Stas Bekman wrote:
Stas Bekman wrote:
Steve Hay wrote:
Kurt George Gjerde wrote:
[...]
Sorry. But yes, the following two scripts reproduces the error.
Thanks, Kurt. They don't segfault for me on linux. So I may need to try to adjust your test.
They both fail for me too. Only the first one gives the "free to pool during global destruction" error, but both give program error popup windows. Stack traces below from 5.8.3-RC1. (They're both the same, I think.)
Kurt and Steve, no problem, I'll try to fix it and send the patches to p5p. I've talked to Nick and he said that the fixes won't enter 5.8.3 no matter what. But they will be available on CPAN (since Storable has a dual life). First I want to finish the batch of mp2 docs generation. There is a way too many functions. Then I need to handle A-T autoconfig problems and next I'll work on this issue.
I wasn't able to reproduce it on linux, but please try this new version: http://apache.org/~stas/Storable-2.10.tar.gz
I've added a new test t/threads.t which contains one of the "exploits" Kurt has posted, plus added a call to init_perinterp from a newly added CLONE function, which hopefully fixes the problem.
Yes, it does appear to fix the problems -- both the test scripts that Kurt put together now run without error for me.
Great. Thanks for testing, Steve.
However, the new t/threads.t file isn't in that package.
oops, I forgot to update MANIFEST.
If only fixing mp2 was that easy...
If only I could reproduce the segfaults you reported on mp2. I have some ideas of the possible culprits, all related to CLONE.
As I'm still unable to write a test for Storable that will make it fail under linux I'm digging into the XS, trying to make it to.
__________________________________________________________________ Stas Bekman JAm_pH ------> Just Another mod_perl Hacker http://stason.org/ mod_perl Guide ---> http://perl.apache.org mailto:[EMAIL PROTECTED] http://use.perl.org http://apacheweek.com http://modperlbook.org http://apache.org http://ticketmaster.com
--------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
