Do it for all packages that
* mention perl, or
* have a directory name starting with p5-*, or
* depend on a package starting with p5-
like last time, for 5.18, where this didn't lead to complaints.
Let me know if you have any this time.
Update LICENSE
Upstream change:
2013-07-12 Jarkko Hietaniemi <jhi@Vredefort-2.local>
* Address rt.cpan.org #86857: t/setrlimit.t failure
Skip setrlimit() tests also if the hard limit is RLIM_INFINITY.
* Release 1.2907.
2013-07-07 Jarkko Hietaniemi <jhi@iki.fi>
* Skip number-of-files and stack setrlimit tests in Cygwin:
the constants exist but the implementation doesn't really
seem to be there. Document this in the pod. Testing by dsolimano.
* For setrlimit tests involving byte sizes (memory or file size)
round the byte sizes down to 4 kB boundaries.
* Fix some pod nits.
* Release 1.2906.
2013-06-22 Jarkko Hietaniemi <jhi@iki.fi>
* Address rt.cpan.org #86334: typo fixes
* Address rt.cpan.org #62504: Disable user/system
timing tests if workload accumulate too little time
* Attempt harder in getrusage.t to burn enough user and system time,
refactor the "burn some cpu" utility into t/burn.pl, use that also
in t/times.t.
* Fix a testing problem in NetBSD for t/setrlimit.t,
the tests were misnumbered.
* Silence compilation warnings coming from Perl 5.18.0.
* Clarify the license to be Artistic 2.0 or LGPL 2.0.
* Regenerate ppport.h with Perl 5.18.0.
* Release 1.2905.
a) refer 'perl' in their Makefile, or
b) have a directory name of p5-*, or
c) have any dependency on any p5-* package
Like last time, where this caused no complaints.
to trigger/signal a rebuild for the transition 5.10.1 -> 5.12.1.
The list of packages is computed by finding all packages which end
up having either of PERL5_USE_PACKLIST, BUILDLINK_API_DEPENDS.perl,
or PERL5_PACKLIST defined in their make setup (tested via
"make show-vars VARNAMES=..."), minus the packages updated after
the perl package update.
sno@ was right after all, obache@ kindly asked and he@ led the
way. Thanks!
- updating package to 1.29.03
- removed patch which is applied upstream
Upstream changes:
2009-03-31 Jarkko Hietaniemi <jhi@iki.fi>
* Address rt.cpan.org #44082:
"Bug in compiler detection (Makefile.PL)" (for AIX xlc)
Patch submitted in the report by rehsack at cpan.org.
* Release 1.2903.
Pkgsrc changes:
o Remove no longer needed patch
Upstream changes:
2008-12-30 Jarkko Hietaniemi <jhi@iki.fi>
* Lowered the "0.95 of the current limit" down to
"0.75 of the current limit" in setrlimit.t so that
more memory-constrained systems (like 2MB or less of
default stack) have slightly more chance of completing the test.
The 0.75 also has more chance of aligning with page sizes.
* The scalar context return of getrlimit() was documented
to return an object, even though it really returned the
soft limit, reported by anicka@suse.cz. Fixed the documentation.
* Judging by the cpantesters reports no netbsd has a functional
RLIMIT_STACK, so skipping the test in all of them.
* Release 1.2902.
to trigger/signal a rebuild for the transition 5.8.8 -> 5.10.0.
The list of packages is computed by finding all packages which end
up having either of PERL5_USE_PACKLIST, BUILDLINK_API_DEPENDS.perl,
or PERL5_PACKLIST defined in their make setup (tested via
"make show-vars VARNAMES=...").
that's an accident waiting to happen on next version number "increase").
Also included is a bugfix for one of the self-tests, a comparison
which was the wrong way. Reported upstream.
Tests OK on i386/4.0 if you bump the stack limit to something more
than the default 2MB.
Changes:
2008-01-29 Jarkko Hietaniemi <jhi@iki.fi>
* netbsd-alpha does not have a functional RLIMIT_STACK,
reported by David Cantrell. Skip the test, and document
as a known issue.
* Release 1.2901.
2008-01-27 Jarkko Hietaniemi <jhi@iki.fi>
* Rewrite the setrlimit.t test. Simplify it a lot; now all
RLIM_INFINITY limits are simply skipped; for those that
are limited we try setting only the soft limit, and accept
a value lower than we tried to set, since various systems
might either enforce lower limits, or round up the limits.
Added documentation on setrlimit() about this.
* Allow calling getrlimit/setrlimit and getpriority/setpriority
with the string names ("RLIMIT_NPROC", for example).
* Include ppport.h (generated with perl 5.8.8), reshuffle
system header includes appropriately.
* Tested also with older Perls, 5.005_05 and 5.6.2, in OS X.
(5.005_04 didn't work, it creates but then cannot dynaload
the .bundle files, go figure.)
* Minor documentation tweaks and copyright year bumps.
* Release 1.29.
2006-05-26 Jarkko Hietaniemi
* Tweak the regexp parsing the ps -o output in setpriority.t,
from Alexey Tourbin and Rafael Garcia-Suarez. Still just
a best-effort attempt, but one can only try.
* Release 1.28.
2006-05-25 Jarkko Hietaniemi
* Address rt.cpan.org #13130 and #19412: try to cope with
an already reniced shell running the setpriority.t. The
tricks used are trying "ps -o pid,nice" and looking for $$,
and if that fails, then trying whether the nice(1) is the
GNU one, and if so, running it without arguments.
* Release 1.27.
2006-05-21 Jarkko Hietaniemi
* Address rt.cpan.org #19298: bug in getrusage(): 'inblock' was
misspelt as 'inlock' (the end result being that people looking
for the 'inblock' field got zero)
* For Mac OS X reword the message about the RLIM_NPROC test failure
(or rather, the possible failure). Still a mystery under what
exact conditions the test fails or succeeds.
* Add Test::Pod and Test::Pod::Coverage tests.
* Release 1.26.
2006-04-09 Jarkko Hietaniemi
* Address rt.cpan.org #13131: setrlimit.t tries to increase hard
limits, cannot raise limits higher than the hard limit.
* Address rt.cpan.org #13130: getpriority.t and setpriority.t
assume priority = 0, solution hopefully portable. Now should
work under e.g. "nice make test".
* Add more verbosity to the case of the known rlimit.t test
failure in Mac OS X.
* Release 1.25.
developer is officially maintaining the package.
The rationale for changing this from "tech-pkg" to "pkgsrc-users" is
that it implies that any user can try to maintain the package (by
submitting patches to the mailing list). Since the folks most likely
to care about the package are the folks that want to use it or are
already using it, this would leverage the energy of users who aren't
developers.
module directory has changed (eg. "darwin-2level" vs.
"darwin-thread-multi-2level").
binary packages of perl modules need to be distinguishable between
being built against threaded perl and unthreaded perl, so bump the
PKGREVISION of all perl module packages and introduce
BUILDLINK_RECOMMENDED for perl as perl>=5.8.5nb5 so the correct
dependencies are registered and the binary packages are distinct.
addresses PR pkg/28619 from H. Todd Fujinaka.
Change log:
Sun Feb 29 16:51:34 2004 Jarkko Hietaniemi <jhi@iki.fi>
* No functional changes but document the known fact that
under at least Mac OS X 10.3.2 the t/setrlimit.t subtest #8
may fail because of an OS bug. Also reorder the documentation
a bit to collect all the known issues under one heading.
* Release 1.24.
changes since 1.15:
Tue Oct 7 21:54:04 2003 Jarkko Hietaniemi <jhi@iki.fi>
* Makefile.PL: use archlibexp instead of archlib in cc -I.../CORE
so that Perls installed under ~user work.
* Release 1.23.
Sun Feb 9 12:12:42 2003 Jarkko Hietaniemi <jhi@iki.fi>
* Add RLIMIT_LOCKS.
* Fix a typo s/RLIM_SAVEWD_MAX/RLIMIT_SAVED_MAX/
* Sort the @EXPORT list.
* Release 1.22.
Thu Dec 12 01:15:44 2002 Jarkko Hietaniemi <jhi@iki.fi>
* Patches for NCR MP-RAS from grommel@sears.com.
* Release 1.21.
Tue Nov 26 04:20:38 2002 Jarkko Hietaniemi <jhi@iki.fi>
* Add a warning about the time/load-sensitivity of the
tests to INSTALL and Makefile.PL.
* Rewrite the getrusage.t and times.t tests a bit to
better work on a fast machine (IOW, try spending a
bit more time).
* Release 1.20.
Fri Nov 22 17:57:44 2002 Jarkko Hietaniemi <jhi@iki.fi>
* Fixed a test problem found by Alain Barbet: in NetBSD
the RLIMIT_STACK wants to be aligned. Fixed also the
other memory-related limit tests to allow for alignment.
* Release 1.19.
Wed Nov 20 16:42:50 2002 Jarkko Hietaniemi <jhi@iki.fi>
* Problem found by Jeff Boes, forwarded and analyzed
by Slaven Rezic: if one has a fast machine and/or
the granularity of times() is low, one could get false
negatives from far() in getrusage.t. Fixed the same
problem in times.t.
* Bumped the copyright statement years.
* Release 1.18.
Thu Sep 19 16:53:26 2002 Jarkko Hietaniemi <jhi@iki.fi>
* Patch from Miles Egan to use -I$archlib/CORE instead
of -I$installarchlib/CORE in Makefile.PL, since they
might be different but the first one is the one apps
are supposed to be using.
* Sanity check for getrlimit/setrlimit/getpriority/setpriority
so that one won't use the string "RLIMIT_..."/"PRIO_..." when
one is supposed to be using the constant RLIMIT_.../PRIO_...
* Release 1.17.
Sun Sep 15 18:44:36 2002 Jarkko Hietaniemi <jhi@iki.fi>
* Patch from Allen Smith to avoid -lbsd in IRIX
since it potentially brings in harmful side effects
regarding setpgrp/setgroups.
* Mention the use of VERBOSE in the probe failure message.
* Release 1.16.
Changes:
* Add COPYRIGHT AND LICENSE to Resource.pm.
* A space between the 2>/dev/null and the command is nice.
* I suck. Vivek Khera reported NINE MONTHS ago that
I was missing sub isrss(). My memory was jogged by
a message from Jeff Boes ONE MONTH ago. My apologies.
The automatic truncation in gensolpkg doesn't work for packages which
have the same package name for the first 5-6 chars.
e.g. amanda-server and amanda-client would be named amanda and amanda.
Now, we add a SVR4_PKGNAME and use amacl for amanda-client and amase for
amanda-server.
All svr4 packages also have a vendor tag, so we have to reserve some chars
for this tag, which is normaly 3 or 4 chars. Thats why we can only use 6
or 5 chars for SVR4_PKGNAME. I used 5 for all the packages, to give the
vendor tag enough room.
All p5-* packages and a few other packages have now a SVR4_PKGNAME.
This module provides a Perl interface to part of the BSD process
resource library. It allows the use of the {get,set}rlimit and
getrusage BSD C Library routines from perl.
Provided by Nathan Ahlstrom <nrahlstr@winternet.com> in PR #12630.
