re2 0.1.0 → 0.1.1

data/Rakefile

@@ -16,6 +16,10 @@ Rake::TestTask.new do |t|
   t.verbose = true
 end
 
+task :valgrind do
+  system "valgrind  --tool=memcheck --leak-check=full --show-reachable=no --num-callers=15 --track-fds=yes --workaround-gcc296-bugs=yes --max-stackframe=7304328 --dsymutil=yes --track-origins=yes --log-file=report.txt ruby test/leak.rb"
+end
+
 task :test => :compile
 task :default => :test
 

data/ext/re2/re2.cc

@@ -19,6 +19,10 @@ extern "C" {
   #  define RSTRING_LEN(x) (RSTRING(x)->len)
   #endif
 
+  #if !defined(RSTRING_PTR)
+  #  define RSTRING_PTR(x) (RSTRING(x)->ptr)
+  #endif
+
   typedef struct {
     RE2 *pattern;
   } re2_pattern;
@@ -141,9 +145,9 @@ extern "C" {
     VALUE array = rb_ary_new2(m->number_of_matches);
     for (i = 0; i < m->number_of_matches; i++) {
       if (m->matches[i].empty()) {
-        rb_ary_store(array, i, Qnil);
+        rb_ary_push(array, Qnil);
       } else {
-        rb_ary_store(array, i, rb_str_new2(m->matches[i].as_string().c_str()));
+        rb_ary_push(array, rb_str_new(m->matches[i].data(), m->matches[i].size()));
       }
     }
     return array;
@@ -158,7 +162,7 @@ extern "C" {
     if (nth >= m->number_of_matches || m->matches[nth].empty()) {
       return Qnil;
     } else {
-      return rb_str_new2(m->matches[nth].as_string().c_str());
+      return rb_str_new(m->matches[nth].data(), m->matches[nth].size());
     }
   }
 
@@ -236,27 +240,29 @@ extern "C" {
 
     Data_Get_Struct(self, re2_matchdata, m);
 
-    result = rb_str_new2("#<RE2::MatchData");
+    result = rb_str_new("#<RE2::MatchData", 16);
 
     for (i = 0; i < m->number_of_matches; i++) {
-      rb_str_cat2(result, " ");
+      rb_str_cat(result, " ", 1);
 
       if (i > 0) {
         char buf[sizeof(i)*3+1];
         snprintf(buf, sizeof(buf), "%d", i);
         rb_str_cat2(result, buf);
-        rb_str_cat2(result, ":");
+        rb_str_cat(result, ":", 1);
       }
 
       match = re2_matchdata_nth_match(i, self);
 
       if (match == Qnil) {
-        rb_str_cat2(result, "nil");
+        rb_str_cat(result, "nil", 3);
       } else {
-        rb_str_append(result, rb_str_inspect(match));
+        rb_str_cat(result, "\"", 1);
+        rb_str_cat(result, RSTRING_PTR(match), RSTRING_LEN(match));
+        rb_str_cat(result, "\"", 1);
       }
     }
-    rb_str_cat2(result, ">");
+    rb_str_cat(result, ">", 1);
 
     return result;
   }
@@ -409,11 +415,11 @@ extern "C" {
   re2_regexp_inspect(VALUE self)
   {
     re2_pattern *p;
-    VALUE result = rb_str_new2("#<RE2::Regexp /");
+    VALUE result = rb_str_new("#<RE2::Regexp /", 15);
 
     Data_Get_Struct(self, re2_pattern, p);
-    rb_str_cat2(result, p->pattern->pattern().c_str());
-    rb_str_cat2(result, "/>");
+    rb_str_cat(result, p->pattern->pattern().data(), p->pattern->pattern().size());
+    rb_str_cat(result, "/>", 2);
 
     return result;
   }
@@ -431,7 +437,7 @@ extern "C" {
   {
     re2_pattern *p;
     Data_Get_Struct(self, re2_pattern, p);
-    return rb_str_new2(p->pattern->pattern().c_str());
+    return rb_str_new(p->pattern->pattern().data(), p->pattern->pattern().size());
   }
 
   /*
@@ -665,7 +671,7 @@ extern "C" {
   {
     re2_pattern *p;
     Data_Get_Struct(self, re2_pattern, p);
-    return rb_str_new2(p->pattern->error().c_str());
+    return rb_str_new(p->pattern->error().data(), p->pattern->error().size());
   }
 
   /*
@@ -679,7 +685,7 @@ extern "C" {
   {
     re2_pattern *p;
     Data_Get_Struct(self, re2_pattern, p);
-    return rb_str_new2(p->pattern->error_arg().c_str());
+    return rb_str_new(p->pattern->error_arg().data(), p->pattern->error_arg().size());
   }
 
   /*
@@ -746,7 +752,7 @@ extern "C" {
         BOOL2RUBY(p->pattern->options().one_line()));
 
     /* This is a read-only hash after all... */
-    OBJ_FREEZE(options);
+    rb_obj_freeze(options);
 
     return options;
   }
@@ -844,7 +850,8 @@ extern "C" {
       Data_Get_Struct(matchdata, re2_matchdata, m);
       m->matches = new (std::nothrow) re2::StringPiece[n];
       m->regexp = self;
-      m->string = rb_str_dup_frozen(text);
+      m->string = rb_str_dup(text);
+      rb_str_freeze(m->string);
 
       if (m->matches == 0) {
         rb_raise(rb_eNoMemError, "not enough memory to allocate StringPieces for matches");
@@ -896,6 +903,10 @@ extern "C" {
   static VALUE
   re2_Replace(VALUE self, VALUE str, VALUE pattern, VALUE rewrite)
   {
+
+    /* Look out for frozen strings. */
+    rb_check_frozen(str);
+
     UNUSED(self);
     VALUE repl;
     re2_pattern *p;
@@ -916,7 +927,10 @@ extern "C" {
     repl = rb_str_new(str_as_string.c_str(), str_as_string.length());
 
     /* Replace the original string with the replacement. */
-    rb_str_update(str, 0, RSTRING_LEN(str), repl);
+    if (RSTRING_LEN(str) != RSTRING_LEN(repl)) {
+      rb_str_resize(str, RSTRING_LEN(repl));
+    }
+    memcpy(RSTRING_PTR(str), RSTRING_PTR(repl), RSTRING_LEN(repl));
 
     return str;
   }
@@ -939,6 +953,10 @@ extern "C" {
   static VALUE
   re2_GlobalReplace(VALUE self, VALUE str, VALUE pattern, VALUE rewrite)
   {
+
+    /* Look out for frozen strings. */
+    rb_check_frozen(str);
+
     UNUSED(self);
 
     /* Convert all the inputs to be pumped into RE2::GlobalReplace. */
@@ -959,7 +977,10 @@ extern "C" {
     repl = rb_str_new(str_as_string.c_str(), str_as_string.length());
 
     /* Replace the original string with the replacement. */
-    rb_str_update(str, 0, RSTRING_LEN(str), repl);
+    if (RSTRING_LEN(str) != RSTRING_LEN(repl)) {
+      rb_str_resize(str, RSTRING_LEN(repl));
+    }
+    memcpy(RSTRING_PTR(str), RSTRING_PTR(repl), RSTRING_LEN(repl));
 
     return str;
   }
@@ -979,7 +1000,8 @@ extern "C" {
   {
     UNUSED(self);
     re2::StringPiece unquoted_as_string_piece(StringValuePtr(unquoted));
-    return rb_str_new2(RE2::QuoteMeta(unquoted_as_string_piece).c_str());
+    std::string quoted_string = RE2::QuoteMeta(unquoted_as_string_piece);
+    return rb_str_new(quoted_string.data(), quoted_string.size());
   }
 
   void

data/test/leak.rb

@@ -0,0 +1,32 @@
+$:.unshift(File.join(File.dirname(__FILE__), '..', 'lib'))
+require 're2'
+
+loops = if ENV["LOOPS"]
+  ENV["LOOPS"].to_i
+else
+  150
+end
+
+puts "Looping #{loops} times..."
+
+loops.times do
+  r = RE2::Regexp.new('woo(oo)(o+)(\d*)')
+  r.ok?
+  r.match('wooo')
+  r.match('wooo', 0)
+  r.match('wooo', 1)
+  r.match('wooo', 2)
+  r =~ 'woooo'
+  r =~ 'bob'
+  RE2::Replace("woo", "woo", "bob")
+  RE2::GlobalReplace("woo", "woo", "bob")
+  m = r.match('woooooo1234')
+  m[0]
+  m[1]
+  m[0,2]
+  m[1..-1]
+  m.string
+  m.regexp
+  m.inspect
+  sleep 0.1
+end

data/test/re2_test.rb

@@ -76,6 +76,16 @@ class RE2Test < Test::Unit::TestCase
     assert_equal name_id, name.object_id
   end
 
+  def test_replace_with_a_frozen_string
+    frozen_name = "Arnold".freeze
+    assert frozen_name.frozen?
+
+    # Ruby 1.9 raises a RuntimeError instead of a TypeError.
+    assert_raise TypeError, RuntimeError do
+      RE2::Replace(frozen_name, "o", "a")
+    end
+  end
+
   def test_global_replace
     assert_equal "waa", RE2::GlobalReplace("woo", "o", "a")
     assert_equal "hoo", RE2::GlobalReplace("woo", "w", "h")
@@ -88,6 +98,16 @@ class RE2Test < Test::Unit::TestCase
     assert_equal name_id, name.object_id
   end
 
+  def test_global_replace_with_a_frozen_string
+    frozen_name = "Arnold".freeze
+    assert frozen_name.frozen?
+
+    # Ruby 1.9 raises a RuntimeError instead of a TypeError.
+    assert_raise TypeError, RuntimeError do
+      RE2::GlobalReplace(frozen_name, "o", "a")
+    end
+  end
+
   def test_compiling
     r = RE2::Regexp.new("woo")
     assert r.ok?

metadata

@@ -3,48 +3,49 @@ name: re2
 version: !ruby/object:Gem::Version 
   prerelease: false
   segments: 
-  - 0
-  - 1
-  - 0
-  version: 0.1.0
+    - 0
+    - 1
+    - 1
+  version: 0.1.1
 platform: ruby
 authors: 
-- Paul Mucur
+  - Paul Mucur
 autorequire: 
 bindir: bin
 cert_chain: []
 
-date: 2010-08-28 00:00:00 +01:00
+date: 2010-09-26 00:00:00 +01:00
 default_executable: 
 dependencies: 
-- !ruby/object:Gem::Dependency 
-  name: rake-compiler
-  prerelease: false
-  requirement: &id001 !ruby/object:Gem::Requirement 
-    none: false
-    requirements: 
-    - - ">="
-      - !ruby/object:Gem::Version 
-        segments: 
-        - 0
-        version: "0"
-  type: :development
-  version_requirements: *id001
+  - !ruby/object:Gem::Dependency 
+    name: rake-compiler
+    prerelease: false
+    requirement: &id001 !ruby/object:Gem::Requirement 
+      none: false
+      requirements: 
+        - - ">="
+          - !ruby/object:Gem::Version 
+            segments: 
+              - 0
+            version: "0"
+    type: :development
+    version_requirements: *id001
 description: Ruby bindings to re2, "an efficient, principled regular expression library".
 email: ruby.re2@librelist.com
 executables: []
 
 extensions: 
-- ext/re2/extconf.rb
+  - ext/re2/extconf.rb
 extra_rdoc_files: []
 
 files: 
-- ext/re2/extconf.rb
-- ext/re2/re2.cc
-- LICENSE.txt
-- README.md
-- Rakefile
-- test/re2_test.rb
+  - ext/re2/extconf.rb
+  - ext/re2/re2.cc
+  - LICENSE.txt
+  - README.md
+  - Rakefile
+  - test/re2_test.rb
+  - test/leak.rb
 has_rdoc: true
 homepage: http://github.com/mudge/re2
 licenses: []
@@ -53,23 +54,23 @@ post_install_message:
 rdoc_options: []
 
 require_paths: 
-- lib
+  - lib
 required_ruby_version: !ruby/object:Gem::Requirement 
   none: false
   requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      segments: 
-      - 0
-      version: "0"
+    - - ">="
+      - !ruby/object:Gem::Version 
+        segments: 
+          - 0
+        version: "0"
 required_rubygems_version: !ruby/object:Gem::Requirement 
   none: false
   requirements: 
-  - - ">="
-    - !ruby/object:Gem::Version 
-      segments: 
-      - 0
-      version: "0"
+    - - ">="
+      - !ruby/object:Gem::Version 
+        segments: 
+          - 0
+        version: "0"
 requirements: []
 
 rubyforge_project: 
@@ -78,4 +79,5 @@ signing_key:
 specification_version: 3
 summary: Ruby bindings to re2.
 test_files: 
-- test/re2_test.rb
+  - test/re2_test.rb
+  - test/leak.rb