regexec.c: Rmv obsolete recursion checks: GH #8369 #19636
Conversation
There is no recursion to exceed limits of. This fixes Perl#8369
| if (cur_curlyx->u.curlyx.count >= /*max*/ARG2(cur_curlyx->u.curlyx.me)) { | ||
| /* Maximum greed exceeded */ | ||
| if (cur_curlyx->u.curlyx.count >= REG_INFTY | ||
| && ckWARN(WARN_REGEXP) | ||
| && !reginfo->warned) | ||
| { | ||
| reginfo->warned = TRUE; | ||
| Perl_warner(aTHX_ packWARN(WARN_REGEXP), | ||
| "Complex regular subexpression recursion " | ||
| "limit (%d) exceeded", | ||
| REG_INFTY - 1); | ||
| } | ||
| cur_curlyx->u.curlyx.count--; | ||
| CACHEsayNO; | ||
| } |
There was a problem hiding this comment.
I think this removes too much: the outer block looks to me like it is checking whether we've reached the max allowed for this (minimal) repetition. I can have a dig to verify, but I expect @demerphq will know.
If I'm correct, then we should really have a test that fails when you remove this. (I'm a bit surprised we don't have any tests of the warning, for that matter.)
There was a problem hiding this comment.
Thank you, I believe you are correct.
There was a problem hiding this comment.
@khwilliamson I think this got merged without fixing the above over-removal, please could you verify?
|
The test suite only exercises this code via the fold_grind tests, and only then because it is using a ?? quantifier |
I fixed this locally as a result of Hugo van der Sanden's comment in #19636, but neglected to commit it.
|
Thanks for finding this. Fixed by bdadb24 I'll change the issue title to indicate we need tests for this |
|
I instead created a new ticket #19781 |
I fixed this locally as a result of Hugo van der Sanden's comment in Perl#19636, but neglected to commit it.
There is no recursion to exceed limits of.
This fixes #8369