Bug #116724 for Sys-PageCache: Unescaped left brace in regex is illegal here in regex

Thu Aug 04 17:41:47 2016 SREZIC [...] cpan.org - Ticket created

Subject:

Unescaped left brace in regex is illegal here in regex

Tests fail with perl 5.25.x because of making unescaped { illegal: Unescaped left brace in regex is illegal here in regex; marked by <-- HERE in m/\\x{ <-- HERE ([0-9a-z]+)}/ at t/Util.pm line 21. Compilation failed in require at t/09_page_size.t line 3. BEGIN failed--compilation aborted at t/09_page_size.t line 3. t/09_page_size.t ........ Dubious, test returned 255 (wstat 65280, 0xff00) No subtests run In earlier versions (5.24.x ...) there was just a warning.

Fri Aug 05 01:49:40 2016 hirose31 [...] gmail.com - Correspondence added

I've released v0.03. I hope this release fix this problem. https://metacpan.org/diff/file?source=HIROSE/Sys-PageCache-0.02&target=HIROSE/Sys-PageCache-0.03#t/Util.pm Thanks for your report!

Fri Aug 05 01:49:40 2016 The RT System itself - Status changed from 'new' to 'open'

Fri Aug 05 01:50:15 2016 hirose31 [...] gmail.com - Status changed from 'open' to 'resolved'

Fri Aug 05 02:14:27 2016 SREZIC [...] cpan.org - Correspondence added

On 2016-08-05 01:49:40, HIROSE wrote: Show quoted text

> I've released v0.03. I hope this release fix this problem.

Yes, this particular problem seems to be fixed. However, CPAN Testers report a lot of (random?) failures: http://fast-matrix.cpantesters.org/?dist=Sys-PageCache%200.03

Fri Aug 05 04:40:00 2016 hirose31 [...] gmail.com - Correspondence added

I think the root cause of "mmap: Cannot allocate memory" depends on testing environment. eg: set very low resource limit.

Mon Aug 08 02:08:43 2016 SREZIC [...] cpan.org - Correspondence added

On 2016-08-05 04:40:00, HIROSE wrote: Show quoted text

> I think the root cause of "mmap: Cannot allocate memory" depends on > testing environment. eg: set very low resource limit.

It's true that smoker systems (including mine) work with resource limits. But I get failures regardless if the limits are set or not. Here's a fail/pass table sorted by operating system, whether it's a VM or a physical system, and amount of memory. Unfortunately I don't see a pattern here --- it may fail or pass for VMs, and the amount of memory does not seem to matter: FreeBSD 9.2 (VM, 1GB RAM): fails FreeBSD 9.2 (physical, 16 GB RAM): fails FreeBSD 10.1 (VM, 1GB RAM): fails Debian/wheezy (VM, 2GB RAM): passes Debian/jessie (VM, 32GB RAM): fails Debian/jessie (pyhsical, 32GB RAM): passes (This is with various perl versions ranging between 5.8 and 5.25, and I say said, it does not matter if there are any memory limits active)

Mon Aug 08 02:08:49 2016 SREZIC [...] cpan.org - Status changed from 'resolved' to 'open'

Mon Jul 24 15:53:24 2017 SREZIC [...] cpan.org - Correspondence added

On 2016-08-08 02:08:43, SREZIC wrote: Show quoted text

> On 2016-08-05 04:40:00, HIROSE wrote:

> > I think the root cause of "mmap: Cannot allocate memory" depends on > > testing environment. eg: set very low resource limit.

> > It's true that smoker systems (including mine) work with resource > limits. But I get failures regardless if the limits are set or not. > > Here's a fail/pass table sorted by operating system, whether it's a VM > or a physical system, and amount of memory. Unfortunately I don't see > a pattern here --- it may fail or pass for VMs, and the amount of > memory does not seem to matter: > > FreeBSD 9.2 (VM, 1GB RAM): fails > FreeBSD 9.2 (physical, 16 GB RAM): fails > FreeBSD 10.1 (VM, 1GB RAM): fails > Debian/wheezy (VM, 2GB RAM): passes > Debian/jessie (VM, 32GB RAM): fails > Debian/jessie (pyhsical, 32GB RAM): passes > > (This is with various perl versions ranging between 5.8 and 5.25, and > I say said, it does not matter if there are any memory limits active)

The initial issue is resolved, I'll create another ticket for the remaining problems.

Mon Jul 24 15:53:29 2017 SREZIC [...] cpan.org - Status changed from 'open' to 'resolved'

Mon Jul 24 15:53:30 2017 SREZIC [...] cpan.org - Fixed in 0.03 added