bpo-11105: use a lower recursion limit for infinite recursion tests#26550
bpo-11105: use a lower recursion limit for infinite recursion tests#26550isidentical merged 1 commit into
Conversation
|
@Fidget-Spinner does this patch prevent the crash for your end (do you have resources to test it out?). I highly doubt so (since these functions are giant I don't expect them to be inlined), though if it does then it would be good to know. |
Sorry, something went wrong.
Thanks for working on this! Yup I can test it out, I use MSVC often. Your doubts were right, this is sadly still causing a stack overflow. BTW, testing with buildbots should catch it, the |
Sorry, something went wrong.
Cool! I was worried that they are not available on the public list, but apparently, they are! Thanks for letting me know and testing this out. |
Sorry, something went wrong.
b246c21 to
7f83161
Compare
June 5, 2021 20:29
7f83161 to
d7b496f
Compare
June 5, 2021 20:47
d7b496f to
f31c841
Compare
June 5, 2021 22:39
|
🤖 New build scheduled with the buildbot fleet by @isidentical for commit f31c8411cc51590aee9bc8c2b3042911316b546e 🤖 If you want to schedule another build, you need to add the ":hammer: test-with-buildbots" label again. |
Sorry, something went wrong.
|
Seems like the error in |
Sorry, something went wrong.
I just checked on main branch, this specific test also causes a stack overflow there. So maybe that test failure was previously masked by this one. Your patch fixes the AST stack overflow stack overflow issue for windows. Thanks! Something I found strange is that the exceptions recursion test was added in 4e7a69b but it passed all along. I'm starting to suspect maybe there's some other change elsewhere that caused both |
Sorry, something went wrong.
|
Ok, bisect tells me that this commit caused |
Sorry, something went wrong.
f31c841 to
1bf3518
Compare
June 8, 2021 11:08
|
🤖 New build scheduled with the buildbot fleet by @isidentical for commit 1bf3518db90f3e1c6c8eb1d8daf925e0ae5ee649 🤖 If you want to schedule another build, you need to add the ":hammer: test-with-buildbots" label again. |
Sorry, something went wrong.
|
@isidentical I think if you rebase to the latest 6d518bb, test with buildbots should mostly pass. commit f3fa63e was reverted. |
Sorry, something went wrong.
1bf3518 to
e01e32d
Compare
June 8, 2021 11:29
|
🤖 New build scheduled with the buildbot fleet by @isidentical for commit e01e32d 🤖 If you want to schedule another build, you need to add the ":hammer: test-with-buildbots" label again. |
Sorry, something went wrong.
Fidget-Spinner
left a comment
There was a problem hiding this comment.
LGTM. The windows 10 tests are passing. Thanks for your efforts. Without your help I would've taken a lot longer to track down the other windows buildbot failures.
Sorry, something went wrong.
|
Thanks @isidentical for the PR 🌮🎉.. I'm working now to backport this PR to: 3.10. |
Sorry, something went wrong.
|
Thanks @isidentical for the PR 🌮🎉.. I'm working now to backport this PR to: 3.9. |
Sorry, something went wrong.
|
Sorry, @isidentical, I could not cleanly backport this to |
Sorry, something went wrong.
|
Sorry @isidentical, I had trouble checking out the |
Sorry, something went wrong.
(cherry picked from commit e58d762) Co-authored-by: Batuhan Taskaya <batuhan@python.org>
(cherry picked from commit e58d762) Co-authored-by: Batuhan Taskaya <batuhan@python.org>
https://bugs.python.org/issue11105