diff options
Diffstat (limited to 'ext/psych/psych_parser.c')
| -rw-r--r-- | ext/psych/psych_parser.c | 575 |
1 files changed, 575 insertions, 0 deletions
diff --git a/ext/psych/psych_parser.c b/ext/psych/psych_parser.c new file mode 100644 index 0000000000..00a2207b58 --- /dev/null +++ b/ext/psych/psych_parser.c @@ -0,0 +1,575 @@ +#include <psych.h> + +VALUE cPsychParser; + +static ID id_read; +static ID id_path; +static ID id_empty; +static ID id_start_stream; +static ID id_end_stream; +static ID id_start_document; +static ID id_end_document; +static ID id_alias; +static ID id_scalar; +static ID id_start_sequence; +static ID id_end_sequence; +static ID id_start_mapping; +static ID id_end_mapping; +static ID id_event_location; + +#define PSYCH_TRANSCODE(_str, _yaml_enc, _internal_enc) \ + do { \ + rb_enc_associate_index((_str), (_yaml_enc)); \ + if(_internal_enc) \ + (_str) = rb_str_export_to_enc((_str), (_internal_enc)); \ + } while (0) + +static int io_reader(void * data, unsigned char *buf, size_t size, size_t *read) +{ + VALUE io = (VALUE)data; + VALUE string = rb_funcall(io, id_read, 1, SIZET2NUM(size)); + + *read = 0; + + if(! NIL_P(string)) { + char * str = StringValuePtr(string); + size_t len = (size_t)RSTRING_LEN(string); + + /* IO#read(size) is documented to return at most `size` bytes, but a + * misbehaving IO-like object may return more. Clamp the copy to the + * buffer libyaml gave us to avoid writing past its end, rounding down + * to a character boundary so a multibyte character is never split. */ + if(len > size) { + rb_encoding * enc = rb_enc_get(string); + len = (size_t)(rb_enc_left_char_head(str, str + size, str + len, enc) - str); + } + + *read = len; + memcpy(buf, str, len); + } + + return 1; +} + +static void dealloc(void * ptr) +{ + yaml_parser_t * parser; + + parser = (yaml_parser_t *)ptr; + yaml_parser_delete(parser); + xfree(parser); +} + +#if 0 +static size_t memsize(const void *ptr) +{ + const yaml_parser_t *parser = ptr; + /* TODO: calculate parser's size */ + return 0; +} +#endif + +static const rb_data_type_t psych_parser_type = { + "Psych/parser", + {0, dealloc, 0,}, + 0, 0, +#ifdef RUBY_TYPED_FREE_IMMEDIATELY + RUBY_TYPED_FREE_IMMEDIATELY, +#endif +}; + +static VALUE allocate(VALUE klass) +{ + yaml_parser_t * parser; + VALUE obj = TypedData_Make_Struct(klass, yaml_parser_t, &psych_parser_type, parser); + + yaml_parser_initialize(parser); + + return obj; +} + +static VALUE make_exception(yaml_parser_t * parser, VALUE path) +{ + if (parser->error == YAML_MEMORY_ERROR) { + return rb_eNoMemError; + } else { + size_t line, column; + VALUE ePsychSyntaxError; + + line = parser->context_mark.line + 1; + column = parser->context_mark.column + 1; + + ePsychSyntaxError = rb_const_get(mPsych, rb_intern("SyntaxError")); + + return rb_funcall(ePsychSyntaxError, rb_intern("new"), 6, + path, + SIZET2NUM(line), + SIZET2NUM(column), + SIZET2NUM(parser->problem_offset), + parser->problem ? rb_usascii_str_new2(parser->problem) : Qnil, + parser->context ? rb_usascii_str_new2(parser->context) : Qnil); + } +} + +static VALUE transcode_string(VALUE src, int * parser_encoding) +{ + int utf8 = rb_utf8_encindex(); + int utf16le = rb_enc_find_index("UTF-16LE"); + int utf16be = rb_enc_find_index("UTF-16BE"); + int source_encoding = rb_enc_get_index(src); + + if (source_encoding == utf8) { + *parser_encoding = YAML_UTF8_ENCODING; + return src; + } + + if (source_encoding == utf16le) { + *parser_encoding = YAML_UTF16LE_ENCODING; + return src; + } + + if (source_encoding == utf16be) { + *parser_encoding = YAML_UTF16BE_ENCODING; + return src; + } + + src = rb_str_export_to_enc(src, rb_utf8_encoding()); + RB_GC_GUARD(src); + + *parser_encoding = YAML_UTF8_ENCODING; + return src; +} + +static VALUE transcode_io(VALUE src, int * parser_encoding) +{ + VALUE io_external_encoding; + int io_external_enc_index; + + io_external_encoding = rb_funcall(src, rb_intern("external_encoding"), 0); + + /* if no encoding is returned, assume ascii8bit. */ + if (NIL_P(io_external_encoding)) { + io_external_enc_index = rb_ascii8bit_encindex(); + } else { + io_external_enc_index = rb_to_encoding_index(io_external_encoding); + } + + /* Treat US-ASCII as utf_8 */ + if (io_external_enc_index == rb_usascii_encindex()) { + *parser_encoding = YAML_UTF8_ENCODING; + return src; + } + + if (io_external_enc_index == rb_utf8_encindex()) { + *parser_encoding = YAML_UTF8_ENCODING; + return src; + } + + if (io_external_enc_index == rb_enc_find_index("UTF-16LE")) { + *parser_encoding = YAML_UTF16LE_ENCODING; + return src; + } + + if (io_external_enc_index == rb_enc_find_index("UTF-16BE")) { + *parser_encoding = YAML_UTF16BE_ENCODING; + return src; + } + + /* Just guess on ASCII-8BIT */ + if (io_external_enc_index == rb_ascii8bit_encindex()) { + *parser_encoding = YAML_ANY_ENCODING; + return src; + } + + /* If the external encoding is something we don't know how to handle, + * fall back to YAML_ANY_ENCODING. */ + *parser_encoding = YAML_ANY_ENCODING; + + return src; +} + +static VALUE protected_start_stream(VALUE pointer) +{ + VALUE *args = (VALUE *)pointer; + return rb_funcall(args[0], id_start_stream, 1, args[1]); +} + +static VALUE protected_start_document(VALUE pointer) +{ + VALUE *args = (VALUE *)pointer; + return rb_funcall3(args[0], id_start_document, 3, args + 1); +} + +static VALUE protected_end_document(VALUE pointer) +{ + VALUE *args = (VALUE *)pointer; + return rb_funcall(args[0], id_end_document, 1, args[1]); +} + +static VALUE protected_alias(VALUE pointer) +{ + VALUE *args = (VALUE *)pointer; + return rb_funcall(args[0], id_alias, 1, args[1]); +} + +static VALUE protected_scalar(VALUE pointer) +{ + VALUE *args = (VALUE *)pointer; + return rb_funcall3(args[0], id_scalar, 6, args + 1); +} + +static VALUE protected_start_sequence(VALUE pointer) +{ + VALUE *args = (VALUE *)pointer; + return rb_funcall3(args[0], id_start_sequence, 4, args + 1); +} + +static VALUE protected_end_sequence(VALUE handler) +{ + return rb_funcall(handler, id_end_sequence, 0); +} + +static VALUE protected_start_mapping(VALUE pointer) +{ + VALUE *args = (VALUE *)pointer; + return rb_funcall3(args[0], id_start_mapping, 4, args + 1); +} + +static VALUE protected_end_mapping(VALUE handler) +{ + return rb_funcall(handler, id_end_mapping, 0); +} + +static VALUE protected_empty(VALUE handler) +{ + return rb_funcall(handler, id_empty, 0); +} + +static VALUE protected_end_stream(VALUE handler) +{ + return rb_funcall(handler, id_end_stream, 0); +} + +static VALUE protected_event_location(VALUE pointer) +{ + VALUE *args = (VALUE *)pointer; + return rb_funcall3(args[0], id_event_location, 4, args + 1); +} + +static VALUE parse(VALUE self, VALUE handler, VALUE yaml, VALUE path) +{ + yaml_parser_t * parser; + yaml_event_t event; + int done = 0; + int state = 0; + int parser_encoding = YAML_ANY_ENCODING; + int encoding = rb_utf8_encindex(); + rb_encoding * internal_enc = rb_default_internal_encoding(); + + TypedData_Get_Struct(self, yaml_parser_t, &psych_parser_type, parser); + + yaml_parser_delete(parser); + yaml_parser_initialize(parser); + + if (rb_respond_to(yaml, id_read)) { + yaml = transcode_io(yaml, &parser_encoding); + yaml_parser_set_encoding(parser, parser_encoding); + yaml_parser_set_input(parser, io_reader, (void *)yaml); + } else { + StringValue(yaml); + yaml = transcode_string(yaml, &parser_encoding); + yaml_parser_set_encoding(parser, parser_encoding); + yaml_parser_set_input_string( + parser, + (const unsigned char *)RSTRING_PTR(yaml), + (size_t)RSTRING_LEN(yaml) + ); + } + + while(!done) { + VALUE event_args[5]; + VALUE start_line, start_column, end_line, end_column; + + if(parser->error || !yaml_parser_parse(parser, &event)) { + VALUE exception; + + exception = make_exception(parser, path); + yaml_parser_delete(parser); + yaml_parser_initialize(parser); + + rb_exc_raise(exception); + } + + start_line = SIZET2NUM(event.start_mark.line); + start_column = SIZET2NUM(event.start_mark.column); + end_line = SIZET2NUM(event.end_mark.line); + end_column = SIZET2NUM(event.end_mark.column); + + event_args[0] = handler; + event_args[1] = start_line; + event_args[2] = start_column; + event_args[3] = end_line; + event_args[4] = end_column; + rb_protect(protected_event_location, (VALUE)event_args, &state); + + switch(event.type) { + case YAML_STREAM_START_EVENT: + { + VALUE args[2]; + + args[0] = handler; + args[1] = INT2NUM(event.data.stream_start.encoding); + rb_protect(protected_start_stream, (VALUE)args, &state); + } + break; + case YAML_DOCUMENT_START_EVENT: + { + VALUE args[4]; + /* Get a list of tag directives (if any) */ + VALUE tag_directives = rb_ary_new(); + /* Grab the document version */ + VALUE version = event.data.document_start.version_directive ? + rb_ary_new3( + (long)2, + INT2NUM(event.data.document_start.version_directive->major), + INT2NUM(event.data.document_start.version_directive->minor) + ) : rb_ary_new(); + + if(event.data.document_start.tag_directives.start) { + yaml_tag_directive_t *start = + event.data.document_start.tag_directives.start; + yaml_tag_directive_t *end = + event.data.document_start.tag_directives.end; + for(; start != end; start++) { + VALUE handle = Qnil; + VALUE prefix = Qnil; + if(start->handle) { + handle = rb_str_new2((const char *)start->handle); + PSYCH_TRANSCODE(handle, encoding, internal_enc); + } + + if(start->prefix) { + prefix = rb_str_new2((const char *)start->prefix); + PSYCH_TRANSCODE(prefix, encoding, internal_enc); + } + + rb_ary_push(tag_directives, rb_ary_new3((long)2, handle, prefix)); + } + } + args[0] = handler; + args[1] = version; + args[2] = tag_directives; + args[3] = event.data.document_start.implicit == 1 ? Qtrue : Qfalse; + rb_protect(protected_start_document, (VALUE)args, &state); + } + break; + case YAML_DOCUMENT_END_EVENT: + { + VALUE args[2]; + + args[0] = handler; + args[1] = event.data.document_end.implicit == 1 ? Qtrue : Qfalse; + rb_protect(protected_end_document, (VALUE)args, &state); + } + break; + case YAML_ALIAS_EVENT: + { + VALUE args[2]; + VALUE alias = Qnil; + if(event.data.alias.anchor) { + alias = rb_str_new2((const char *)event.data.alias.anchor); + PSYCH_TRANSCODE(alias, encoding, internal_enc); + } + + args[0] = handler; + args[1] = alias; + rb_protect(protected_alias, (VALUE)args, &state); + } + break; + case YAML_SCALAR_EVENT: + { + VALUE args[7]; + VALUE anchor = Qnil; + VALUE tag = Qnil; + VALUE plain_implicit, quoted_implicit, style; + VALUE val = rb_str_new( + (const char *)event.data.scalar.value, + (long)event.data.scalar.length + ); + + PSYCH_TRANSCODE(val, encoding, internal_enc); + + if(event.data.scalar.anchor) { + anchor = rb_str_new2((const char *)event.data.scalar.anchor); + PSYCH_TRANSCODE(anchor, encoding, internal_enc); + } + + if(event.data.scalar.tag) { + tag = rb_str_new2((const char *)event.data.scalar.tag); + PSYCH_TRANSCODE(tag, encoding, internal_enc); + } + + plain_implicit = + event.data.scalar.plain_implicit == 0 ? Qfalse : Qtrue; + + quoted_implicit = + event.data.scalar.quoted_implicit == 0 ? Qfalse : Qtrue; + + style = INT2NUM(event.data.scalar.style); + + args[0] = handler; + args[1] = val; + args[2] = anchor; + args[3] = tag; + args[4] = plain_implicit; + args[5] = quoted_implicit; + args[6] = style; + rb_protect(protected_scalar, (VALUE)args, &state); + } + break; + case YAML_SEQUENCE_START_EVENT: + { + VALUE args[5]; + VALUE anchor = Qnil; + VALUE tag = Qnil; + VALUE implicit, style; + if(event.data.sequence_start.anchor) { + anchor = rb_str_new2((const char *)event.data.sequence_start.anchor); + PSYCH_TRANSCODE(anchor, encoding, internal_enc); + } + + tag = Qnil; + if(event.data.sequence_start.tag) { + tag = rb_str_new2((const char *)event.data.sequence_start.tag); + PSYCH_TRANSCODE(tag, encoding, internal_enc); + } + + implicit = + event.data.sequence_start.implicit == 0 ? Qfalse : Qtrue; + + style = INT2NUM(event.data.sequence_start.style); + + args[0] = handler; + args[1] = anchor; + args[2] = tag; + args[3] = implicit; + args[4] = style; + + rb_protect(protected_start_sequence, (VALUE)args, &state); + } + break; + case YAML_SEQUENCE_END_EVENT: + rb_protect(protected_end_sequence, handler, &state); + break; + case YAML_MAPPING_START_EVENT: + { + VALUE args[5]; + VALUE anchor = Qnil; + VALUE tag = Qnil; + VALUE implicit, style; + if(event.data.mapping_start.anchor) { + anchor = rb_str_new2((const char *)event.data.mapping_start.anchor); + PSYCH_TRANSCODE(anchor, encoding, internal_enc); + } + + if(event.data.mapping_start.tag) { + tag = rb_str_new2((const char *)event.data.mapping_start.tag); + PSYCH_TRANSCODE(tag, encoding, internal_enc); + } + + implicit = + event.data.mapping_start.implicit == 0 ? Qfalse : Qtrue; + + style = INT2NUM(event.data.mapping_start.style); + + args[0] = handler; + args[1] = anchor; + args[2] = tag; + args[3] = implicit; + args[4] = style; + + rb_protect(protected_start_mapping, (VALUE)args, &state); + } + break; + case YAML_MAPPING_END_EVENT: + rb_protect(protected_end_mapping, handler, &state); + break; + case YAML_NO_EVENT: + rb_protect(protected_empty, handler, &state); + break; + case YAML_STREAM_END_EVENT: + rb_protect(protected_end_stream, handler, &state); + done = 1; + break; + } + yaml_event_delete(&event); + if (state) rb_jump_tag(state); + } + + return self; +} + +/* + * call-seq: + * parser.mark # => #<Psych::Parser::Mark> + * + * Returns a Psych::Parser::Mark object that contains line, column, and index + * information. + */ +static VALUE mark(VALUE self) +{ + VALUE mark_klass; + VALUE args[3]; + yaml_parser_t * parser; + + TypedData_Get_Struct(self, yaml_parser_t, &psych_parser_type, parser); + mark_klass = rb_const_get_at(cPsychParser, rb_intern("Mark")); + args[0] = SIZET2NUM(parser->mark.index); + args[1] = SIZET2NUM(parser->mark.line); + args[2] = SIZET2NUM(parser->mark.column); + + return rb_class_new_instance(3, args, mark_klass); +} + +void Init_psych_parser(void) +{ +#undef rb_intern +#if 0 + mPsych = rb_define_module("Psych"); +#endif + + cPsychParser = rb_define_class_under(mPsych, "Parser", rb_cObject); + rb_define_alloc_func(cPsychParser, allocate); + + /* Any encoding: Let the parser choose the encoding */ + rb_define_const(cPsychParser, "ANY", INT2NUM(YAML_ANY_ENCODING)); + + /* UTF-8 Encoding */ + rb_define_const(cPsychParser, "UTF8", INT2NUM(YAML_UTF8_ENCODING)); + + /* UTF-16-LE Encoding with BOM */ + rb_define_const(cPsychParser, "UTF16LE", INT2NUM(YAML_UTF16LE_ENCODING)); + + /* UTF-16-BE Encoding with BOM */ + rb_define_const(cPsychParser, "UTF16BE", INT2NUM(YAML_UTF16BE_ENCODING)); + + rb_require("psych/syntax_error"); + + rb_define_private_method(cPsychParser, "_native_parse", parse, 3); + rb_define_method(cPsychParser, "mark", mark, 0); + + id_read = rb_intern("read"); + id_path = rb_intern("path"); + id_empty = rb_intern("empty"); + id_start_stream = rb_intern("start_stream"); + id_end_stream = rb_intern("end_stream"); + id_start_document = rb_intern("start_document"); + id_end_document = rb_intern("end_document"); + id_alias = rb_intern("alias"); + id_scalar = rb_intern("scalar"); + id_start_sequence = rb_intern("start_sequence"); + id_end_sequence = rb_intern("end_sequence"); + id_start_mapping = rb_intern("start_mapping"); + id_end_mapping = rb_intern("end_mapping"); + id_event_location = rb_intern("event_location"); +} |