|
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 27 May 2016. |
|||
|
01:48
ilbot3 joined
|
|||
| moderator | 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 | ||
|
05:35
domidumont joined
05:40
domidumont joined
06:15
domidumont joined
06:52
domidumont joined
07:56
Cheery_ joined
08:00
llfourn joined
12:53
sufrostico joined
12:59
cognominal joined
14:25
sufrostico joined
15:47
sufrostico joined
|
|||
| ugexe | nine: will there ultimately be any difference between requiring a module by name vs by path? | 16:50 | |
| nine | Haven't given requiring modules by path much thought lately | 16:51 | |
| ugexe | sometime in the last month(?) `perl6 -I. 'require <Build.pm>; ::("Build").new;'" starting giving "Failure\\nin any at gen/moar/m-Metamodel.nqp line 3055". Its fixed with s/require <Build.pm>/require ::("Build")/ | 16:54 | |
| gist.github.com/ugexe/fa833539cf77...24e9b982d6 # heres the the details, maybe you can make sense of it | 17:00 | ||
|
17:19
sufrostico joined
17:39
domidumont joined
17:50
sufrostico joined
|
|||
| llfourn | ugexe: if you rm -rf .precomp does it work? | 17:58 | |
| ugexe | no, but let me add no precompilation to the Build.pm file first | 17:59 | |
| llfourn | yeah that should do it | 18:00 | |
| ugexe | same deal | ||
| llfourn | hmm | ||
| ugexe | seems like its screwing up the symbol name | ||
| llfourn | oh wait | 18:01 | |
| ugexe | require <Build.pm>; itself doesnt error. its only once you try using a symbol from that | ||
| llfourn | Build.pm won't work that way | ||
| actually it should nvm | |||
| it won't return Build but GLOBAL::Build should be there | |||
| what is the contents of Build.pm? | |||
| ugexe | its from Inline::Perl5 | 18:02 | |
| but the contents don't seem to matter | |||
| llfourn | m: require Test; say ::('Test') | ||
| camelia | rakudo-moar 4a7eaa: OUTPUTĀ«(Test)ā¤Ā» | ||
| llfourn | m: say $*CWD | 18:03 | |
| camelia | rakudo-moar 4a7eaa: OUTPUTĀ«"/home/camelia".IOā¤Ā» | ||
| ugexe | require <Build.pm>; say GLOBAL::Build; # no such symbol Build | ||
| llfourn | where is Test.pm relative to where we are? | ||
| require './rakudo/lib/Test.pm'; #maybe | |||
| m: require './rakudo/lib/Test.pm'; #maybe | |||
| camelia | rakudo-moar 4a7eaa: OUTPUTĀ«Could not find ./rakudo/lib/Test.pm in:ā¤ /home/camelia/.perl6ā¤ /home/camelia/rakudo-m-inst-1/share/perl6/siteā¤ /home/camelia/rakudo-m-inst-1/share/perl6/vendorā¤ /home/camelia/rakudo-m-inst-1/share/perl6ā¤ CompUnit::Repository::Absolā¦Ā» | ||
| llfourn | m: require './rakudo-m-inst-1/lib/Test.pm'; #maybe | 18:04 | |
| camelia | rakudo-moar 4a7eaa: OUTPUTĀ«Could not find ./rakudo-m-inst-1/lib/Test.pm in:ā¤ /home/camelia/.perl6ā¤ /home/camelia/rakudo-m-inst-1/share/perl6/siteā¤ /home/camelia/rakudo-m-inst-1/share/perl6/vendorā¤ /home/camelia/rakudo-m-inst-1/share/perl6ā¤ CompUnit::Repositoā¦Ā» | ||
| llfourn | m: require './rakudo-m-inst-1/lib/Test.pm6'; #maybe | ||
| camelia | rakudo-moar 4a7eaa: OUTPUTĀ«Could not find ./rakudo-m-inst-1/lib/Test.pm6 in:ā¤ /home/camelia/.perl6ā¤ /home/camelia/rakudo-m-inst-1/share/perl6/siteā¤ /home/camelia/rakudo-m-inst-1/share/perl6/vendorā¤ /home/camelia/rakudo-m-inst-1/share/perl6ā¤ CompUnit::Repositā¦Ā» | ||
| ugexe | github.com/ugexe/zef/commit/e9efde...0e34c78bf7 # this shows the workaround from filename to short name that was required | 18:08 | |
| llfourn | :\\ it should work and we have tests for it | ||
| ugexe | yeah it used to work. the error of just "Failed" makes it even stranger | 18:09 | |
| llfourn | m: say './rakudo-m-inst-1/lib/Test.pm6'.e; | 18:10 | |
| camelia | rakudo-moar 4a7eaa: OUTPUTĀ«Method 'e' not found for invocant of class 'Str'ā¤ in block <unit> at /tmp/lmnvTZkUe6 line 1ā¤ā¤Ā» | ||
| llfourn | m: say './rakudo-m-inst-1/lib/Test.pm6'.IO.e; | ||
| camelia | rakudo-moar 4a7eaa: OUTPUTĀ«Falseā¤Ā» | ||
| ugexe | m: require "/home/camelia/rakudo-m-inst-1/share/perl6/sources/C712FE6969F786C9380D643DF17E85D06868219E"; say ::("Test"); # this works, so maybe its class names? | 18:16 | |
| camelia | rakudo-moar 4a7eaa: OUTPUTĀ«(Test)ā¤Ā» | ||
| ugexe | nope, because this doesnt fail | 18:18 | |
| m: require "/home/camelia/rakudo-m-inst-1/share/perl6/sources/4CE2F6717405FCBF7F9D888C3BDF76C1127F8CB1"; say ::("TAP::Harness").new; | |||
| camelia | rakudo-moar 4a7eaa: OUTPUTĀ«Type check failed in assignment to $!output; expected IO::Handle but got IO::Handle.new(:path(...ā¤ in block <unit> at /tmp/PIzNFxF_hd line 1ā¤ā¤Ā» | ||
| llfourn | ugexe: it seems you need absolute path... | 18:20 | |
| but that can't be right becasue the tests don't have abs path | |||
| ugexe | it seems to have something to do with `use` statements in the Build.pm | 18:21 | |
| if i leave the `use` statements and empty the class Build, it still fails. If I remove the `use` statements as well it works | |||
| but it must be something specific about the `use`d modules, because adding a `use Test;` to the to the top of the Build.pm in place of the others and it works still | 18:22 | ||
|
18:23
FROGGS joined
|
|||
| llfourn | It's weird I'm just having problems requiring things by relative paths in general | 18:26 | |
| ugexe | even with just the basename that exists in the CWD? | 18:27 | |
| llfourn | yeah if I cd into lib and do require "Foo.pm6"; I get Couldn't find Foo.pm6 error. | 18:28 | |
| but "Foo.pm6".IO.e returns true | |||
| ugexe | -I. ? | ||
| llfourn | you need that? | 18:29 | |
| ugexe | i have it in this code base, so i imagine it was needed at one point at least | ||
| llfourn | it works -- but I don't get why I need it :S | ||
| oh well that solves that problem | |||
| I'm off gl figuring out that require bug :) | 18:31 | ||
| ugexe | you'll be back! | 18:33 | |
| fwiw it works if i remove `use Panda::Common; use Panda::Builder; use Shell::Command;` (leaving use v6 and use LibraryMake), but the RAKUDO_MODULE_DEBUG doesn't show anything fishy going on | 18:34 | ||
| hmm i might be misreading this output, but it looks like maybe the ordering is off for a dependency of a dependency | 18:44 | ||
| it worked in 2016.04, so sometime after that | 18:53 | ||
|
18:54
hankache joined
|
|||
| nine | A bit of a stretch, but maybe it's rt.perl.org/Ticket/Display.html?id=128156 after all? | 18:54 | |
| ugexe | i thought that at first but the error doesnt match | 19:00 | |
| it started immediately after relocatable precomp | 19:31 | ||
|
19:40
sufrostico joined
|
|||
| nine | Ok, first fix available in rt128156_fix_precomp_deps_validation | 19:45 | |
| ugexe: can you try anyway? | |||
| ugexe | sure, building now | 19:46 | |
| same error | 19:57 | ||
|
20:09
sufrostico joined
|
|||
| nine | Ok, thanks for testing | 20:17 | |
|
20:18
sufrostico joined
|
|||
| ugexe | github.com/rakudo/rakudo/commit/bf...167ff13ac4 # its this commit | 20:26 | |
| nine | An educated guess would be the change in CompUnit | 20:32 | |
|
20:35
sufrostico joined
|
|||