Bug #108955 for BSD-Resource: No constants exported for the new RLIMIT values

Sun Nov 15 03:58:30 2015 ntyni [...] iki.fi - Ticket created

Subject:

No constants exported for the new RLIMIT values

The new rlimits like RLIMIT_MSGQUEUE and RLIMIT_NICE don't have exported constants like the old ones, even though they're in the get_rlimits() hash. % perl -MBSD::Resource -le 'print RLIMIT_CORE' 4 % perl -MBSD::Resource -le 'print RLIMIT_NICE' RLIMIT_NICE is not a valid BSD::Resource macro at -e line 1. % perl -MBSD::Resource -le 'print ((get_rlimits())->{RLIMIT_NICE})' 13 As it happens, this breaks the mod_perl2 test suite (in t/modules/apache_resource.t), because Apache2::Resource now generates a fatal warning.

Sun Nov 15 04:45:39 2015 CARNIL [...] cpan.org - Cc CARNIL added

Sun Nov 15 08:13:10 2015 JHI [...] cpan.org - Correspondence added

On Sun Nov 15 03:58:30 2015, ntyni@iki.fi wrote: Show quoted text

> The new rlimits like RLIMIT_MSGQUEUE and RLIMIT_NICE don't have > exported constants like the old ones, even though they're in the > get_rlimits() hash. > > % perl -MBSD::Resource -le 'print RLIMIT_CORE' > 4 > > % perl -MBSD::Resource -le 'print RLIMIT_NICE' > RLIMIT_NICE is not a valid BSD::Resource macro at -e line 1. > > % perl -MBSD::Resource -le 'print ((get_rlimits())->{RLIMIT_NICE})' > 13 > > As it happens, this breaks the mod_perl2 test suite (in > t/modules/apache_resource.t), because Apache2::Resource now generates > a fatal warning.

What operating system is this? That it would be only those two constants and not the others added in 1.2908 would be strange, since all of them were added using the same pattern. (Which is the same pattern used before.)

Sun Nov 15 08:13:10 2015 The RT System itself - Status changed from 'new' to 'open'

Sun Nov 15 10:23:43 2015 ntyni [...] iki.fi - Correspondence added

Subject:	Re: [rt.cpan.org #108955] No constants exported for the new RLIMIT values
Date:	Sun, 15 Nov 2015 17:23:26 +0200
To:	Jarkko_Hietaniemi via RT <bug-BSD-Resource [...] rt.cpan.org>
From:	Niko Tyni <ntyni [...] iki.fi>

On Sun, Nov 15, 2015 at 08:13:10AM -0500, Jarkko_Hietaniemi via RT wrote: Show quoted text

> > What operating system is this?

Debian unstable, amd64. Show quoted text

> That it would be only those two constants and not the others added in 1.2908 would be strange, since all of them were added using the same pattern. (Which is the same pattern used before.)

Sure it applies to all the new ones, those two were just examples. Sorry for being unclear. % perl -MBSD::Resource -le 'my $r = get_rlimits(); for my $res (keys %$r) { my $val = eval "&BSD::Resource::${res}()"; print "$res $r->{$res} <$val>" }' RLIMIT_FSIZE 1 <1> RLIMIT_RTTIME 15 <> RLIMIT_SIGPENDING 11 <> RLIMIT_OFILE 7 <7> RLIMIT_VMEM 9 <9> RLIMIT_RSS 5 <5> RLIMIT_LOCKS 10 <10> RLIMIT_CORE 4 <4> RLIMIT_DATA 2 <2> RLIMIT_OPEN_MAX 7 <7> RLIMIT_RTPRIO 14 <> RLIMIT_AS 9 <9> RLIMIT_NICE 13 <> RLIMIT_NPROC 6 <6> RLIMIT_CPU 0 <0> RLIMIT_STACK 3 <3> RLIMIT_NOFILE 7 <7> RLIMIT_MEMLOCK 8 <8> RLIMIT_MSGQUEUE 12 <> -- Niko Tyni ntyni@debian.org

Sun Nov 15 10:30:59 2015 JHI [...] cpan.org - Correspondence added

On Sun Nov 15 10:23:43 2015, ntyni@iki.fi wrote: Show quoted text

> On Sun, Nov 15, 2015 at 08:13:10AM -0500, Jarkko_Hietaniemi via RT > wrote:

> > > > What operating system is this?

> > Debian unstable, amd64. >

> > That it would be only those two constants and not the others added in > > 1.2908 would be strange, since all of them were added using the same > > pattern. (Which is the same pattern used before.)

> > Sure it applies to all the new ones, those two were just examples. > Sorry > for being unclear. > > % perl -MBSD::Resource -le 'my $r = get_rlimits(); for my $res (keys > %$r) { my $val = eval "&BSD::Resource::${res}()"; print "$res $r-

> >{$res} <$val>" }'

> RLIMIT_FSIZE 1 <1> > RLIMIT_RTTIME 15 <> > RLIMIT_SIGPENDING 11 <> > RLIMIT_OFILE 7 <7> > RLIMIT_VMEM 9 <9> > RLIMIT_RSS 5 <5> > RLIMIT_LOCKS 10 <10> > RLIMIT_CORE 4 <4> > RLIMIT_DATA 2 <2> > RLIMIT_OPEN_MAX 7 <7> > RLIMIT_RTPRIO 14 <> > RLIMIT_AS 9 <9> > RLIMIT_NICE 13 <> > RLIMIT_NPROC 6 <6> > RLIMIT_CPU 0 <0> > RLIMIT_STACK 3 <3> > RLIMIT_NOFILE 7 <7> > RLIMIT_MEMLOCK 8 <8> > RLIMIT_MSGQUEUE 12 <>

Hmm. In this case it smells like the t/modules/apache_resource.t needs some looking into. What is it actually trying to test? And what does "man getrlimit" say about those missing constants? What it now looks like is that those constants are defined in the system headers (since they get compiled in) but then ... they do not work?

Sun Nov 15 10:50:17 2015 ntyni [...] iki.fi - Correspondence added

Subject:	Re: [rt.cpan.org #108955] No constants exported for the new RLIMIT values
Date:	Sun, 15 Nov 2015 17:49:57 +0200
To:	Jarkko_Hietaniemi via RT <bug-BSD-Resource [...] rt.cpan.org>
From:	Niko Tyni <ntyni [...] iki.fi>

On Sun, Nov 15, 2015 at 10:30:59AM -0500, Jarkko_Hietaniemi via RT wrote: Show quoted text

> <URL: https://rt.cpan.org/Ticket/Display.html?id=108955 >

Show quoted text

> Hmm. In this case it smells like the t/modules/apache_resource.t needs some looking into. What is it actually trying to test?

Yes, the test should probably use the get_rlimits() hash values instead of the exported constants, if the lack of the latter is intentional. I assumed it isn't. Show quoted text

> And what does "man getrlimit" say about those missing constants? What it now looks like is that those constants are defined in the system headers (since they get compiled in) but then ... they do not work?

I think the problem is just that the new RLIMIT_* names aren't listed in constant() in Resource.xs like the old ones are. -- Niko Tyni ntyni@debian.org

Sun Nov 15 10:59:43 2015 JHI [...] cpan.org - Correspondence added

On Sun Nov 15 10:50:17 2015, ntyni@iki.fi wrote: Show quoted text

> I think the problem is just that the new RLIMIT_* names aren't listed > in constant() in Resource.xs like the old ones are.

Ahh. That might be it. I'll roll out a new release within the next 24h.

Sun Nov 15 17:33:37 2015 gregoa [...] cpan.org - Cc GREGOA added

Tue Feb 02 10:31:42 2016 JHI [...] cpan.org - Correspondence added

I believe this is fixed in 1.2909.

Tue Feb 02 10:31:43 2016 JHI [...] cpan.org - Status changed from 'open' to 'resolved'

Tue Feb 02 10:31:43 2016 JHI [...] cpan.org - Fixed in 1.2909 added