Merge strscan-3.0.5

Merged: https://github.com/ruby/ruby/pull/6890

ext/strscan/extconf.rb

@@ -1,5 +1,9 @@
 # frozen_string_literal: true
 require 'mkmf'
-$INCFLAGS << " -I$(top_srcdir)" if $extmk
-have_func("onig_region_memsize", "ruby.h")
-create_makefile 'strscan'
+if RUBY_ENGINE == 'ruby'
+  $INCFLAGS << " -I$(top_srcdir)" if $extmk
+  have_func("onig_region_memsize", "ruby.h")
+  create_makefile 'strscan'
+else
+  File.write('Makefile', dummy_makefile("").join)
+end

ext/strscan/strscan.c

@@ -22,7 +22,7 @@ extern size_t onig_region_memsize(const struct re_registers *regs);
 
 #include <stdbool.h>
 
-#define STRSCAN_VERSION "3.0.1"
+#define STRSCAN_VERSION "3.0.5"
 
 /* =======================================================================
                          Data Type Definitions
@@ -435,11 +435,11 @@ strscan_get_pos(VALUE self)
  *
  * In short, it's a 0-based index into the string.
  *
- *   s = StringScanner.new("abcädeföghi")
- *   s.charpos           # -> 0
- *   s.scan_until(/ä/)   # -> "abcä"
- *   s.pos               # -> 5
- *   s.charpos           # -> 4
+ *   s = StringScanner.new("abc\u00e4def\u00f6ghi")
+ *   s.charpos                # -> 0
+ *   s.scan_until(/\u00e4/)   # -> "abc\u00E4"
+ *   s.pos                    # -> 5
+ *   s.charpos                # -> 4
  */
 static VALUE
 strscan_get_charpos(VALUE self)
@@ -1458,6 +1458,54 @@ strscan_fixed_anchor_p(VALUE self)
     return p->fixed_anchor_p ? Qtrue : Qfalse;
 }
 
+typedef struct {
+    VALUE self;
+    VALUE captures;
+} named_captures_data;
+
+static int
+named_captures_iter(const OnigUChar *name,
+                    const OnigUChar *name_end,
+                    int back_num,
+                    int *back_refs,
+                    OnigRegex regex,
+                    void *arg)
+{
+    named_captures_data *data = arg;
+
+    VALUE key = rb_str_new((const char *)name, name_end - name);
+    VALUE value = RUBY_Qnil;
+    int i;
+    for (i = 0; i < back_num; i++) {
+        value = strscan_aref(data->self, INT2NUM(back_refs[i]));
+    }
+    rb_hash_aset(data->captures, key, value);
+    return 0;
+}
+
+/*
+ * call-seq:
+ *   scanner.named_captures -> hash
+ *
+ * Returns a hash of string variables matching the regular expression.
+ *
+ *   scan = StringScanner.new('foobarbaz')
+ *   scan.match?(/(?<f>foo)(?<r>bar)(?<z>baz)/)
+ *   scan.named_captures # -> {"f"=>"foo", "r"=>"bar", "z"=>"baz"}
+ */
+static VALUE
+strscan_named_captures(VALUE self)
+{
+    struct strscanner *p;
+    GET_SCANNER(self, p);
+    named_captures_data data;
+    data.self = self;
+    data.captures = rb_hash_new();
+    onig_foreach_name(RREGEXP_PTR(p->regex), named_captures_iter, &data);
+
+    return data.captures;
+}
+
 /* =======================================================================
                               Ruby Interface
    ======================================================================= */
@@ -1468,6 +1516,8 @@ strscan_fixed_anchor_p(VALUE self)
  * StringScanner provides for lexical scanning operations on a String.  Here is
  * an example of its usage:
  *
+ *   require 'strscan'
+ *
  *   s = StringScanner.new('This is an example string')
  *   s.eos?               # -> false
  *
@@ -1650,4 +1700,6 @@ Init_strscan(void)
     rb_define_method(StringScanner, "inspect",     strscan_inspect,     0);
 
     rb_define_method(StringScanner, "fixed_anchor?", strscan_fixed_anchor_p, 0);
+
+    rb_define_method(StringScanner, "named_captures", strscan_named_captures, 0);
 }

ext/strscan/strscan.gemspec

@@ -16,13 +16,26 @@ Gem::Specification.new do |s|
   s.summary = "Provides lexical scanning operations on a String."
   s.description = "Provides lexical scanning operations on a String."
 
-  s.require_path = %w{lib}
-  s.files = %w{ext/strscan/extconf.rb ext/strscan/strscan.c}
-  s.extensions = %w{ext/strscan/extconf.rb}
+  files = [
+    "COPYING",
+    "LICENSE.txt",
+  ]
+  if RUBY_ENGINE == "jruby"
+    s.require_paths = %w{ext/jruby/lib lib}
+    files << "ext/jruby/lib/strscan.rb"
+    files << "lib/strscan.jar"
+    s.platform = "java"
+  else
+    s.require_paths = %w{lib}
+    files << "ext/strscan/extconf.rb"
+    files << "ext/strscan/strscan.c"
+    s.extensions = %w{ext/strscan/extconf.rb}
+  end
+  s.files = files
   s.required_ruby_version = ">= 2.4.0"
 
-  s.authors = ["Minero Aoki", "Sutou Kouhei"]
-  s.email = [nil, "kou@cozmixng.org"]
+  s.authors = ["Minero Aoki", "Sutou Kouhei", "Charles Oliver Nutter"]
+  s.email = [nil, "kou@cozmixng.org", "headius@headius.com"]
   s.homepage = "https://github.com/ruby/strscan"
   s.licenses = ["Ruby", "BSD-2-Clause"]
 end

test/strscan/test_ractor.rb

@@ -3,7 +3,7 @@ require 'test/unit'
 
 class TestStringScannerRactor < Test::Unit::TestCase
   def setup
-    pend unless defined? Ractor
+    omit "Ractor not defined" unless defined? Ractor
   end
 
   def test_ractor

test/strscan/test_stringscanner.rb

@@ -207,6 +207,8 @@ class TestStringScanner < Test::Unit::TestCase
   end
 
   def test_charpos_not_use_string_methods
+    omit "not supported on TruffleRuby" if RUBY_ENGINE == "truffleruby"
+
     string = +'abcädeföghi'
     scanner = create_string_scanner(string)
 
@@ -567,6 +569,8 @@ class TestStringScanner < Test::Unit::TestCase
   end
 
   def test_invalid_encoding_string
+    omit "no encoding check on TruffleRuby for scan(String)" if RUBY_ENGINE == "truffleruby"
+
     str = "\xA1\xA2".dup.force_encoding("euc-jp")
     ss = create_string_scanner(str)
     assert_raise(Encoding::CompatibilityError) do
@@ -712,6 +716,8 @@ class TestStringScanner < Test::Unit::TestCase
   end
 
   def test_aref_without_regex
+    omit "#[:missing] always raises on TruffleRuby if matched" if RUBY_ENGINE == "truffleruby"
+
     s = create_string_scanner('abc')
     s.get_byte
     assert_nil(s[:c])
@@ -757,6 +763,33 @@ class TestStringScanner < Test::Unit::TestCase
     assert_equal(false, StringScanner.new("a", fixed_anchor: nil).fixed_anchor?)
     assert_equal(false, StringScanner.new("a", fixed_anchor: false).fixed_anchor?)
   end
+
+  def test_scan_aref_repeatedly
+    s = StringScanner.new('test string')
+    assert_equal "test",   s.scan(/\w(\w)(\w*)/)
+    assert_equal "test",   s[0]
+    assert_equal "e",      s[1]
+    assert_equal "st",     s[2]
+    assert_nil             s.scan(/\w+/)
+    assert_nil             s[0]
+    assert_nil             s[1]
+    assert_nil             s[2]
+    assert_equal " ",      s.scan(/\s+/)
+    assert_equal " ",      s[0]
+    assert_nil             s[1]
+    assert_nil             s[2]
+    assert_equal "string", s.scan(/\w(\w)(\w*)/)
+    assert_equal "string", s[0]
+    assert_equal "t",      s[1]
+    assert_equal "ring",   s[2]
+  end
+
+  def test_named_captures
+    omit("not implemented on TruffleRuby") if ["truffleruby"].include?(RUBY_ENGINE)
+    scan = StringScanner.new("foobarbaz")
+    assert_equal(9, scan.match?(/(?<f>foo)(?<r>bar)(?<z>baz)/))
+    assert_equal({"f" => "foo", "r" => "bar", "z" => "baz"}, scan.named_captures)
+  end
 end
 
 class TestStringScannerFixedAnchor < TestStringScanner