diff options
| author | Peter Zhu <[email protected]> | 2024-07-29 20:45:26 -0400 |
|---|---|---|
| committer | GitHub <[email protected]> | 2024-07-29 17:45:26 -0700 |
| commit | ce565cd4b851977bf37a470bee54e441bb60486d (patch) | |
| tree | 1a7fea867b14c3a8a88dc07c3fc3c29fa3ebcb4a /re.c | |
| parent | 6d744837abc3f7f71a1f10c7ca399201f6f05e43 (diff) | |
[Bug #20653] Fix memory leak in String#start_with? when regexp times out (#11255)
Fix memory leak in String#start_with? when regexp times out
[Bug #20653]
This commit refactors how Onigmo handles timeout. Instead of raising a
timeout error, onig_search will return a ONIGERR_TIMEOUT which the
caller can free memory, and then raise a timeout error.
This fixes a memory leak in String#start_with when the regexp times out.
For example:
regex = Regexp.new("^#{"(a*)" * 10_000}x$", timeout: 0.000001)
str = "a" * 1000000 + "x"
10.times do
100.times do
str.start_with?(regex)
rescue
end
puts `ps -o rss= -p #{$$}`
end
Before:
33216
51936
71152
81728
97152
103248
120384
133392
133520
133616
After:
14912
15376
15824
15824
16128
16128
16144
16144
16160
16160
Diffstat (limited to 're.c')
| -rw-r--r-- | re.c | 63 |
1 files changed, 15 insertions, 48 deletions
@@ -1671,10 +1671,16 @@ rb_reg_onig_match(VALUE re, VALUE str, if (result < 0) { onig_region_free(regs, 0); - if (result != ONIG_MISMATCH) { + switch (result) { + case ONIG_MISMATCH: + break; + case ONIGERR_TIMEOUT: + rb_raise(rb_eRegexpTimeoutError, "regexp match timeout"); + default: { onig_errmsg_buffer err = ""; onig_error_code_to_str((UChar*)err, (int)result); rb_reg_raise(err, re); + } } } @@ -1735,23 +1741,6 @@ reg_onig_search(regex_t *reg, VALUE str, struct re_registers *regs, void *args_p ONIG_OPTION_NONE); } -struct rb_reg_onig_match_args { - VALUE re; - VALUE str; - struct reg_onig_search_args args; - struct re_registers regs; - - OnigPosition result; -}; - -static VALUE -rb_reg_onig_match_try(VALUE value_args) -{ - struct rb_reg_onig_match_args *args = (struct rb_reg_onig_match_args *)value_args; - args->result = rb_reg_onig_match(args->re, args->str, reg_onig_search, &args->args, &args->regs); - return Qnil; -} - /* returns byte offset */ static long rb_reg_search_set_match(VALUE re, VALUE str, long pos, int reverse, int set_backref_str, VALUE *set_match) @@ -1762,38 +1751,22 @@ rb_reg_search_set_match(VALUE re, VALUE str, long pos, int reverse, int set_back return -1; } - struct rb_reg_onig_match_args args = { - .re = re, - .str = str, - .args = { - .pos = pos, - .range = reverse ? 0 : len, - }, - .regs = {0} + struct reg_onig_search_args args = { + .pos = pos, + .range = reverse ? 0 : len, }; + struct re_registers regs = {0}; - /* If there is a timeout set, then rb_reg_onig_match could raise a - * Regexp::TimeoutError so we want to protect it from leaking memory. */ - if (rb_reg_match_time_limit) { - int state; - rb_protect(rb_reg_onig_match_try, (VALUE)&args, &state); - if (state) { - onig_region_free(&args.regs, false); - rb_jump_tag(state); - } - } - else { - rb_reg_onig_match_try((VALUE)&args); - } + OnigPosition result = rb_reg_onig_match(re, str, reg_onig_search, &args, ®s); - if (args.result == ONIG_MISMATCH) { + if (result == ONIG_MISMATCH) { rb_backref_set(Qnil); return ONIG_MISMATCH; } VALUE match = match_alloc(rb_cMatch); rb_matchext_t *rm = RMATCH_EXT(match); - rm->regs = args.regs; + rm->regs = regs; if (set_backref_str) { RB_OBJ_WRITE(match, &RMATCH(match)->str, rb_str_new4(str)); @@ -1810,7 +1783,7 @@ rb_reg_search_set_match(VALUE re, VALUE str, long pos, int reverse, int set_back rb_backref_set(match); if (set_match) *set_match = match; - return args.result; + return result; } long @@ -4672,12 +4645,6 @@ rb_reg_timeout_p(regex_t *reg, void *end_time_) return false; } -void -rb_reg_raise_timeout(void) -{ - rb_raise(rb_eRegexpTimeoutError, "regexp match timeout"); -} - /* * call-seq: * Regexp.timeout -> float or nil |