Perl 6 - the future is here, just unevenly distributed

IRC log for #moarvm, 2016-04-25

| Channels | #moarvm index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:35 lizmat joined #moarvm
01:00 lizmat joined #moarvm
03:00 matiaslina joined #moarvm
05:07 cognominal joined #moarvm
06:09 domidumont joined #moarvm
06:15 domidumont joined #moarvm
07:01 brrt joined #moarvm
07:01 brrt \o
07:02 brrt http://www.freelists.org/post/luajit/Random-failures-in-compiled-code <- fortunately i'm not the only one that sometimes miscompiles code :-)
07:13 brrt the infiniloop in nqp building on reframe-jit is not so easy to repeat
07:14 brrt ... let's try ASAN
07:18 zakharyas joined #moarvm
07:23 brrt does the --asan option work on OS X?
07:23 brrt clearly, it does not
07:26 brrt fix is easy fortunately :-)
07:51 lizmat joined #moarvm
07:59 brrt infiniloop does originate from something ASAN is catching
08:34 mtj_ joined #moarvm
08:58 mtj_ joined #moarvm
09:48 vendethiel- joined #moarvm
11:10 leont_ joined #moarvm
13:00 khagan joined #moarvm
14:07 leont_ joined #moarvm
14:21 tomboy64 joined #moarvm
14:21 tomboy64 psch: mind taking a look at that? https://bpaste.net/show/0337fa47a038
14:34 cognominal joined #moarvm
14:37 brrt left #moarvm
14:37 brrt joined #moarvm
14:37 brrt //quit
14:47 nebuchadnezzar joined #moarvm
14:57 domidumont joined #moarvm
14:59 tomboy64 joined #moarvm
15:13 psch tomboy64: okay, i took a look.  i'm neither particularly perl5 savvy nor familiar with the moarvm build process, though... :)
15:14 psch jnthn: did you see tomboy64++'s patch for a pkg-config aware moarvm Configure.pl?
15:15 leont_ joined #moarvm
15:19 jnthn psch: No
15:19 jnthn Link?
15:19 psch an hour ago
15:19 psch 14:21 < tomboy64> psch: mind taking a look at that? https://bpaste.net/show/0337fa47a038
15:20 jnthn oh, no wonder I didn't pay any attention, it was addressed to you ;)
15:21 psch i just send tomboy64 here vOv
15:21 psch *sent
15:21 psch the day before yesterday, i think.. :)
15:21 * jnthn doesn't really know much about pkg-config and prefers to pretend not to know anything about the build process :P
15:22 jnthn On line 45,
15:22 jnthn +            $config{cincludes} .= ' ' . "$result";
15:22 jnthn Oh, includes, not flags
15:23 jnthn Though the invocation of pkg-config is --cflags :)
15:23 tomboy64 the problem is, if you build against native libraries, Configure.pl doesn't get the includes for libffi right
15:24 tomboy64 except for libtommath, the proper way for that would be to run pkg-config
15:24 tomboy64 ExtUtils-PkgConfig would be a perl wrapper for that but i was told to use native pkg-config or a perl-core package
15:24 * tomboy64 shrugs
15:29 jnthn Only thing I'd really note is that if it detects pkg-config but then pkg-config doesn't know about the package in question, it'll blow up
15:29 jnthn Whereas before it'd just carry on
15:29 tomboy64 no, it will continue silently
15:29 jnthn I dunno if that means "it'd be doomed anyway so this is fail fast"
15:29 jnthn Oh?
15:30 jnthn +        my $result = `$config{pkgconfig} --cflags libuv`;
15:30 jnthn We then test $? and die otherwise
15:30 tomboy64 there a check for $? before
15:30 jnthn Does it come back with 0 if pkg-config doesn't know the package?
15:30 tomboy64 it comes back with a non-zero result
15:30 tomboy64 try it out ;)
15:31 tomboy64 pkg-config --cflags blablubb
15:31 tomboy64 echo $?
15:32 jnthn Right, so
15:32 jnthn +    if ($config{pkgconfig_works}) {
15:32 jnthn +        my $result = `$config{pkgconfig} --cflags libuv`;
15:32 jnthn +        if ( $? == 0 ) {
15:32 jnthn And later
15:32 tomboy64 https://bpaste.net/show/72fc34c83e53 <--- i made a little change here, forgot to upload that
15:32 jnthn oh, it's a print, not a die
15:32 jnthn d'oh
15:32 jnthn OK, no problem. I just can't read, apparently :)
15:32 tomboy64 :)
15:32 tomboy64 it first checks if pkgconfig works in general
15:33 tomboy64 if pkgconfig works, check for the lib. if that's not there, use the original behavior.
15:33 jnthn yeah, I got that part. I just wrongly thought that if it did work but didn't not he package, we bailed.
15:33 jnthn But it just notes it and goes on
15:33 tomboy64 well, that would be an option ...
15:33 tomboy64 silent errors are rarely good :-}
15:33 jnthn It would, but sounds less robust :)
15:33 jnthn I think I like noting it
15:33 jnthn As it does now
15:34 tomboy64 only problem left is, tommath doesn't register .pc files
15:34 tomboy64 e.g. to pkg-config it's invisible
15:34 tomboy64 (no idea why)
15:37 jnthn OK. I'm fine with that patch going in
15:38 tomboy64 great :)
15:38 ilmari tomboy64: because nobody submitted patches to them? ;-P
15:40 tomboy64 ilmari: mebbe :)
15:48 tomboy64 jnthn: psch: what about patches for nqp?
15:49 tomboy64 is this the right place?
15:49 psch tomboy64: no, i think that would be #p6dev, ideally with a PR to http://github.com/perl6/nqp
15:49 tomboy64 psch: roger that
15:51 jnthn tomboy64: If it's comfortable to PR that patch, please do so...otherwise I (or somebody else) can apply it manually later
15:51 jnthn (Or just push it to a git remote somewhere is as good as the PR)
15:51 tomboy64 i can do a pr
15:51 * jnthn bbl
15:51 jnthn Thanks!
15:51 tomboy64 i will just need to make sure it actually works the way i intend it to
17:46 zakharyas joined #moarvm
18:26 Ven joined #moarvm
21:38 zakharyas joined #moarvm
22:23 lizmat joined #moarvm
22:29 tomboy64 joined #moarvm
22:42 tomboy64 joined #moarvm
23:01 tomboy64 joined #moarvm
23:27 cognominal joined #moarvm
23:55 cognominal joined #moarvm

| Channels | #moarvm index | Today | | Search | Google Search | Plain-Text | summary