* re.c (rb_reg_s_union, Init_Regexp): new method `Regexp.union'.

* lib/pathname.rb (realpath): examine Dir.pwd because it may have
  symlinks.


git-svn-id: svn+ssh://ci.ruby-lang.org/ruby/trunk@4856 b2dd03c8-39d4-4d8f-98ff-823fe69b080e

3 files changed, 120 insertions, 32 deletions

diff --git a/ChangeLog b/ChangeLog
index 17284bdea7..ea03e547a0 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,10 @@
+Wed Oct 29 17:27:05 2003  Tanaka Akira  <akr@m17n.org>
+
+	* re.c (rb_reg_s_union, Init_Regexp): new method `Regexp.union'.
+
+	* lib/pathname.rb (realpath): examine Dir.pwd because it may have
+	  symlinks.
+
 Wed Oct 29 17:16:31 2003  Nobuyoshi Nakada  <nobu@ruby-lang.org>
 
 	* eval.c (rb_longjmp): must not disturb original jump.
diff --git a/lib/pathname.rb b/lib/pathname.rb
index 4822a4a172..b4462bfeaf 100644
--- a/lib/pathname.rb
+++ b/lib/pathname.rb
@@ -105,27 +105,29 @@ class Pathname
     Pathname.new(path)
   end
 
-  # realpath returns real pathname of self in actual filesystem.
+  # realpath returns a real pathname of self in actual filesystem.
+  # The real pathname doesn't contain a symlink and useless dots.
   #
   # If false is given for the optional argument force_absolute,
   # it may return relative pathname.
   # Otherwise it returns absolute pathname.
   def realpath(force_absolute=true)
-    # Check file existence at first by File.stat.
-    # This test detects ELOOP.
-    #
-    # /tmp/a -> a
-    # /tmp/b -> b/b
-    # /tmp/c -> ./c
-    # /tmp/d -> ../tmp/d
-
-    File.stat(@path)
-
-    top = %r{\A/} =~ @path ? '/' : ''
-    unresolved = @path.scan(%r{[^/]+})
+    if %r{\A/} =~ @path
+      top = '/'
+      unresolved = @path.scan(%r{[^/]+})
+    elsif force_absolute
+      # Although POSIX getcwd returns a pathname which contains no symlink,
+      # 4.4BSD-Lite2 derived getcwd may return the environment variable $PWD
+      # which may contain a symlink.
+      # So the return value of Dir.pwd should be examined.
+      top = '/'
+      unresolved = Dir.pwd.scan(%r{[^/]+}) + @path.scan(%r{[^/]+})
+    else
+      top = ''
+      unresolved = @path.scan(%r{[^/]+})
+    end
     resolved = []
-    checked_path = {}
-    
+
     until unresolved.empty?
       case unresolved.last
       when '.'
@@ -133,34 +135,36 @@ class Pathname
       when '..'
         resolved.unshift unresolved.pop
       else
-        path = top + unresolved.join('/')
-        raise Errno::ELOOP.new(path) if checked_path[path]
-        checked_path[path] = true
-        if File.lstat(path).symlink?
-          link = File.readlink(path)
-          if %r{\A/} =~ link
+        loop_check = {}
+        while (stat = File.lstat(path = top + unresolved.join('/'))).symlink?
+          symlink_id = "#{stat.dev}:#{stat.ino}"
+          raise Errno::ELOOP.new(path) if loop_check[symlink_id]
+          loop_check[symlink_id] = true
+          if %r{\A/} =~ (link = File.readlink(path))
             top = '/'
             unresolved = link.scan(%r{[^/]+})
           else
-            unresolved.pop
-            unresolved.concat link.scan(%r{[^/]+})
+            unresolved[-1,1] = link.scan(%r{[^/]+})
           end
+        end
+        next if (filename = unresolved.pop) == '.'
+        if filename != '..' && resolved.first == '..'
+          resolved.shift
         else
-          resolved.unshift unresolved.pop
+          resolved.unshift filename
         end
       end
     end
+
+    if top == '/'
+      resolved.shift while resolved[0] == '..'
+    end
     
     if resolved.empty?
-      path = top.empty? ? '.' : top
+      Pathname.new(top.empty? ? '.' : '/')
     else
-      path = top + resolved.join('/')
+      Pathname.new(top + resolved.join('/'))
     end
-    
-    # Note that Dir.pwd has no symlinks.
-    path = File.join(Dir.pwd, path) if %r{\A/} !~ path && force_absolute
-
-    Pathname.new(path).cleanpath
   end
 
   # parent method returns parent directory, i.e. ".." is joined at last.
@@ -297,7 +301,7 @@ end
 # IO
 class Pathname
   # Pathname#each_line iterates over lines of the file.
-  # It's yields String objects for each line.
+  # It's yields a String object for each line.
   #
   # This method is exist since 1.8.1.
   def each_line(*args, &block) IO.foreach(@path, *args, &block) end
diff --git a/re.c b/re.c
index 7460d6ae3d..9fb770015a 100644
--- a/re.c
+++ b/re.c
@@ -1485,6 +1485,82 @@ rb_reg_options(re)
 }
 
 static VALUE
+rb_reg_s_union(argc, argv)
+    int argc;
+    VALUE *argv;
+{
+    if (argc == 0) {
+        VALUE args[1];
+        args[0] = rb_str_new2("(?!)");
+        return rb_class_new_instance(1, args, rb_cRegexp);
+    }
+    else if (argc == 1) {
+        VALUE v;
+        v = rb_check_convert_type(argv[0], T_REGEXP, "Regexp", "to_regexp");
+        if (!NIL_P(v))
+            return v;
+        else {
+            VALUE args[1];
+            args[0] = rb_reg_s_quote(argc, argv);
+            return rb_class_new_instance(1, args, rb_cRegexp);
+        }
+    }
+    else {
+        int i, kcode = -1;
+        VALUE kcode_re = Qnil;
+        VALUE source = rb_str_buf_new(0);
+        VALUE args[3];
+        for (i = 0; i < argc; i++) {
+            volatile VALUE v;
+            if (0 < i)
+                rb_str_buf_cat2(source, "|");
+            v = rb_check_convert_type(argv[i], T_REGEXP, "Regexp", "to_regexp");
+            if (!NIL_P(v)) {
+                if (FL_TEST(v, KCODE_FIXED)) {
+                    if (kcode == -1) {
+                        kcode_re = v;
+                        kcode = RBASIC(v)->flags & KCODE_MASK;
+                    }
+                    else if ((RBASIC(v)->flags & KCODE_MASK) != kcode) {
+                        volatile VALUE str1, str2;
+                        str1 = rb_inspect(kcode_re);
+                        str2 = rb_inspect(v);
+                        rb_raise(rb_eArgError, "mixed kcode: %s and %s",
+                            RSTRING(str1)->ptr, RSTRING(str2)->ptr);
+                    }
+                }
+                v = rb_reg_to_s(v);
+            }
+            else {
+                args[0] = argv[i];
+                v = rb_reg_s_quote(1, args);
+            }
+            rb_str_buf_append(source, v);
+        }
+        args[0] = source;
+        args[1] = Qnil;
+        switch (kcode) {
+          case -1:
+            args[2] = Qnil;
+            break;
+          case KCODE_NONE:
+            args[2] = rb_str_new2("n");
+            break;
+          case KCODE_EUC:
+            args[2] = rb_str_new2("e");
+            break;
+          case KCODE_SJIS:
+            args[2] = rb_str_new2("s");
+            break;
+          case KCODE_UTF8:
+            args[2] = rb_str_new2("u");
+            break;
+        }
+        return rb_class_new_instance(3, args, rb_cRegexp);
+    }
+}
+
+static VALUE
 rb_reg_init_copy(copy, re)
     VALUE copy, re;
 {
@@ -1732,6 +1808,7 @@ Init_Regexp()
     rb_define_singleton_method(rb_cRegexp, "compile", rb_class_new_instance, -1);
     rb_define_singleton_method(rb_cRegexp, "quote", rb_reg_s_quote, -1);
     rb_define_singleton_method(rb_cRegexp, "escape", rb_reg_s_quote, -1);
+    rb_define_singleton_method(rb_cRegexp, "union", rb_reg_s_union, -1);
     rb_define_singleton_method(rb_cRegexp, "last_match", rb_reg_s_last_match, -1);
 
     rb_define_method(rb_cRegexp, "initialize", rb_reg_initialize_m, -1);