|
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 20 February 2016. |
|||
|
00:13
autarch joined
|
|||
| ugexe | looks like it should work. having $version/$!version in there makes its slightly harder to follow with .version-matcher in the same method, but it doesnt make much sense to name them something else either | 02:25 | |
|
02:48
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 | ||
| ugexe | i wonder if this can't somehow use :api<0> and :api<1> for CURI itself. otherwise it will probably get ugly supported multiple formats in various CURI methods | 03:13 | |
| vs 2 separate CURI, each implementing only the version format they support | 03:14 | ||
| s/vs/ie/ | |||
|
04:34
cognominal joined
|
|||
| nine | Yes, I've thought about using polymorphism, too. For the little changes we have for now, it's probably overkill, but once we got more changes and formats, a little refactoring in that direction will help. | 09:28 | |
| The functionality can stay within private classes, so it doesn't change the API. | 09:29 | ||
|
11:10
hankache joined
11:32
domidumont joined
12:19
FROGGS joined
12:53
FROGGS joined
14:19
autarch joined
|
|||
| nine | Store logical file names in precomp files | 16:25 | |
| We now use logical file names for annotating sources and construct absolute paths when printing stack traces. This avoids storing absolute paths inprecompiled files which is a problem for packaging modules for Linux distributions. | |||
| It also allows for us to add the name of the requested module to the backtrace output, alleviating the problem with the undeciferable SHA-1 file names. | |||
| Review: github.com/rakudo/rakudo/commit/37...a920730d01 | |||
|
17:28
domidumont joined
|
|||
| b2gills | nine: I think you should look into Pinto/Stratopan for managing CPAN modules at work | 17:42 | |
| ugexe | heh, ive been using something similar to this (split on #) for filenames to allow commit ids in the uri/path | 17:51 | |
| never even occured it could work for precomp | 17:53 | ||
| mst | b2gills: carton and a local::lib is generally quite sufficient | 17:56 | |
| the pinto author isn't really a toolchain person so it's not always systems quality code | |||
| if you've got a bunch of your own distributions and don't want to run a minicpan, pinto's likely a reasonable approach | 17:58 | ||
| but you don't *usually* need it | |||
|
18:01
sufrostico joined
|
|||
| b2gills | I just compile my own Perls in /opt/perl-5.22.0 etc and have a small tool that creates a symlink in /opt/perl/bin/perl-5.22.0 that points to the perl binary. So I don't really know how well any tool used to manage Perl works. | 18:01 | |
| mst | honestly, in production I do basically that | 18:04 | |
| then stack a local::lib per app atop them with carton | |||
| you can use plenv if you like but I've never really felt the need | 18:05 | ||
| and on debian systems half the time I just stack a local::lib atop the system perl | |||
| because it's generally good enough for my purposes | |||
|
18:42
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 | ||
| mst | nine: so, yes, approve entirely :D | 18:53 | |
| ugexe | is there a name/term for a module that is installed and nothing else depends on? like orphan, but without knowing if the package was originally installed as a dependency or not? | 19:14 | |
| mst | depending on which way you look at it, root or leaf perhaps | 19:20 | |
| ugexe | ah yes, `leaves` is what fedora/yum uses | 19:26 | |
|
20:30
sufrostico joined
|
|||