|
Fire is step THREE! | github.com/perl6/toolchain-bikeshed | Channel logs: irclog.perlgeek.de/perl6-toolchain/today | useful prior art: metacpan.org/pod/CPAN::Meta::Spec Set by moderator on 8 February 2016. |
|||
| patrickz | gotta go to sleep now. Will be back tomorrow... | 00:01 | |
|
01:00
sufrostico joined
|
|||
| ugexe | patrickz: i added ZEF_BUILDPM_DEBUG=1 to output the relevant build info so i can manually emulate the Build.pm process | 01:32 | |
| from the output | |||
| the only thing i can think of would be a rakudo bug somehow related to a rw class attribute not updating its value on reassignment, but thats grasping at straws | 01:35 | ||
| jdv79: `list` command is enabled for cpan. im using 100 results, if thats too high let me know | 01:36 | ||
|
03:24
Kassandry joined
|
|||
| jdv79 | ugexe: i don't see it in the zef repo | 05:25 | |
| ugexe | the commit message doesnt show the `zef list` part of the message because of the ` i guess. it just says "for CPAN (most recent 100)" | 05:28 | |
| github.com/ugexe/zef/commit/19f21f...a0c67c7335 | 05:29 | ||
| jdv79 | oh, my mistake. thanks! | ||
| its curious that you have mirrors in the cpan code. the p5 mc has no mirror afaik. | 05:37 | ||
| i guess we could if we dumped the db periodically for mc6. idk how realistic that is what with user info and lag issues. | 05:38 | ||
| ugexe | its mostly because thats the framework that was already there for p6c where i do use the mirrors. but, something like mirrors could be used for absolutifying the download_url against cpan/backpan mirrors | 05:42 | |
| jdv79 | right | 05:43 | |
|
07:03
domidumont joined
07:09
domidumont joined
07:48
FROGGS joined
09:41
leont joined
11:14
FROGGS joined
12:07
ribasushi joined
12:40
sufrostico joined
14:06
prammer joined
14:28
sufrostico joined
15:03
domidumont joined
15:36
domidumont joined
16:06
domidumont joined
17:27
prammer joined
17:37
prammer joined
18:01
Kassandry joined
18:27
leont joined
18:52
patrickz joined
19:06
FROGGS joined
19:10
domidumont joined,
prammer joined
19:13
sufrostico joined
|
|||
| patrickz | ugexe: paste.debian.net/379645/ | 19:17 | |
| It seems the build silently failed. My perl was borked | 19:18 | ||
| I missed the -fPIC option... | 19:22 | ||
| jdv79 | i've done that | 19:25 | |
| nine | I wonder if that is easily detectable | 19:43 | |
| Seems like perl -MConfig -E 'say %Config{cccdlflags}' gives me "cccdlflags-fPIC". But I don't know if it always does that for a perl built with -Duseshrplib | 19:44 | ||
| Unfortunatly -Duseshrplib is not the only way to get a suitable libperl, as we can use a libperl.a as well. | 19:46 | ||
|
19:52
prammer joined
19:55
leont joined
20:02
prammer joined
20:12
prammer joined
|
|||
| jdv79 | does $Config{libperl} not cover both? | 20:20 | |
|
21:13
prammer joined
|
|||
| ugexe | patrickz: ahh, glad we know. its hard to handle Build.pm because some people die, some dont return any value from the method, and some dont seem to handle their errors so the script still exits 0 even though it fails and outputs on stderr | 21:28 | |
| leont wonders what the question was that jdv79 answered | 21:29 | ||
|
21:39
prammer joined
21:42
b2gills1 joined
21:49
ribasushi joined
|
|||
| jdv79 | leont: irclog.perlgeek.de/perl6-toolchain/...i_12019353 | 22:07 | |
| leont | Right, there's a log :-) | 22:13 | |
| ugexe | hmm seems like im getting some of that run/shell giving me the wrong exitcode stuff | 22:14 | |
| leont | Using the configuration values for compiling it tricky, they don't always mean what you'd expect | 22:15 | |
| cccdlflags will always contain whatever flags are needed on that platform for compiling object files for a shared object | 22:16 | ||
| (erm, if that perl is shrplib) | 22:17 | ||
|
22:21
b2gills joined
22:23
camelia joined
|
|||
| ugexe | seems i can get the correct exitcode by switching `$proc.out.close` to `$proc.out.close unless +@err;`, where `my @err = $proc.err.lines;` | 23:12 | |
| leont | Could it be a pipe buffering issue? | 23:25 | |
| ugexe | dunno. if i simply don't call $proc.out.close it also works. but not closing $proc.out when the process was a success has its own ramifications on exitcode | 23:40 | |
| jdv79 | ugexe: paste.scsys.co.uk/505240 | 23:48 | |
| you master branch isn't terribly stable lately i've noticed:) | 23:49 | ||
| i just wanted to try the list feature | |||
| oh. nm. my fault. | 23:51 | ||
| Skipping... (use :force to override) is kinda weird when someone doesn't know that --force is :force maybe | 23:52 | ||
| patrickz | All of this instability stuff will be loads better once cpan lifts of because people will start doing actual *releases* of their modules with version number and all. | 23:56 | |
|
23:57
ranguard joined
|
|||
| jdv79 | its a long road to that point | 23:57 | |
| ugexe | yeah its because the message is generated in Zef::App. if zef::app was being used without bin/zef i wouldnt want it to say --force. i do agree with you however, i just picked something to fill in until i pass exceptions back or something | 23:58 | |
| what really surprises me is `zef :force -v install xxx` actually does a force install | |||
| jdv79 | its not an exceptional cause though is it? | ||
| really? | |||
| ugexe | yeah | 23:59 | |