Perl 6 - the future is here, just unevenly distributed

IRC log for #metacpan, 2014-08-12

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

All times shown according to UTC.

Time Nick Message
00:22 neilb joined #metacpan
00:22 SineSwiper joined #metacpan
00:29 neilb joined #metacpan
01:26 FROGGS__ joined #metacpan
01:45 metacpan joined #metacpan
01:45 metacpan [metacpan-web] oalders created oalders/cookie-secret (+2 new commits): http://git.io/V14YQw
01:45 metacpan metacpan-web/oalders/cookie-secret e68d51d Olaf Alders: Adds Config::JFDI and Data::Printer to cpanfile.
01:45 metacpan metacpan-web/oalders/cookie-secret 736ca5b Olaf Alders: Use a secret when creating cookies.
01:45 metacpan left #metacpan
01:45 dipsy [ Comparing e68d51d7fbe8^...736ca5b4fae4 · CPAN-API/metacpan-web · GitHub ]
01:49 klapperl joined #metacpan
03:58 theory joined #metacpan
04:04 metacpan joined #metacpan
04:04 metacpan [metacpan-web] oalders pushed 2 new commits to oalders/cookie-secret: http://git.io/ms2Ngg
04:04 metacpan metacpan-web/oalders/cookie-secret 683b5e2 Olaf Alders: Adds Git::Sub to cpanfile.
04:04 metacpan metacpan-web/oalders/cookie-secret 0e85a7b Olaf Alders: Get proper path to app when run under test harness.
04:04 metacpan left #metacpan
04:04 dipsy [ Comparing 736ca5b4fae4...0e85a7b36859 · CPAN-API/metacpan-web · GitHub ]
04:06 oiami joined #metacpan
05:13 metacpan joined #metacpan
05:13 metacpan [metacpan-web] rwstauner pushed 1 new commit to master: http://git.io/pd7Cng
05:13 metacpan metacpan-web/master 6fc860e Randy Stauner: Merge gh-1230 to add more feed data
05:13 metacpan left #metacpan
05:13 dipsy [ Merge gh-1230 to add more feed data · 6fc860e · CPAN-API/metacpan-web · GitHub ]
05:19 rwstauner oops
05:19 metacpan joined #metacpan
05:19 metacpan [metacpan-web] rwstauner pushed 13 new commits to master: http://git.io/YPjp8g
05:19 metacpan metacpan-web/master 6940399 oiami: Make feed title more descriptive and add info about author ++ modules
05:19 metacpan metacpan-web/master bb218a3 oiami: Add new RSS feed icon
05:19 metacpan metacpan-web/master 8882fef oiami: Create author page title and add RSS feed subscription button
05:19 metacpan left #metacpan
05:19 dipsy [ Comparing 6fc860e00227...5edf24e9ae50 · CPAN-API/metacpan-web · GitHub ]
05:20 rwstauner good to know: having a pre-commit hook will @#$% up your merge commit if you had to resolve conflicts manually
05:24 * rwstauner &
06:40 dp_ joined #metacpan
07:52 n0body joined #metacpan
08:07 neilb joined #metacpan
08:17 FROGGS joined #metacpan
08:49 neilb joined #metacpan
11:36 rashi joined #metacpan
13:40 sid joined #metacpan
13:59 neilb joined #metacpan
14:06 neilb joined #metacpan
14:45 metacpan joined #metacpan
14:45 metacpan [metacpan-web] rwstauner pushed 3 new commits to master: http://git.io/Uz8vXg
14:45 metacpan metacpan-web/master 2caf237 oiami: Fix wrong element order, let anchor cover icon
14:45 metacpan metacpan-web/master b5cafd6 oiami: Change to use all icons from FontAwesome
14:45 metacpan metacpan-web/master bfc0a9c oiami: Fix misplaced text
14:45 metacpan left #metacpan
14:45 dipsy [ Comparing 5edf24e9ae50...bfc0a9c5369d · CPAN-API/metacpan-web · GitHub ]
16:02 kentnl is there a straight forward way to get a feed of $RELEASES with modules and versions of those modules in the result?
16:03 kentnl sort of like how it currently expands dependencies to long form, and kinda how theres a 'provides => [ Foo , Bar ]' , but I want 'provides => [{ Foo => 5.0 }, { Bar => 5.0 }]
16:06 theory joined #metacpan
16:33 metacpan joined #metacpan
16:33 metacpan [metacpan-web] oalders force-pushed oalders/cookie-secret from 0e85a7b to afbd283: http://git.io/SL2Lgw
16:33 metacpan metacpan-web/oalders/cookie-secret 009ccfa Olaf Alders: Adds Config::JFDI and Data::Printer to cpanfile.
16:33 metacpan metacpan-web/oalders/cookie-secret 4a9bca6 Olaf Alders: Use a secret when creating cookies.
16:33 metacpan metacpan-web/oalders/cookie-secret f0769c5 Olaf Alders: Get proper path to app when run under test harness.
16:33 metacpan left #metacpan
16:33 dipsy [ Commits · CPAN-API/metacpan-web · GitHub ]
16:42 neilb joined #metacpan
16:45 metacpan joined #metacpan
16:45 metacpan [metacpan-web] oalders force-pushed oalders/cookie-secret from afbd283 to 39eb7a7: http://git.io/SL2Lgw
16:45 metacpan metacpan-web/oalders/cookie-secret df81d32 Olaf Alders: Bump Plack::Middleware::Session to 0.22
16:45 metacpan metacpan-web/oalders/cookie-secret 39eb7a7 Olaf Alders: JSON serialization for cookies.
16:45 metacpan left #metacpan
16:45 dipsy [ Commits · CPAN-API/metacpan-web · GitHub ]
16:52 neilb joined #metacpan
17:19 trs kentnl: the dependencies in a release document are from CPAN::Meta, i.e. parsing META.{yml,json}
17:20 trs metacpan simply stores that data structure.
17:20 kentnl trs: it shows it in a different format from CPAN::Meta for some reason, but that was not really my point.
17:20 trs you can't go from recent releases -> modules in a single query, because ES doesn't do joins
17:20 trs but you can do it in two queries.
17:21 kentnl is there an easy way to get the release path from a module?
17:21 * kentnl would be happy enough with that
17:21 kentnl as it is, I can only see 'author' and 'release', but I can't do anything useful there without assuming that $AUTHOR hasn't used any fany cpan path storage tricks and assuming that $AUTHOR has used a sane file extension
17:21 trs kentnl: see the "release" and "author" keys for a module document.
17:22 kentnl ie: sometimes FOO/Bar.tar.gz # is sane, but there's lots of FOO/My-Thingy/Bar.tar.gz and FOO/Whatever.zip
17:22 trs kentnl: while $author/$release won't get you a filename, it _will_ let you find it out via metacpan's api
17:22 trs /release/$author/$release
17:23 trs and release documents have a download_url
17:23 kentnl yeah, I was just hoping to avoid more queries because what I'm doing is already slow enough
17:23 trs well, sorry to burst your bubble. :)
17:23 trs ES doesn't do joins, so many queries are common.
17:24 kentnl Thats a nightmare when you have large ping and a large query response time :(
17:24 trs caching?
17:24 dipsy caching is hard to get right. :(
17:24 kentnl ^
17:24 kentnl and what I'm doing is kinda not caching friendly
17:24 trs nod.
17:24 kentnl ie: I'm trying to implement cpan-outdated using the metacpan API
17:25 kentnl and if what I write proves *slower* than that, there's no point in doing it
17:25 trs maybe you could implement using miyagawa's service used by cpanm ;)
17:26 trs or just download 02packages and call it a day?
17:26 kentnl ( well, I can do things they can't like supporting --dev , and "dont look further back than a week", but those are secondary benefits ime )
17:26 trs nod
17:26 kentnl cpan-outdated does just that. Which is part of its slowness ;)
17:27 trs I can't imagine that using a local flat file with all the data is going to be slower than making lots of queries.
17:27 trs why is using 02packages slow?
17:27 kentnl the other part of its slowness is that it iterates all of 02packages and checks *everything* in it for being installed, which is needless checks if you have a way to go "Only X have been released since we last checked"
17:27 kentnl =~ cpan-changes loads all of @INC into ram.
17:27 trs wth does it iterate 02packages instead of iterating what's on your disk?
17:28 * kentnl could be wrong, I'll double check
17:28 kentnl but either way, you still have to load all @INC into memory
17:28 kentnl and that's slow
17:28 trs you don't have to do it at the same time...
17:29 trs and this is a trivially parallelizable problem.
17:29 kentnl no, but you still have to do it *prior* to cpan-outdated terminating.
17:29 kentnl its not a ram problem, but an IO problem
17:29 kentnl because even if you don't load it into *perl* memory, it will load into VFS Cache memory
17:29 trs nod.
17:30 trs so your proposal is to look up the latest releases since last week (or whatever) and then only check those?  with the idea being that you run it every week?
17:31 kentnl yeah, or every other day, or whatever $INTERVAL
17:31 trs sure
17:31 kentnl some cases its because "yes, cpan-outdated, thats nice, keep reminding me about this thing that is broken and I've tried to update, but it fails, so I don't want to update that today, please remind me when author releases a non broken version"
17:41 * kentnl will just have to write the version and emit "Foo@1.24" output only for the time being #pragmatism #wow #hashtags
17:41 metacpan joined #metacpan
17:41 metacpan [metacpan-web] oalders pushed 1 new commit to oalders/cookie-secret: http://git.io/66E1LQ
17:41 metacpan metacpan-web/oalders/cookie-secret e5d5ce8 vagrant: Bump version of Module::Build::Tiny to 0.037
17:41 metacpan left #metacpan
17:42 dipsy [ Bump version of Module::Build::Tiny to 0.037 · e5d5ce8 · CPAN-API/metacpan-web · GitHub ]
17:57 mo joined #metacpan
18:01 mo joined #metacpan
18:14 ingy neilb: so I can start my Massive Attack at 5pm Friday in Seattle?
18:16 ingy aka #INGYDOSPAUSE
18:16 neilb heh. What's the time in seattle now?
18:16 neilb you -7?
18:20 kentnl ingy: http://kentfredric.github.io/metacpan-status/  # I have the countdown to UTC CPANDay here
18:20 dipsy [ Once-A-Week/Month/Day countdowns ]
18:22 * kentnl thinks he'll get his once-a-week neocpanism out before then *just* in case we do manage to break PAUSE
18:37 * ether finally clues in that I should allow javascript for kentnl's page
18:37 ether all this time I thought he was giving out config data
18:39 ingy kentnl: could you up the precision?
18:39 oalders :)
18:40 ingy neilb: yes
18:40 oalders kentnl: is that url for public consumption then?
18:41 ingy ether: don't do it. internet-cooties--
18:41 neilb ingy: then yes, from 5pm friday you can start uploading!
18:42 * ingy needs to have a target goal
18:42 neilb Combined with Kent, who can start early, by the time I've woken up, maybe ...
18:44 ingy I can maybe get first 100 before bed
18:44 ingy I'm not fully automated yet :(
18:44 ingy next year!
18:51 ingy hmmm https://gist.github.com/e7d7626e9303ef3bb3c8
18:51 dipsy [ foo.pl ]
19:43 kentnl oalders: yes.
19:43 oalders cool
19:45 kentnl ingy: I could, but I don't want to thrash the browser. When it reaches closer to the deadline, the precision factors automatically scale
19:45 kentnl so when you have an hour left, it becomes shown in x.y minutes
19:45 kentnl and when you have a minute left, it shows seconds
19:49 ingy kentnl: I meant to picoseconds plz
19:50 kentnl browsers don't even refresh that fast:p
19:50 ingy WELL FIX THAT!
19:50 kentnl min tick is like 5/10 msec
19:51 kentnl its a good thing really, because otherwise travis would probably attempt to use high resolution timings for no reason and crash my browser faster than it already does
19:53 kentnl ether: also, you may have thought this was a blank page if you never enabled JS on it: http://kentnl.github.io/#/simple
19:53 dipsy [ KENTNL Project Statuses ]
19:54 neilb joined #metacpan
20:01 neilb_ joined #metacpan
20:04 neilb_ joined #metacpan
20:05 neilb_ joined #metacpan
20:08 neilb__ joined #metacpan
20:11 neilb joined #metacpan
21:00 neilb joined #metacpan
21:22 neilb joined #metacpan
21:25 neilb_ joined #metacpan
21:48 grantm joined #metacpan
22:59 daxim joined #metacpan

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