\x00\xFF\xFD"); + } + else if (encidx == ENCINDEX_UTF_32LE) { + DEFAULT_REPLACE_CHAR("\xFD\xFF\x00\x00"); + } + else { + DEFAULT_REPLACE_CHAR("?"); + } + + while (p < e) { + int ret = rb_enc_precise_mbclen(p, e, enc); + if (MBCLEN_NEEDMORE_P(ret)) { + break; + } + else if (MBCLEN_CHARFOUND_P(ret)) { + p += MBCLEN_CHARFOUND_LEN(ret); + } + else if (MBCLEN_INVALID_P(ret)) { + const char *q = p; + long clen = rb_enc_mbmaxlen(enc); + if (NIL_P(buf)) buf = rb_str_buf_new(RSTRING_LEN(str)); + if (p > p1) rb_str_buf_cat(buf, p1, p - p1); + + if (e - p < clen) clen = e - p; + if (clen <= mbminlen * 2) { + clen = mbminlen; + } + else { + clen -= mbminlen; + for (; clen > mbminlen; clen-=mbminlen) { + ret = rb_enc_precise_mbclen(q, q + clen, enc); + if (MBCLEN_NEEDMORE_P(ret)) break; + if (MBCLEN_INVALID_P(ret)) continue; + UNREACHABLE; + } + } + if (rep) { + rb_str_buf_cat(buf, rep, replen); + } + else { + repl = rb_yield(rb_enc_str_new(p, clen, enc)); str_mod_check(str, sp, slen); - repl = str_compat_and_valid(repl, enc); - rb_str_buf_cat(buf, RSTRING_PTR(repl), RSTRING_LEN(repl)); - } - p += clen; - p1 = p; - } - else { - UNREACHABLE; - } - } - if (NIL_P(buf)) { - if (p == e) { - ENC_CODERANGE_SET(str, ENC_CODERANGE_VALID); - return Qnil; - } - buf = rb_str_buf_new(RSTRING_LEN(str)); - } - if (p1 < p) { - rb_str_buf_cat(buf, p1, p - p1); - } - if (p < e) { - if (rep) { - rb_str_buf_cat(buf, rep, replen); - } - else { - repl = rb_yield(rb_enc_str_new(p, e-p, enc)); + repl = str_compat_and_valid(repl, enc); + rb_str_buf_cat(buf, RSTRING_PTR(repl), RSTRING_LEN(repl)); + } + p += clen; + p1 = p; + } + else { + UNREACHABLE; + } + } + if (NIL_P(buf)) { + if (p == e) { + ENC_CODERANGE_SET(str, ENC_CODERANGE_VALID); + return Qnil; + } + buf = rb_str_buf_new(RSTRING_LEN(str)); + } + if (p1 < p) { + rb_str_buf_cat(buf, p1, p - p1); + } + if (p < e) { + if (rep) { + rb_str_buf_cat(buf, rep, replen); + } + else { + repl = rb_yield(rb_enc_str_new(p, e-p, enc)); str_mod_check(str, sp, slen); - repl = str_compat_and_valid(repl, enc); - rb_str_buf_cat(buf, RSTRING_PTR(repl), RSTRING_LEN(repl)); - } - } - cr = ENC_CODERANGE_VALID; + repl = str_compat_and_valid(repl, enc); + rb_str_buf_cat(buf, RSTRING_PTR(repl), RSTRING_LEN(repl)); + } + } + cr = ENC_CODERANGE_VALID; } ENCODING_CODERANGE_SET(buf, rb_enc_to_index(enc), cr); return buf; @@ -11079,8 +11434,8 @@ unicode_normalize_common(int argc, VALUE *argv, VALUE str, ID id) VALUE argv2[2]; if (!UnicodeNormalizeRequired) { - rb_require("unicode_normalize/normalize.rb"); - UnicodeNormalizeRequired = 1; + rb_require("unicode_normalize/normalize.rb"); + UnicodeNormalizeRequired = 1; } argv2[0] = str; if (rb_check_arity(argc, 0, 1)) argv2[1] = argv[0]; @@ -11174,17 +11529,17 @@ rb_str_unicode_normalized_p(int argc, VALUE *argv, VALUE str) /********************************************************************** * Document-class: Symbol * - * Symbol objects represent named identifiers inside the Ruby interpreter. + * A +Symbol+ object represents a named identifier inside the Ruby interpreter. * - * You can create a \Symbol object explicitly with: + * You can create a +Symbol+ object explicitly with: * * - A {symbol literal}[rdoc-ref:syntax/literals.rdoc@Symbol+Literals]. * - * The same Symbol object will be + * The same +Symbol+ object will be * created for a given name or string for the duration of a program's * execution, regardless of the context or meaning of that name. Thus * if <code>Fred</code> is a constant in one context, a method in - * another, and a class in a third, the Symbol <code>:Fred</code> + * another, and a class in a third, the +Symbol+ <code>:Fred</code> * will be the same object in all three contexts. * * module One @@ -11227,18 +11582,18 @@ rb_str_unicode_normalized_p(int argc, VALUE *argv, VALUE str) * local_variables * # => [:seven] * - * Symbol objects are different from String objects in that - * Symbol objects represent identifiers, while String objects - * represent text or data. + * A +Symbol+ object differs from a String object in that + * a +Symbol+ object represents an identifier, while a String object + * represents text or data. * * == What's Here * - * First, what's elsewhere. \Class \Symbol: + * First, what's elsewhere. \Class +Symbol+: * * - Inherits from {class Object}[rdoc-ref:Object@What-27s+Here]. * - Includes {module Comparable}[rdoc-ref:Comparable@What-27s+Here]. * - * Here, class \Symbol provides methods that are useful for: + * Here, class +Symbol+ provides methods that are useful for: * * - {Querying}[rdoc-ref:Symbol@Methods+for+Querying] * - {Comparing}[rdoc-ref:Symbol@Methods+for+Comparing] @@ -11298,9 +11653,6 @@ rb_str_unicode_normalized_p(int argc, VALUE *argv, VALUE str) * symbol == object -> true or false * * Returns +true+ if +object+ is the same object as +self+, +false+ otherwise. - * - * Symbol#=== is an alias for Symbol#==. - * */ #define sym_equal rb_obj_equal @@ -11309,14 +11661,14 @@ static int sym_printable(const char *s, const char *send, rb_encoding *enc) { while (s < send) { - int n; - int c = rb_enc_precise_mbclen(s, send, enc); + int n; + int c = rb_enc_precise_mbclen(s, send, enc); - if (!MBCLEN_CHARFOUND_P(c)) return FALSE; - n = MBCLEN_CHARFOUND_LEN(c); - c = rb_enc_mbc_to_codepoint(s, send, enc); - if (!rb_enc_isprint(c, enc)) return FALSE; - s += n; + if (!MBCLEN_CHARFOUND_P(c)) return FALSE; + n = MBCLEN_CHARFOUND_LEN(c); + c = rb_enc_mbc_to_codepoint(s, send, enc); + if (!rb_enc_isprint(c, enc)) return FALSE; + s += n; } return TRUE; } @@ -11334,8 +11686,8 @@ rb_str_symname_p(VALUE sym) ptr = RSTRING_PTR(sym); len = RSTRING_LEN(sym); if ((resenc != enc && !rb_str_is_ascii_only_p(sym)) || len != (long)strlen(ptr) || - !rb_enc_symname2_p(ptr, len, enc) || !sym_printable(ptr, ptr + len, enc)) { - return FALSE; + !rb_enc_symname2_p(ptr, len, enc) || !sym_printable(ptr, ptr + len, enc)) { + return FALSE; } return TRUE; } @@ -11355,18 +11707,18 @@ rb_str_quote_unprintable(VALUE str) ptr = RSTRING_PTR(str); len = RSTRING_LEN(str); if ((resenc != enc && !rb_str_is_ascii_only_p(str)) || - !sym_printable(ptr, ptr + len, enc)) { - return rb_str_escape(str); + !sym_printable(ptr, ptr + len, enc)) { + return rb_str_escape(str); } return str; } -MJIT_FUNC_EXPORTED VALUE +VALUE rb_id_quote_unprintable(ID id) { VALUE str = rb_id2str(id); if (!rb_str_symname_p(str)) { - return rb_str_escape(str); + return rb_str_escape(str); } return str; } @@ -11392,66 +11744,46 @@ sym_inspect(VALUE sym) char *dest; if (!rb_str_symname_p(str)) { - str = rb_str_inspect(str); - len = RSTRING_LEN(str); - rb_str_resize(str, len + 1); - dest = RSTRING_PTR(str); - memmove(dest + 1, dest, len); + str = rb_str_inspect(str); + len = RSTRING_LEN(str); + rb_str_resize(str, len + 1); + dest = RSTRING_PTR(str); + memmove(dest + 1, dest, len); } else { - rb_encoding *enc = STR_ENC_GET(str); - RSTRING_GETMEM(str, ptr, len); - str = rb_enc_str_new(0, len + 1, enc); - dest = RSTRING_PTR(str); - memcpy(dest + 1, ptr, len); + rb_encoding *enc = STR_ENC_GET(str); + VALUE orig_str = str; + + len = RSTRING_LEN(orig_str); + str = rb_enc_str_new(0, len + 1, enc); + + // Get data pointer after allocation + ptr = RSTRING_PTR(orig_str); + dest = RSTRING_PTR(str); + memcpy(dest + 1, ptr, len); + + RB_GC_GUARD(orig_str); } dest[0] = ':'; + + RUBY_ASSERT_BUILTIN_TYPE(str, T_STRING); + return str; } -/* - * call-seq: - * to_s -> string - * - * Returns a string representation of +self+ (not including the leading colon): - * - * :foo.to_s # => "foo" - * - * Symbol#id2name is an alias for Symbol#to_s. - * - * Related: Symbol#inspect, Symbol#name. - */ - VALUE rb_sym_to_s(VALUE sym) { return str_new_shared(rb_cString, rb_sym2str(sym)); } -/* - * call-seq: - * to_sym -> self - * - * Returns +self+. - * - * Symbol#intern is an alias for Symbol#to_sym. - * - * Related: String#to_sym. - */ - -static VALUE -sym_to_sym(VALUE sym) -{ - return sym; -} - -MJIT_FUNC_EXPORTED VALUE +VALUE rb_sym_proc_call(ID mid, int argc, const VALUE *argv, int kw_splat, VALUE passed_proc) { VALUE obj; if (argc < 1) { - rb_raise(rb_eArgError, "no receiver given"); + rb_raise(rb_eArgError, "no receiver given"); } obj = argv[0]; return rb_funcall_with_block_kw(obj, mid, argc - 1, argv + 1, passed_proc, kw_splat); @@ -11465,8 +11797,6 @@ rb_sym_proc_call(ID mid, int argc, const VALUE *argv, int kw_splat, VALUE passed * * :foo.succ # => :fop * - * Symbol#next is an alias for Symbol#succ. - * * Related: String#succ. */ @@ -11498,7 +11828,7 @@ static VALUE sym_cmp(VALUE sym, VALUE other) { if (!SYMBOL_P(other)) { - return Qnil; + return Qnil; } return rb_str_cmp_m(rb_sym2str(sym), rb_sym2str(other)); } @@ -11515,7 +11845,7 @@ static VALUE sym_casecmp(VALUE sym, VALUE other) { if (!SYMBOL_P(other)) { - return Qnil; + return Qnil; } return str_casecmp(rb_sym2str(sym), rb_sym2str(other)); } @@ -11532,7 +11862,7 @@ static VALUE sym_casecmp_p(VALUE sym, VALUE other) { if (!SYMBOL_P(other)) { - return Qnil; + return Qnil; } return str_casecmp_p(rb_sym2str(sym), rb_sym2str(other)); } @@ -11608,9 +11938,6 @@ sym_aref(int argc, VALUE *argv, VALUE sym) * length -> integer * * Equivalent to <tt>self.to_s.length</tt>; see String#length. - * - * Symbol#size is an alias for Symbol#length. - * */ static VALUE @@ -11715,7 +12042,7 @@ sym_start_with(int argc, VALUE *argv, VALUE sym) /* * call-seq: - * end_with?(*string_or_regexp) -> true or false + * end_with?(*strings) -> true or false * * * Equivalent to <tt>self.to_s.end_with?</tt>; see String#end_with?. @@ -11746,12 +12073,12 @@ static VALUE string_for_symbol(VALUE name) { if (!RB_TYPE_P(name, T_STRING)) { - VALUE tmp = rb_check_string_type(name); - if (NIL_P(tmp)) { - rb_raise(rb_eTypeError, "%+"PRIsVALUE" is not a symbol", - name); - } - name = tmp; + VALUE tmp = rb_check_string_type(name); + if (NIL_P(tmp)) { + rb_raise(rb_eTypeError, "%+"PRIsVALUE" is not a symbol", + name); + } + name = tmp; } return name; } @@ -11760,7 +12087,7 @@ ID rb_to_id(VALUE name) { if (SYMBOL_P(name)) { - return SYM2ID(name); + return SYM2ID(name); } name = string_for_symbol(name); return rb_intern_str(name); @@ -11770,7 +12097,7 @@ VALUE rb_to_symbol(VALUE name) { if (SYMBOL_P(name)) { - return name; + return name; } name = string_for_symbol(name); return rb_str_intern(name); @@ -11815,7 +12142,7 @@ rb_interned_str_cstr(const char *ptr) VALUE rb_enc_interned_str(const char *ptr, long len, rb_encoding *enc) { - if (UNLIKELY(rb_enc_autoload_p(enc))) { + if (enc != NULL && UNLIKELY(rb_enc_autoload_p(enc))) { rb_enc_autoload(enc); } @@ -11833,10 +12160,11 @@ void Init_String(void) { rb_cString = rb_define_class("String", rb_cObject); - assert(rb_vm_fstring_table()); + RUBY_ASSERT(rb_vm_fstring_table()); st_foreach(rb_vm_fstring_table(), fstring_set_class_i, rb_cString); rb_include_module(rb_cString, rb_mComparable); rb_define_alloc_func(rb_cString, empty_str_alloc); + rb_define_singleton_method(rb_cString, "new", rb_str_s_new, -1); rb_define_singleton_method(rb_cString, "try_convert", rb_str_s_try_convert, 1); rb_define_method(rb_cString, "initialize", rb_str_init, -1); rb_define_method(rb_cString, "initialize_copy", rb_str_replace, 1); @@ -11881,6 +12209,7 @@ Init_String(void) rb_define_method(rb_cString, "freeze", rb_str_freeze, 0); rb_define_method(rb_cString, "+@", str_uplus, 0); rb_define_method(rb_cString, "-@", str_uminus, 0); + rb_define_method(rb_cString, "dup", rb_str_dup_m, 0); rb_define_alias(rb_cString, "dedup", "-@"); rb_define_method(rb_cString, "to_i", rb_str_to_i, -1); @@ -12008,11 +12337,7 @@ Init_String(void) rb_define_method(rb_cSymbol, "==", sym_equal, 1); rb_define_method(rb_cSymbol, "===", sym_equal, 1); rb_define_method(rb_cSymbol, "inspect", sym_inspect, 0); - rb_define_method(rb_cSymbol, "to_s", rb_sym_to_s, 0); - rb_define_method(rb_cSymbol, "id2name", rb_sym_to_s, 0); rb_define_method(rb_cSymbol, "name", rb_sym2str, 0); /* in symbol.c */ - rb_define_method(rb_cSymbol, "intern", sym_to_sym, 0); - rb_define_method(rb_cSymbol, "to_sym", sym_to_sym, 0); rb_define_method(rb_cSymbol, "to_proc", rb_sym_to_proc, 0); /* in proc.c */ rb_define_method(rb_cSymbol, "succ", sym_succ, 0); rb_define_method(rb_cSymbol, "next", sym_succ, 0); |