Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2015-07-18

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

All times shown according to UTC.

Time Nick Message
02:34 mtj- joined #metacpan
02:50 mtj- joined #metacpan
05:57 chansen joined #metacpan
07:16 oiami joined #metacpan
09:32 haarg getting back to CPAN::Changes
09:32 haarg the current indent setting continue to piss me off
09:33 haarg they are imo ugly and also don't make any sense with nesting
11:46 punter joined #metacpan
12:49 vroom joined #metacpan
12:49 haarg kentnl: https://github.com/haarg/CPAN-Changelog
12:50 haarg can you test that against the stuff you have that subclasses CPAN::Changes::* and tell me what you think?
12:51 haarg there are some minor test breakages, but are unlikely to effect anything but the tests
12:56 vroom joined #metacpan
13:17 vroom joined #metacpan
14:52 oiami joined #metacpan
16:33 kentnl haarg: install of that repo fails tests: Can't open corpus/DBIx-Class: No such file or directory at t/parse.t line 10.
16:34 haarg ignore test failures for now
16:34 haarg there's still some file layout stuff that i need to fix
16:40 kentnl there's 2 minor whitespace changes between this and the previous release.
16:40 haarg one is a trailing newline
16:40 kentnl Annoyingly, my guard code that is supposed to turn off the test if the format changes is not triggering :(
16:40 haarg the other is an indentation change when you serialize a group
16:41 haarg if you serialize a release or a whole changelog, the indentation should be consistent with the old code
16:41 haarg also, ->serialize(width => 120)
16:42 haarg doesn't use Text::Wrap
16:42 kentnl yeah. My guard does whole-log as a comparison, but the test only does group serialization :)
16:42 haarg basically, this seems close enough to the old style that it won't cause any meaningful churn
16:42 haarg and you should be able to trivially adjust your tests
16:43 haarg since i don't think anyone else relies on the specific group serialization
17:02 haarg kentnl: so does this seem reasonable if i released it as a new version of CPAN::Changes?
17:03 haarg i still have to decide on some code org stuff, but the current behavior seems acceptable to me
17:03 kentnl go for it, I was about to go re-code all my stuff to guard against weird side effects, but it will be easier if I'm writing them to assume 0.5 output instead of 0.4 output.
17:05 haarg ok.  it might be a bit before i get it finished though.
17:08 kentnl what will that version release as?
17:09 * kentnl assumes what I'm seeing will ship as 0.500001
17:12 haarg 0.500_001 initially
17:15 kentnl hmm. Strange. There's some bug I've hit where my workflow thing is getting empty changefiles emitted.
17:18 kentnl haarg: ugh. I have code somewhere directly setting $changes->{'releases'}->{ $version } = $object
17:19 haarg you... probably shouldn't have done that
17:19 kentnl what should I do instead?
17:20 haarg what are you replacing the release with?
17:24 kentnl a CPAN::Changes::Release object of some kind.
17:25 haarg previously, you should have been able to do ->add_release
17:25 haarg in the new version, that isn't removing the old releases, which i'll change i guess
17:26 haarg need to think about how i'm handling sorting for releases
17:27 kentnl yeah. I tried bodging it with push, and ... got the verison order backwards :D
17:29 haarg the other option is my @releases = $changes->releases; push @releases, ...; $changes->releases(\@releases)
17:29 kentnl and the other thing I think I've spotted is previously, attach_group deduplicated on identical names. Now it doesn't.
17:30 haarg ah, i'll fix that
17:30 haarg most of this code is backcompat code :/
17:31 haarg with a better data model, most of the complexity disappears
17:31 haarg ... and then i have to add it back in
17:33 haarg kentnl: if you are feeling ambitious, you could write some tests for this stuff :)
18:23 haarg backlog mostly cleaned out.  i guess i should work on asset caching on metacpan.
22:30 jnbek joined #metacpan
22:39 sivoais joined #metacpan

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