Perl 6 - the future is here, just unevenly distributed

IRC log for #inline, 2014-11-15

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

All times shown according to UTC.

Time Nick Message
00:00 ether ok done
00:02 ether Makefile.PL bits look good?
00:02 ingy so next
00:02 ingy yes
00:02 ingy before the after_build
00:02 ingy want to do one thing
00:02 ether so, you want to call some arbitrary code after .pms are written to disk?
00:02 ingy not yet
00:04 ingy can you do that?
00:04 ether one plugin calling into another plugin is *weird*
00:04 ether as in really really weird
00:04 ingy ok but
00:04 ether I've done it once as a dirty filthy hack and had to connect up all the other dzil hooks to pull it off
00:04 ingy subclassing the main plugin from MM:Awesome is bad
00:04 ether no it isn't
00:04 ether well not really
00:05 ether we could just munge Makefile.PL directly, instead of using MMA ot do it
00:05 ether I've written a few plugins that do that (in order to allow the user to use Module::Build or Module::Build::tiny instead, if they liked)
00:06 ether that's pretty easy if it's ok to just append to Makefile.PL, rather than inserting stuff into hte middle
00:06 ingy I need to do both
00:06 ether MMA gets complicated because it needs to insert WriteMakefile() args, modify sharedir parameters etc
00:06 ether if you're just adding content to the end, easy peasy
00:07 ingy not just adding to the end
00:07 ingy anyway, this is working but seems wrong
00:07 ether is this still accurate(ish)?  https://gist.github.com/anonymous/f0193b87412dcafebffc
00:07 ingy but let's keep going
00:08 ingy I think it's time to do the after_build
00:08 ingy let me try
00:11 ether after_build gets no args other than $self
00:11 ingy not true :)
00:11 ether oh heyyeah :)
00:11 ingy and that's the info I need :)
00:11 ether it gets passed the build_root
00:11 ether a ha
00:11 ether it's just *so wrong* to be doing that though :)
00:12 * ether is twitching :)
00:12 * ether is a rules girl :)
00:12 ingy ?
00:12 ether after_build is look-but-don't-touch
00:12 ether but carry on
00:12 ether you rule breaker you
00:14 ingy :D
00:14 ingy well we can make more rules later
00:15 ingy ether: how do I get a file path string from that?
00:18 ether sorry.. I broke $work build.. back :)
00:19 * ether twitches more at Path::Class :D
00:19 ingy ask first
00:19 ether sorry
00:19 * ether is bad
00:20 ingy ok how do I get the stringification?
00:20 ingy you can drive
00:22 ether I can?
00:22 ingy hold on
00:25 ether brb 5 min; build really broken
00:25 ingy ok
00:28 ether its $build_root->stringify btw
00:28 ingy ok
00:28 ingy I broke something
00:31 GitHub63 [inline-c-pm] jkeenan opened issue #30: Why does Inline::C require an upgrade to ExtUtils::MakeMaker? http://git.io/wt3mow
00:32 ingy ether: what did I do?
00:34 ether what's the problem?
00:35 ingy that
00:37 ether back
00:37 ingy yay
00:37 ether I didn't see it?
00:37 ingy figgered it out
00:37 ingy my bad
00:38 ingy wrong branch
00:38 ether :)
00:38 ether ->modules
00:38 ether unless there should only be one?
00:38 ether it sounded like there could be 1 or more
00:38 ingy I want the 'module' attribnute
00:38 ingy from the dist.ini
00:39 ether yeah, the accessor is ->modules, because it returns a list
00:39 ether can there be more than one module?
00:39 ether or just only one, ever?
00:39 ether if just one, I can simplify some stuff
00:39 ether e.g. the map in _build_footer, and the attribute declaration
00:40 ether if you could have more than one module, you'll need a loop in after_build
00:40 ingy where is the accessor called 'modules'
00:40 ingy coming from?
00:40 ingy it just adds an 's'?
00:42 GitHub23 [inline-c-pm] karenetheridge comment on issue #30: >  My belief is that if a CPAN distribution requires a module distributed with the core, it should not force an upgrade of that distribution at least until a new stable version of Perl is released on an annual basis... https://github.com/ingydotnet/inline-c-pm/issues/30#issuecomment-63151378
00:42 ether it's coming from the 'handles'
00:42 ether the attribute itself is actually a listref
00:42 ether but the 'modules' accessor gives you a list, not a listref
00:43 ingy got it
00:43 ether yeah, right there :)
00:44 ingy ok here goes
00:45 ingy mungtastic
00:47 ether cool!
00:48 ingy not quite working
00:48 ether beep at me if you need me again.. putting out a fire
00:48 * ether hears your console beeps :)
00:48 ingy ether: like this?
00:48 ingy not sure how to beep
00:49 ingy ether: I'm not sure the build_dir is written at after_build time
00:52 ingy gah
00:52 ingy ok
00:54 ether it is: https://metacpan.org/source/RJBS/Dist-Zilla-5.025/lib/Dist/Zilla/Dist/Builder.pm#L340
00:54 ether files are written out, then after_builds are run
00:54 ether what are you seeing?
00:54 ingy it was my mistake again
00:55 ether if you just do a 'dzil build', it'll be put in Dist-Name-1.23, not .build/latest/
00:55 ether oh ok :)
00:55 ingy getting there
00:56 GitHub113 [inline-c-pm] jkeenan comment on issue #30: I'm not concerned at all with the existence or non-existence of Perl 5.22.  But why does Inline::C need to make use of new features in EU::MM?  Why can't it be written with minimal forced upgrades from the versions distributed with the latest released Perl of modules released as part of core? https://github.com/ingydotnet/inline-c-pm/issues/30#issuecomment-63152159
00:56 ingy ether: wanna add your name here?
01:02 ether where?
01:02 ingy there
01:02 ingy I did it for you
01:03 ingy ok well it'
01:03 ingy ok well it's totally working!
01:03 ingy I'm releasing this
01:03 ether cool!
01:03 ether you don't need Path::Class
01:03 ether we're trying to rip it out from dzil
01:04 GitHub18 [inline-c-pm] Leont comment on issue #30: >  But why does Inline::C need to make use of new features in EU::MM?... http://git.io/rcCTiQ
01:05 GitHub144 [inline-c-pm] karenetheridge comment on issue #30: > I'm not concerned at all with the existence or non-existence of Perl 5.22.... http://git.io/ilLBOQ
01:06 * ether should probably not try to argue with kid51
01:06 leont He has good intentions
01:06 ingy ether: invite him in here
01:06 leont Just keep shouting "why?" at him as if you're a 3 year old :-p
01:06 ingy I'm going to release these modules
01:06 ether he has a funny understanding of release cycles sometimes
01:10 ingy ether: wanna watch a ZillaDist release
01:10 ingy ?
01:10 ether sure!
01:10 ether how do I disconnect from pairup? I need to move it to another window
01:10 ingy leont in the house
01:10 ingy ctl-a d
01:10 leont I'm about to go do bed though
01:11 ether ah it's jsut screen
01:11 ether so ^A^A for me then :)
01:11 ingy tmux
01:11 ingy not screen
01:11 ether gah sorry
01:11 ingy stooop
01:12 ether I moved it from in-screen to out-of-screen, and then promptly banged on ^A to go back to my other window :D
01:13 leont Bad revision, bad boy ;-)
01:14 ingy needs one commit pushed
01:17 ingy leont: we still need that stuff?
01:17 leont You need a -remove = Readme in that filter
01:17 leont Otherwise it will generate two Readmes
01:18 ingy you drive
01:18 leont That was all
01:21 ingy deal with  the opendir later
01:21 * leont &
01:25 GitHub157 [acme-math-xs-pm] ingydotnet pushed 2 new commits to dzil: http://git.io/9LpwKg
01:25 GitHub157 acme-math-xs-pm/dzil d52b843 Ingy döt Net: Clean up dist.ini
01:25 GitHub157 acme-math-xs-pm/dzil 0bcc842 Ingy döt Net: Dzil needs a .gitignore
01:25 ingy ether: I think I need to munge the MANIFEST
01:25 GitHub26 [acme-math-xs-pm] ingydotnet tagged v0.0.1 at 5efae49: http://git.io/BDyQ7g
01:26 ingy^leont ok, we are going to have tagging conflicts
01:26 ingy^leont ether: can I prefix the release tags?
01:26 ingy^leont tearing down this pairup
01:32 ether [Git::Tag] has some format options yeah
01:32 ether or you could use a funky version perhaps?
01:33 ether thanks for the pairup! that was fun
01:39 ingy ether: thanks. this will make for a good report on the grant
01:44 ether :D
01:44 ether thanks for the author shout :)
01:44 ingy :)
09:53 chansen joined #inline
16:25 GitHub132 [inline-module-pm] daoswald opened issue #2: Inline/Module.pod link to Inline/Module/Tutorial.pod doesn't work. http://git.io/fWCyWg
16:30 GitHub71 [inline-module-pm] daoswald opened issue #3: search.cpan.org vs metacpan.org produce different search results. http://git.io/P1zkbQ
17:20 GitHub58 [inline-module-pm] daoswald opened issue #4: Inline::Module::Tutorial discussion of Makefile.PL: Code example is not complete enough. http://git.io/q9BjZw
17:25 davido_ joined #inline
17:25 davido_laptop Hey all.
17:46 davido_laptop ingy Are you able to get together at some point today or tomorrow?
22:59 ingy hi davido__
23:04 davido__ Hey ingy
23:05 davido__ How are you?
23:05 davido__ I encountered an issue in the Math::Prime::FastSieve conversion.  Shouldn't be too hard, but not quite sure of the best way to fix it.
23:08 davido__ If you were to fetch the repo: https://github.com:daoswald/Alt-Math-Prime-FastSieve.git and perform the eumm mantra you'll see.  The simplest failure is t/01-load.t
23:09 davido__ In fact, that's really *the* test to see.  I believe once that one is solved everything else will fall into place for that module.
23:15 ingy :)
23:17 davido__ :)
23:19 ingy davido__: fatal: repository 'https://github.com:daoswald/Alt-Math-Prime-FastSieve.git/' not found
23:19 davido__ grr, just a moment. :)
23:19 davido__ https://github.com/daoswald/Alt-Math-Prime-FastSieve.git
23:19 davido__ or git@github.com:daoswald/Alt-Math-Prime-FastSieve.git
23:20 davido__ :)
23:20 davido__ (never c/p uri's. ;)
23:20 davido__ i mean always c/p, never hand-type
23:22 davido__ If you want to pull it into the pairup I can drive and demonstrate
23:28 ingy davido__: I get https://gist.github.com/395f4670faf2dd418032
23:29 davido__ Oh, t/01can_subs
23:37 davido__ NAME => ... in FastSieve.pm is one problem.
23:56 ingy davido__: what is the basic problem as you understand it?
23:57 davido__ Well, two problems really.
23:58 davido__ First, because i'm using NAME => ... (which I should probably not be using), the .so that gets generated is named FastSieve.so, and I think we're trying to bootstrap FastSieve/Inline.so
23:58 ingy ok
23:58 davido__ I think that's it.
23:59 davido__ but this morning when I was working on it I got pulled away before confirming.
23:59 ingy ok
23:59 davido__ I also think that if we remove NAME => .... we're still looking in the wrong namespace for the functions.
23:59 davido__ (or generating them into the wrong namespace) but I'm not sure about that.
23:59 ingy I need to write a weekly report
23:59 davido__ Does that make sense?

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