MAJOR=4 MINOR=38 DEVNAME=tty38 $ˆl–ßi~” *h¥6 ƒp¸TÅ£Á …u·œ?À¿?÷2009-07-08 Andreas J. Koenig * from cpantesters-discuss 05:04:06 <@ST47> Hey 05:04:21 <@ST47> Anyone ever had a problem with the FTPstats.yml getting corrupted? 05:04:49 <@ST47> It keeps getting messed up and throwing a syntax error through YAML::Syck and throwing my smokers into a useless state 06:28:45 <@ST47> Probably ran out of disk space mid-write... 07:53:33 <@rafl> yeah, i've seen that. 2009-07-02 Andreas J. Koenig * Curtis Jewell reports failing tests that I overlooked on cpantesters. Seems to be that Expect-less systems fail some tests around push/pop on urllist??? http://www.nntp.perl.org/group/perl.cpan.testers/2009/06/msg4356518.html Nope, I can remove Expect and still pass all tests. He is stuck in the test "urllist empty" and then gets the fail for the test "reordering urllist". Skipping the "urllist empty" test when no Expect is installed will most probably "fix" this. I mean, we're fixing a test, not code. Nope. Was a whitespace bug, FIXED. 2009-07-01 Andreas J. Koenig * amend version numbers and let Andy Armstrongs script care for them automatically. 1.9402 was delivered with 3 decreasing version numbers. How come? Status: Decreasing version number ================================= module: CPAN::Distribution version: 1.93 in file: CPAN-1.9402/lib/CPAN/Distribution.pm status: Not indexed because CPAN-1.9304/lib/CPAN.pm in A/AN/ANDK/CPAN-1.9304.tar.gz has a higher version number (1.9304) module: CPAN::Index version: 1.93 in file: CPAN-1.9402/lib/CPAN/Index.pm status: Not indexed because CPAN-1.9304/lib/CPAN.pm in A/AN/ANDK/CPAN-1.9304.tar.gz has a higher version number (1.9304) module: CPAN::LWP::UserAgent version: 1.00 in file: CPAN-1.9402/lib/CPAN/LWP/UserAgent.pm status: Not indexed because CPAN-1.9304/lib/CPAN.pm in A/AN/ANDK/CPAN-1.9304.tar.gz has a higher version number (1.9304) Something like perl-reversion -current 0.0.4 -set 0.0.5 lib/**/*.pm 2009-04-19 Andreas J. Koenig * Acme::Badexample ($VERSION = do { $_++ while 1; return $_ };) breaks us. Could not get a stable solution from BSD::Resource. The problem was that when running the r command after a while the number of $SIG{XCPU} interrupts increased until finally the shell ended itself. Probably a bug somewhere else. Giving up at this state: diff --git a/lib/CPAN/Module.pm b/lib/CPAN/Module.pm index f9520d9..92d53c9 100644 --- a/lib/CPAN/Module.pm +++ b/lib/CPAN/Module.pm @@ -652,9 +652,32 @@ sub available_version { #-> sub CPAN::Module::parse_version ; sub parse_version { my($self,$parsefile) = @_; + our($soft0,$hard0,$used0); + if (CPAN->has_usable("BSD::Resource")) { + ($used0) = BSD::Resource::getrusage(); + ($soft0,$hard0) = BSD::Resource::getrlimit(BSD::Resource::RLIMIT_CPU()) unless $soft0; + warn "used0[$used0]soft0[$soft0]hard0[$hard0]\n"; + my $success = BSD::Resource::setrlimit(BSD::Resource::RLIMIT_CPU(),int(3+$used0),$hard0); + if ($success) { + $SIG{XCPU} = sub { + my $sig = shift; + die +{ SIG => $sig}; + }; + } else { + $CPAN::Frontend->mywarn("Could not set limit\n"); + } + } my $have = eval { MM->parse_version($parsefile); }; if ($@) { - $CPAN::Frontend->mywarn("Error while parsing version number in file '$parsefile'\n"); + if (my $sig = $@->{SIG}) { + $CPAN::Frontend->mywarn("Signal $sig caught while parsing version number in file '$parsefile'\n"); + } else { + $CPAN::Frontend->mywarn("Error while parsing version number in file '$parsefile'\n"); + } + } + if ($soft0) { + BSD::Resource::setrlimit(BSD::Resource::RLIMIT_CPU(),$soft0,$hard0); + $S