aboutsummaryrefslogtreecommitdiffstats
diff options
context:
space:
mode:
authorKazuki Yamaguchi <k@rhe.jp>2016-11-16 19:52:02 +0900
committerKazuki Yamaguchi <k@rhe.jp>2016-11-19 13:55:50 +0000
commit39cf3c0a381399f8bf4c7eda89af1caae33078e6 (patch)
tree4267a1feda35d17e463a6c64c94070eb3406c0e6
parentacbaa8c4269263f05a90c959533b5199841b4b1e (diff)
downloadruby-topic/re-matchdata-fixes.tar.gz
re.c: fix for RMatch::regexp == Qniltopic/re-matchdata-fixes
So, don't blindly assume RMatch::regexp contains an instance of Regexp and do check properly, because it may be nil as of r45451.
-rw-r--r--re.c31
-rw-r--r--test/ruby/test_regexp.rb22
2 files changed, 39 insertions, 14 deletions
diff --git a/re.c b/re.c
index 069a9bc..1e576d7 100644
--- a/re.c
+++ b/re.c
@@ -1088,8 +1088,7 @@ match_regexp(VALUE match)
static VALUE
match_names(VALUE match)
{
- match_check(match);
- return rb_reg_names(RMATCH(match)->regexp);
+ return rb_reg_names(match_regexp(match));
}
/*
@@ -1116,9 +1115,8 @@ match_backref_number(VALUE match, VALUE backref)
{
const char *name;
int num;
-
struct re_registers *regs = RMATCH_REGS(match);
- VALUE regexp = RMATCH(match)->regexp;
+ VALUE regexp;
match_check(match);
switch (TYPE(backref)) {
@@ -1134,6 +1132,7 @@ match_backref_number(VALUE match, VALUE backref)
break;
}
+ regexp = match_regexp(match);
num = onig_name_to_backref_number(RREGEXP_PTR(regexp),
(const unsigned char*)name,
(const unsigned char*)name + strlen(name),
@@ -1918,7 +1917,7 @@ match_ary_aref(VALUE match, VALUE idx, VALUE result)
static VALUE
match_aref(int argc, VALUE *argv, VALUE match)
{
- VALUE idx, length;
+ VALUE idx, length, regexp;
match_check(match);
rb_scan_args(argc, argv, "11", &idx, &length);
@@ -1928,7 +1927,9 @@ match_aref(int argc, VALUE *argv, VALUE match)
return rb_reg_nth_match(FIX2INT(idx), match);
}
else {
- int num = namev_to_backref_number(RMATCH_REGS(match), RMATCH(match)->regexp, idx);
+ int num;
+ regexp = match_regexp(match);
+ num = namev_to_backref_number(RMATCH_REGS(match), regexp, idx);
if (num >= 0) {
return rb_reg_nth_match(num, match);
}
@@ -1989,7 +1990,8 @@ match_values_at(int argc, VALUE *argv, VALUE match)
rb_ary_push(result, rb_reg_nth_match(FIX2INT(argv[i]), match));
}
else {
- int num = namev_to_backref_number(RMATCH_REGS(match), RMATCH(match)->regexp, argv[i]);
+ VALUE regexp = match_regexp(match);
+ int num = namev_to_backref_number(RMATCH_REGS(match), regexp, argv[i]);
if (num >= 0) {
rb_ary_push(result, rb_reg_nth_match(num, match));
}
@@ -2078,15 +2080,16 @@ match_named_captures_iter(const OnigUChar *name, const OnigUChar *name_end,
static VALUE
match_named_captures(VALUE match)
{
- VALUE hash;
+ VALUE hash, regexp;
struct MEMO *memo;
match_check(match);
+ regexp = match_regexp(match);
hash = rb_hash_new();
memo = MEMO_NEW(hash, match, 0);
- onig_foreach_name(RREGEXP(RMATCH(match)->regexp)->ptr, match_named_captures_iter, (void*)memo);
+ onig_foreach_name(RREGEXP(regexp)->ptr, match_named_captures_iter, (void*)memo);
return hash;
}
@@ -2947,12 +2950,14 @@ rb_reg_equal(VALUE re1, VALUE re2)
static VALUE
match_hash(VALUE match)
{
+ VALUE regexp;
const struct re_registers *regs;
st_index_t hashval;
match_check(match);
+ regexp = match_regexp(match);
hashval = rb_hash_start(rb_str_hash(RMATCH(match)->str));
- hashval = rb_hash_uint(hashval, reg_hash(RMATCH(match)->regexp));
+ hashval = rb_hash_uint(hashval, reg_hash(regexp));
regs = RMATCH_REGS(match);
hashval = rb_hash_uint(hashval, regs->num_regs);
hashval = rb_hash_uint(hashval, rb_memhash(regs->beg, regs->num_regs * sizeof(*regs->beg)));
@@ -2974,11 +2979,15 @@ static VALUE
match_equal(VALUE match1, VALUE match2)
{
const struct re_registers *regs1, *regs2;
+ VALUE reg1, reg2;
+
if (match1 == match2) return Qtrue;
if (!RB_TYPE_P(match2, T_MATCH)) return Qfalse;
if (!RMATCH(match1)->regexp || !RMATCH(match2)->regexp) return Qfalse;
if (!rb_str_equal(RMATCH(match1)->str, RMATCH(match2)->str)) return Qfalse;
- if (!rb_reg_equal(RMATCH(match1)->regexp, RMATCH(match2)->regexp)) return Qfalse;
+ reg1 = match_regexp(match1);
+ reg2 = match_regexp(match2);
+ if (!rb_reg_equal(reg1, reg2)) return Qfalse;
regs1 = RMATCH_REGS(match1);
regs2 = RMATCH_REGS(match2);
if (regs1->num_regs != regs2->num_regs) return Qfalse;
diff --git a/test/ruby/test_regexp.rb b/test/ruby/test_regexp.rb
index b0da247..7c09467 100644
--- a/test/ruby/test_regexp.rb
+++ b/test/ruby/test_regexp.rb
@@ -639,14 +639,30 @@ class TestRegexp < Test::Unit::TestCase
end
def test_match_without_regexp
+ # create a MatchData for each assertion because the internal state may change
+ test = proc {|&blk| "abc".sub("a", ""); blk.call($~) }
+
bug10877 = '[ruby-core:68209] [Bug #10877]'
- "abc".sub("a", "")
- assert_raise_with_message(IndexError, /foo/, bug10877) {$~["foo"]}
+ test.call {|m| assert_raise_with_message(IndexError, /foo/, bug10877) {m["foo"]} }
key = "\u{3042}"
[Encoding::UTF_8, Encoding::Shift_JIS, Encoding::EUC_JP].each do |enc|
idx = key.encode(enc)
- assert_raise_with_message(IndexError, /#{idx}/, bug10877) {$~[idx]}
+ test.call {|m| assert_raise_with_message(IndexError, /#{idx}/, bug10877) {m[idx]} }
end
+ test.call {|m| assert_equal(/a/, m.regexp) }
+ test.call {|m| assert_equal("abc", m.string) }
+ test.call {|m| assert_equal(1, m.size) }
+ test.call {|m| assert_equal(0, m.begin(0)) }
+ test.call {|m| assert_equal(1, m.end(0)) }
+ test.call {|m| assert_equal([0, 1], m.offset(0)) }
+ test.call {|m| assert_equal([], m.captures) }
+ test.call {|m| assert_equal([], m.names) }
+ test.call {|m| assert_equal({}, m.named_captures) }
+ test.call {|m| assert_equal(/a/.match("abc"), m) }
+ test.call {|m| assert_equal(/a/.match("abc").hash, m.hash) }
+ test.call {|m| assert_equal("bc", m.post_match) }
+ test.call {|m| assert_equal("", m.pre_match) }
+ test.call {|m| assert_equal(["a", nil], m.values_at(0, 1)) }
end
def test_last_match