Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2016-07-29

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

All times shown according to UTC.

Time Nick Message
04:42 oiami joined #metacpan
05:51 Khisanth joined #metacpan
07:21 neilb joined #metacpan
08:02 Relequestual joined #metacpan
08:12 neilb joined #metacpan
09:49 dolmen joined #metacpan
14:56 AirDisa joined #metacpan
15:07 oalders mickey: the latest MetaCPAN::Client breaks OrePAN2 https://github.com/tokuhirom/OrePAN2/issues/42
15:09 mickey oalders: i'll look into it later, assuming this is the single valued array change... should be easy to fix
15:09 oalders i _think_ we only want _single_valued_arrayref_to_scalar() on v1
15:09 oalders it's not urgent, but it will break some of our $work code when we upgrade modules next :)
15:11 mickey heading home now, will fix it later... i knew it might cause issues but i rather having the single values sorted in the client so it's not a user problem to check how to read the attributes
15:11 oalders yeah, for sure.  the client is there to fix things up for the user
15:16 mst huh
15:16 mst orepan2 looks quite close to something I was considering implementing
15:16 mst except it appears to be a shitload more complicated
15:16 mst what's it using MetaCPAN for?
15:19 y1mmm joined #metacpan
15:24 oalders mst: it's way faster to use MetaCPAN than to unwrap X tarballs to check which modules they provide
15:25 oalders mst: https://github.com/tokuhirom/OrePAN2/issues/11#issuecomment-54884667
15:27 oalders the internals are a bit complicated.  not sure why it was designed this way
15:37 mst I don't understand
15:37 mst oh, hrm, you're doing a full index of pinned dists as well as non-cpan dists?
15:45 mst oalders: ?
15:45 mst maybe I should just dismantle it and read
15:47 oalders mst: correct. when you rebuild your 02packages
15:59 mst so does this work via being basically a minicpan, plus whatever you've added?
15:59 oalders yes
16:00 oalders but a minicpan in the sense that you can just cherry-pick from CPAN
16:00 oalders and then throw in your DarkPAN stuff
16:02 mst I'm kinda confused here
16:03 mst you need to manually inject every dist+version you want installable?
16:06 oalders That's what we do at $work, but you should be able to just use a minicpan as well.  You'd just need OrePAN2 to index it.
16:17 mst oalders: right, I was thinking more of basically having a minicpan, then a darkpan overlay type thing, then do something like orepan's index merging
16:18 oalders mst: that should work
16:19 mst oh, it'll definitely work
16:19 mst I'm just trying to figure out why my gut feeling is still "I'm better writing my own"
16:20 mst and I think the answer is that I want to be slightly more clever along axes where orepan is intentionally stupid, and a lot stupider along axes where orepan is intentionally clever
16:21 mst I suspect the answer here is going to be "write what I was thinking about anyway, document OrePAN2 as being 'my thing is not better, just different, go look at this as well'"
16:24 mst bleh
16:24 * mst hates reinventing wheels when he can't properly articulate why he wants to
16:36 oalders :)
16:36 oalders let me know what you come up with
16:59 ether I still need to rewrite my $work's cpan wrangling code. it is a horrible mess right now that needs to be babied by hand whenever one wants to make a change
16:59 ether at least I've talked the rest of my team into letting me do that work, so I can keep things unbroken
16:59 ether but still
16:59 ether one of the wrapper scripts actually uses Orepan2 for something I think
17:00 ether we stuff our cpan tarballs into git and then install them per-machine, but the inhouse code is just all in $reporoot/lib, not in an installable dist
17:01 mst oalders: ironically, I think if I create what I want to create, you'll be able to use it with an orepan index too ;)
17:01 mst we'll see
17:14 melo joined #metacpan
17:15 mst filed https://github.com/hirose31/IO-File-AtomicChange/issues/2
17:15 mst (because wtf)
17:46 ranguard when you've all re-written the world, please let me know so I can replace our $work thing as well :)
17:48 ranguard mst: it's still http://blogs.perl.org/users/michael_j/2011/04/deploying-perl-code-with-git-locallib-minicpan-and-cpanminus.html if your looking for requirements :)
18:23 mst ranguard: ta
18:26 alh_ joined #metacpan
18:33 neilb joined #metacpan
20:44 ribasushi joined #metacpan
20:50 melo joined #metacpan
22:02 mickey oalders: https://github.com/tokuhirom/OrePAN2/pull/43
23:10 neilb_ joined #metacpan

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