Perl 6 - the future is here, just unevenly distributed

IRC log for #inline, 2014-12-12

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

All times shown according to UTC.

Time Nick Message
00:08 mohawk joined #inline
00:51 GitHub19 [inline-module-pm] ingydotnet pushed 1 new commit to master: http://git.io/elCWJA
00:51 GitHub19 inline-module-pm/master 200046c Ingy döt Net: Fix syntax error
00:53 GitHub129 [inline-module-pm] ingydotnet comment on issue #12: @daoswald, I think PRD is needed to build the installable .so during user
01:28 danaj joined #inline
01:42 GitHub12 [inline-module-pm] ingydotnet pushed 1 new commit to master: http://git.io/AZWVKw
01:42 GitHub12 inline-module-pm/master 024423a Ingy döt Net: Update MANIFEST with proper files in distdir
01:44 GitHub104 [inline-module-pm] ingydotnet comment on issue #5: Fixed in handle_distdir.
01:44 GitHub61 [inline-module-pm] ingydotnet closed issue #5: distdir fixup needs to update MANIFEST http://git.io/RmxJOg
01:58 GitHub135 [inline-module-pm] ingydotnet pushed 1 new commit to master: http://git.io/PCFXAA
01:58 GitHub135 inline-module-pm/master 178f8ad Ingy döt Net: Constant fold out DEBUG code.
02:45 GitHub149 [inline-module-pm] ingydotnet pushed 1 new commit to master: http://git.io/DwcOaQ
02:45 GitHub149 inline-module-pm/master 4e97078 Ingy döt Net: Tame autostub a bit...
02:47 GitHub51 [inline-module-pm] ingydotnet pushed 1 new commit to master: http://git.io/4brvSA
02:47 GitHub51 inline-module-pm/master bbdb516 Ingy döt Net: CPAN Release 0.19...
02:47 GitHub45 [inline-module-pm] ingydotnet tagged 0.19 at master: http://git.io/gdE0Zg
17:23 SysPete joined #inline
19:33 bulk88 ingy what is the schdule for getting 26fork.t Winb32 hang fix out to CPAN?
19:40 GitHub125 [inline-module-pm] ingydotnet opened issue #13: Add a `perl-inline-module-autostub` command http://git.io/j7A75w
19:47 ingy bulk88: I think 27 still fails on linux
19:52 GitHub179 [inline-module-pm] ingydotnet comment on issue #12: @daoswald, never heard back from you. Closing this now. I think PRD needs to be
19:52 GitHub152 [inline-module-pm] ingydotnet closed issue #12: Inline::CPP needs to bundle PRD in inc http://git.io/wuP_Bg
20:06 GitHub191 [inline-module-pm] ingydotnet pushed 1 new commit to master: http://git.io/e3QByA
20:06 GitHub191 inline-module-pm/master 7b75889 Ingy döt Net: Make sure we don't autostub in distdir
20:48 ingy sivoais: are you about?
20:48 ingy I finally have time to look into DZ stuffs
20:57 sivoais ingy: yep, here!
20:58 ingy sivoais: great
21:00 jberger joined #inline
21:00 jberger o/
21:00 leont You want a subclass of Module::Build that overrides ACTION_code and ACTION_distdir to do your things
21:00 ingy hi jberger \o/
21:01 ingy leont++
21:01 jberger ingy: you might want to look at Alien::Base::ModuleBuild
21:01 ingy ok
21:01 ingy leont: seems very simple if that's the case
21:01 jberger it might have nothing to do with what you need, but it probably uses similar mechanisms to do it
21:02 ingy jberger: yeah it would be good just to look at a modern MB based dist/repo
21:02 leont MB has lots of failings, but it is still generally easier to extend than MM
21:03 jberger for as "modern" as MB allows you to be
21:04 jberger its "object model" is possibly its biggest failing
21:04 leont Well, it copied many of MM's mistakes, and not all of its strenghts
21:04 leont It's a god object, really
21:05 ingy I just need to make it work well enough for a toy (Inline/XS) module release (Alt-Acme-Math-XS-ModuleBuild)
21:05 ingy then I'm done, and it's just maintenance mode
21:10 leont 345 methods is at least 300 too many for my taste
21:10 jberger leont would take 45 men to defend sparta
21:13 leont How Lacedaemon to say that ;-)
21:44 GitHub29 [inline-module-pm] ingydotnet pushed 1 new commit to master: http://git.io/EoWt3A
21:44 GitHub29 inline-module-pm/master 26cc216 Ingy döt Net: Make handle_distdir not use @ARGV directly
22:29 ingy leont: do you have a moment to help me with some trivial moose junk?
22:30 ingy^1 joined #inline
22:31 ingy I'm just having trouble with accessors
22:33 ingy ether: ^^ having moose sadness in dz plugin
22:33 ingy just need a hand
22:34 ether I'm here but lots of things going on at once
22:34 ether do a brain dump with high verbosity and I'll comment?
22:35 ingy it would be easy to show in pairup, but if you are too busy, its ok
22:38 ether I'm much more available asynchronously at the moment unfortunately
22:38 ether do you have a stack trace or a code snippet or anything like that?
22:40 ingy it would probably take a few iterations
22:40 ingy I'll rtfm and then ask if I can't figure it out
22:40 ingy although I've already tried 20 reasonable things :(
22:42 sivoais O_o
22:43 ingy sivoais: you could help me :)
22:43 sivoais maybe... :-P I could certainly take a look
22:43 ingy what's you github id again?
22:43 sivoais zmughal
22:44 ingy^1 ssh -t zmughal@104.236.173.100 pairup
22:45 ingy it's the dzil stuff, so apropos
22:45 sivoais woah, coolio
22:45 ingy can you go wider?
22:45 sivoais yes
22:45 ingy great
22:46 ingy you can drive
22:46 ingy if you get stuck on what is up, just ask me here
22:47 sivoais >_< I don't know the tmux bindings. I use screen still :-P
22:47 ingy ctl-a →
22:47 sivoais cool
22:47 ingy vim
22:47 ingy I just guessed at those attributes
22:48 ingy I want 3 dzil keywords
22:48 ingy module,stub.ilsm
22:48 ingy and they can be used once or multi
22:48 ingy and at least one 'module' required
22:49 sivoais hmm, I see
22:49 ingy and I expect to be able to access them and get an array ref
22:50 ingy ilsm defaults to ['Inline::C']
22:51 ingy and stub should default to [map "$_::Inline", @{$_[0]->module}]
22:51 ingy something like that
22:51 sivoais so I see ilsm when I dump $self...
22:51 ingy ctl-a pageup
22:52 sivoais right?
22:52 ingy yeah
22:52 ingy I think it's the 'handles'
22:52 sivoais ah!
22:52 ingy I think you can access ilsms
22:52 ingy but I don't know that stuff
22:53 ingy leont wrote it I think
22:53 sivoais usually folks do all_$attr ? That's what I saw in the docs
22:53 ingy ok go for it
22:54 * leont uses 'handles' a lot, may well be possible
22:55 ingy let me drive a sec
22:55 sivoais ok
23:01 ingy sivoais: if I call another accessor in a default block what happens
23:01 ingy can I guarantee an order?
23:02 sivoais ingy: I've never done that. I usually would do something like a lazy build instead
23:02 ether you shouldn't depend on $self even having a value in a non-lazy default sub.
23:03 ether its state will be uncertain. if you're doing anything complicated at all, make it lazy
23:04 sivoais btw, _build_footer goes away with using the postamble, right? I have a PR that does that
23:04 leont lazy is one of Moose/Moo's weaknesses, but the solution I'd propose would probably be deemed an overkill :-p
23:04 leont (declaring which attributes depend on what other attributes)
23:05 sivoais hehe, I'd love that actually. Something like dataflow programming
23:05 jberger leont: I just came from a system like that, be careful what you wish for
23:05 jberger or at least implement it better than $old_job did
23:06 leont It's essentially IOC
23:06 leont Yeah, implementing it correctly is tricky
23:06 sivoais laziness, impatience, and *hubris*
23:06 sivoais ;-)
23:06 leont Bread::Board::Declare is the closest thing I'm aware of to what I'd like
23:07 jberger you end up needing a dependency graph
23:08 jberger and to implement a system which detects dirty values farther up the graph, you are going to be tempted to do some form of source re-writing
23:08 jberger once you do that, the game is already over, you just don't know it yet
23:09 ingy sivoais: yeah, let's look at your PR
23:09 sivoais here it is <https://github.com/ingydotnet/dist-zilla-plugin-inlinemodule-pm/pull/4>
23:09 ingy sivoais: welcome to git-hub
23:09 ingy fasten seatbelt
23:10 * sivoais hastily grabs seatbelt
23:13 * jberger says a quick prayer for sivoais
23:13 ingy sivoais: which pr?
23:13 sivoais weeee!
23:13 sivoais 4
23:17 ingy looks great
23:17 ingy brb
23:17 sivoais \o/
23:23 sivoais ingy: I think it was trying to merge the bit where I remove Inline::Module::MakeMaker in after_build
23:24 willthechill joined #inline
23:24 sivoais which I guess is no longer necessary :-)
23:25 ingy how do we do lazy build?
23:26 sivoais add lazy => 1, and builder keys to the attributes
23:26 ingy can you do that?
23:26 ingy ie you drive
23:26 sivoais perldoc Moose::Cookbook::Basics::BinaryTree_BuilderAndLazyBuild # for more info..
23:26 sivoais sure
23:27 sivoais which attributes need to be lazy?
23:27 ingy and then move the stub default code into the accessor 'default'
23:27 sivoais I see
23:27 ingy well stub defaults off of module
23:27 ingy so whatever that means :)
23:29 sivoais I think that should do it
23:29 ingy ok let me drive
23:29 sivoais sure
23:31 sivoais ingy: problem. if you do that if @{$self->stub}, it will make it
23:31 sivoais you can add a predicate to do $self->has_stub
23:31 ingy yeah saw it
23:31 ingy let's see if works
23:35 sivoais woo!
23:36 ingy we need to add 2 lines to start of Makefile.PL
23:36 ingy you know how?
23:36 sivoais I'd have to look at MakeMaker::Awesome
23:36 ingy or ask ether or leont  :)
23:37 leont the footer argument to MM::A is your friend
23:37 sivoais ah, you can set the header option... before _build_header => sub { ... } ?
23:37 leont header I meant, #braino
23:38 leont Well, it's not completely at the start of the Makefile.PL, but before most interesting things
23:38 sivoais or just override _build_header, since it doesn't do much by default
23:47 ingy brb
23:54 ingy sivoais: let's release this thing
23:56 sivoais ingy: there are 2 more PRs. One is a doc thing, the other is pretty serious O_O
23:58 ingy ok

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