Perl 6 - the future is here, just unevenly distributed

IRC log for #pr-challenge, 2015-02-12

| Channels | #pr-challenge index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:04 Ptolemarch joined #pr-challenge
00:35 boreas joined #pr-challenge
02:46 Su-Shee_ joined #pr-challenge
02:56 daveh joined #pr-challenge
03:06 vairav joined #pr-challenge
03:10 rvandam joined #pr-challenge
03:18 boreas joined #pr-challenge
03:31 vairav joined #pr-challenge
03:45 Ptolemarch joined #pr-challenge
04:39 n2plato joined #pr-challenge
04:47 BRAD joined #pr-challenge
04:51 vairav joined #pr-challenge
05:34 Ptolemarch joined #pr-challenge
05:56 vairav joined #pr-challenge
06:38 vairav joined #pr-challenge
07:18 boreas joined #pr-challenge
07:23 Ptolemarch joined #pr-challenge
07:35 reneeb joined #pr-challenge
08:05 reneeb joined #pr-challenge
08:08 vairav joined #pr-challenge
08:22 neilb joined #pr-challenge
08:27 neilb_ joined #pr-challenge
09:00 Lee what's the consensus on MANIFEST? my dist is http://cpants.cpanauthors.org/dist/MooseX-GlobRef - the only Kwalitee fail is a missing MANIFEST, but this is auto generated (the HEAD commit explicitly removes the MANIFEST as "Manifest is autogenerated now.")
09:01 ptc Lee: the MANIFEST should still be packaged with the distribution, even if it is automatically generated
09:01 Lee k, thanks
09:02 ptc Lee: it doesn't need to be in the repo anymore, but it should be available after building the distribution tarball
09:02 ptc does your dist use Dist::Zilla?
09:02 pink_mist this is why dzil is so handy; it just takes care of these things for you
09:03 Lee erm looks like Module::Build
09:04 * pink_mist wouldn't suggest making a PR that switches to dzil though; that would be unlikely to fly since it's so big :>
09:04 ptc pink_mist: that's something to discuss with the module author; it could be a good idea in some cases
09:05 Lee yeah, doing that without discussion would be a dick move
09:05 ptc also, maybe the author hasn't got around to it
09:05 pink_mist fair enough
09:06 pink_mist I've seen a lot of resistance to switching to dzil though from a lot of people who aren't already sold on the cool-aid that it is :P
09:07 ptc Lee: if you run 'perl Build.PL; ./Build dist', does the MANIFEST appear in the resulting tarball?
09:07 ptc pink_mist: true.  :-)
09:07 ptc pink_mist: it sometimes feels like such a huge, complex beast to use and configure nicely
09:09 Lee ptc: yes, it's there in the .gz
09:10 Lee could be the missing MYMETA.json from MANIFEST.SKIP
09:12 Ptolemarch joined #pr-challenge
09:15 punter joined #pr-challenge
09:17 Relequestual joined #pr-challenge
09:20 btyler left #pr-challenge
09:22 veryrusty joined #pr-challenge
09:24 sue joined #pr-challenge
09:38 ptc Lee: don't think it's the missing MYMETA.json, as that's also automatically generated by the user after running the build scripts
09:38 ptc it's probably the case that a new release needs to be made :-)
09:40 ptc Lee: just had a look at the cpants output.  The MANIFEST is there, it just doesn't match the contents of the dist
09:43 Lee ptc: so why would that be different if it's automatically generated?
09:44 ptc Lee: dunno.  I see that SIGNATURE appears twice in the MANIFEST on metacpan (https://metacpan.org/source/DEXTER/MooseX-GlobRef-0.0701/MANIFEST)
09:44 ptc Lee: maybe that's a/the problem?
09:44 Lee hmmmmm
09:47 ptc Lee: there's a kwalitee.t test in xt/ .  Does it give helpful output?
09:48 ptc Lee: this test is supposed to be basically the same set of tests as the CPANTS kwalitee tests
09:49 * Lee installs Test::Kwalitee ...
09:49 Lee no probs
09:49 Lee i think it's probably the duplicate SIG - i guess it just needs a new release building
09:52 ptc Lee: also have a look at the output of './Build manifest' to see if the output matches that which one expects
09:55 Lee oh interesting. Test::CheckChanges won't install due to t/multi.t failing (skipped: case insensitive filesystem) but i do my dev on a case sensitive partition
09:55 pink_mist a skip shouldn't count as a fail
09:56 Lee Parse errors: More than one plan found in TAP output
09:56 pink_mist sounds like it might have erroneous tests then
09:57 ptc or maybe it doesn't work well on windows...]
09:57 pink_mist also, are you sure it's not running its tests on a case insensitive filesystem rather than the one you're doing your dev on?
09:57 ptc getting it to build and test without errors on windows is very worthwhile PR :-)
09:57 Lee oh wait, ~/.cpan
09:57 pink_mist indeed
09:57 * Lee symlinks that to a sensible location
09:58 Lee bloody Mac's...
10:00 Lee right, xt failures to investigate and fix
10:02 ptc ye gods, aren't macs supposed to be unix?  I.e. case sensitive?
10:02 Lee no, not out of the box
10:02 Lee hence first step: create a sensible parition for dev :D
10:03 * ptc remembers frustrating development on macs pre MacOSX
10:09 Lee right, META.json is missing from MANIFEST
10:09 Lee ptc++ # much help, thanks
10:12 ptc Lee: no probs :-)
10:21 Lee https://github.com/dex4er/perl-MooseX-GlobRef/pull/3 # hurrah
10:45 sue joined #pr-challenge
11:01 Ptolemarch joined #pr-challenge
11:16 ptc Lee: yay!
11:30 sue joined #pr-challenge
11:39 daveh HTML::Prototype, untouched since 2006
11:39 daveh er, yay?
11:41 ptc daveh: yay, since Lee submitted a PR
12:05 neilb joined #pr-challenge
12:15 athos joined #pr-challenge
12:19 daveh his module? awse
12:19 daveh did pretty well in cpantesters. but that might be because the tests are shit :)
12:37 sue joined #pr-challenge
12:50 Ptolemarch joined #pr-challenge
12:54 boreas joined #pr-challenge
12:59 kanashiro joined #pr-challenge
13:04 boreas joined #pr-challenge
13:56 fredericop joined #pr-challenge
14:14 BooK deven: so, any ideas for your PR?
14:14 tinypig joined #pr-challenge
14:18 sue joined #pr-challenge
14:27 gryphon joined #pr-challenge
14:31 neilb joined #pr-challenge
14:33 punter joined #pr-challenge
14:35 neilb_ joined #pr-challenge
14:36 deven BooK: Haven't had time to work on it yet, but I might get a chance this weekend.  Do you have suggestions?
14:39 Ptolemarch joined #pr-challenge
14:41 sivoais joined #pr-challenge
14:43 BooK deven: not really. I've made a few commits after our discussion, but mostly to package it with dzil
14:44 deven BooK: Are you saying you've left me nothing to do? ;)
14:44 BooK oopsie
14:44 BooK one good way forward it to run the install step in a temporary directory and then commit all those files
14:45 deven BooK: I thought you said something about already doing so?
14:45 BooK no, it's not doing that right now
14:45 deven BooK: So you do still have the issue of unchanged files not ending up in the commit?
14:45 BooK yes, and doing the install in a temporary directory would solve it
14:46 deven Yes, it seems like the obvious solution, if not the only one!
14:46 BooK my previous solution was to wipe the entire repository ;-)
14:46 deven Um, I thought you wanted multiple versions in the same repository?
14:46 BooK https://github.com/book/Git-CPAN-Hook/blob/master/lib/Git/CPAN/Hook.pm#L104 # munging with local::lib should probably happen in this sub, around the if block
14:47 BooK sorry, I meant work tree :-)
14:47 deven Does the module work under Windows?
14:47 BooK would be nice
14:47 deven Meaning it doesn't?
14:47 BooK meaning I don't know, and I don't see why it wouldn't
14:48 deven ah
14:48 BooK so basically, before running the install step, munge stuff so that the files are installed in a temp dir,
14:48 BooK then run the commit step to pick all the files added to the temp dir
14:48 deven CPAN testers doesn't show that answer?
14:48 BooK haven't looked yet
14:48 BooK then possibly "install" the files again, this time on top of the repository
14:49 BooK (and via git, not via CPAN)
14:49 deven I'm assuming you can do the temp install the same way you would with a filesystem like AFS where you install to a different path than you use it from.
14:49 deven (I think AFS was an example where that was needed anyhow.)
14:49 BooK I think local::lib has all the needed bits
14:49 deven probably does
14:49 deven Install again via git?  I'm not sure what you mean by that.
14:50 deven Use a git checkout to deploy the module for the system to use?
14:50 BooK pick up the files in the commit and put then in the work tree
14:50 BooK cherry-pick the commit you just made
14:50 deven hmm
14:51 BooK not sure how that works when there already are some files where you want to put new ones
14:51 BooK but that's a git issue
14:51 deven I guess I'm not entirely seeing the use case you're targeting here
14:51 BooK deven: installing dependencies?
14:52 BooK I need to stay close to the CPAN install step, because I never know how deep in the dependency traversel we currently are
14:52 deven I remember you mentioning git bisect, which does sound interesting, but I guess I'm still a little fuzzy on the overall intent of the module.
14:52 BooK installed distribution = independent tree object
14:53 BooK a bunch of installed modules = stack all those trees on top of each other
14:53 BooK move around freely around different installation sets
14:53 BooK eg for testing purposes
14:54 deven Sort of a "perlbrew for modules" idea?
14:54 BooK stack installation sets as commits on a branch, and move around in the branch using bisect
14:54 BooK kinda
14:54 deven hmm
14:54 BooK so there also some work needed on a command-line tool
14:54 BooK because it could basically become its own cpan client, that only delegates to a real cpan client when whatever dist you want is not in the repo yet
14:57 BooK git cpan create my-moo-worktree ; git cpan install Moo ; git cpan checkout my-moose-worktree ; git cpan install Moose ; ( test my module with moose ) ; git cpan checkout my-moo-worktree ; ( test my module with moo )
14:58 BooK git cpan install-all Moo # I expect this to be quite complex, especially with the combinatorial explosion of all versions of all deps, if we go that way)
14:58 vroom joined #pr-challenge
15:00 BooK the above example was assuming a module that works with both Moo and Moose. there are other ways to do that at the moment (Test modules that pretend a given module is not installed)
15:03 lbrown joined #pr-challenge
15:03 Ptolemarch joined #pr-challenge
15:51 neilb joined #pr-challenge
16:17 kanashiro joined #pr-challenge
16:27 solargroovy joined #pr-challenge
16:29 daveh so did someone say they had HTML::Prototype?
16:39 n2plato joined #pr-challenge
16:50 reneeb daveh: HTML::Prototype has no github repository (at least it's not mentioned in the Metadata). So it's not assigned yet. You can work on it ;-)
17:00 nawglan joined #pr-challenge
17:05 boreas joined #pr-challenge
17:17 n2plato joined #pr-challenge
17:32 jevin joined #pr-challenge
17:33 solargroovy joined #pr-challenge
17:37 vairav joined #pr-challenge
17:40 neilb joined #pr-challenge
18:03 vairav joined #pr-challenge
18:53 spechenin joined #pr-challenge
18:56 n2plato joined #pr-challenge
19:10 gansh joined #pr-challenge
19:14 vairav joined #pr-challenge
19:30 Ptolemarch joined #pr-challenge
19:39 punter joined #pr-challenge
20:03 vairav joined #pr-challenge
20:04 spechenin joined #pr-challenge
20:09 sue joined #pr-challenge
20:29 n2plato joined #pr-challenge
20:34 n2plato joined #pr-challenge
21:13 vairav joined #pr-challenge
21:19 Ptolemarch joined #pr-challenge
21:38 solargroovy joined #pr-challenge
21:40 Ptolemarch joined #pr-challenge
21:47 nawglan left #pr-challenge
22:14 absolut_todd joined #pr-challenge
22:17 vairav joined #pr-challenge
23:11 boreas joined #pr-challenge
23:15 kid51 joined #pr-challenge
23:58 boreas joined #pr-challenge
23:59 boreas_ joined #pr-challenge

| Channels | #pr-challenge index | Today | | Search | Google Search | Plain-Text | summary