Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2015-03-01

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

All times shown according to UTC.

Time Nick Message
00:00 jberger ?
00:00 mohawk sf.net ones
00:01 jberger yeh
00:04 mohawk anyway, temporary top tip - don't manually delete branches on gh as they'll get recreated until tomorrow (or so), just on sf
00:05 mohawk (and in doing so will trigger a travis build)
00:08 sivoais is there anyway of sending GitHub notifications to pdl-devel?
00:08 mohawk let's take a look
00:08 mohawk i can think of a workaround
00:09 sivoais yeah, subscribe a pdl-specific user
00:11 mohawk ie create a "pdl-devel" gh user, give that as its email address
00:12 mohawk sivoais, the coverage jobs take longest so they need initiating first
00:12 mohawk better parallelism :-)
00:13 sivoais actually,  let's not do that. I am following a project that uses a ML and GH heavily and the GH is very noisy
00:14 sivoais yeah,  you mentioned that before. now that it all works, I shall do that
00:14 mohawk also, i don't like non-parallel testing
00:14 mohawk it's taking over 2 minutes
00:14 mohawk can we use that formatter you had before?
00:15 mohawk also, is it possible to have parallel make? after all the blood i sweated in enabling that :-)
00:17 mohawk it might be necessary to bump the EUMM prereq to 7.5.14 though
00:18 sivoais yes, I would prefer parallel as well. We can always keep a non parallel build that is a "fast finish"
00:18 sivoais meaning that it doesn't have to be finished for a build to be marked as finished
00:20 mohawk i'm not sure what that would add
00:21 mohawk i can't imagine any situation where a parallel build works, and a non-parallel doesn't?
00:22 mohawk we got a fail, 5.20, badval=0
00:22 mohawk https://travis-ci.org/PDLPorters/pdl/jobs/52583333#L7472
00:23 mohawk https://travis-ci.org/PDLPorters/pdl/jobs/52583333#L6195
00:23 mohawk also t/transform.t seems to get wrong number of tests there
00:25 mohawk looks like line 214 of t/pdl_from_string.t should skip if !WITH_BADVAL
00:25 mohawk or, you know, something that doesn't break :-)
00:29 mohawk btw, i'm not convinced having "allow failure" on BADVAL=0 is the way forward
00:39 travis-ci joined #pdl
00:39 travis-ci PDLPorters/pdl#22 (travis-ci-more - 316dba2 : Zakariyya Mughal): The build passed.
00:39 travis-ci Change view : https://github.com/PDLPorters/pdl/compare/64d73fe8e02f^...316dba242057
00:39 travis-ci Build details : http://travis-ci.org/PDLPorters/pdl/builds/52583325
00:39 travis-ci left #pdl
00:43 mohawk w00t
00:48 mohawk ...except for the badval stuff :-)
00:51 sivoais mohawk: I'm going to address that in test-cleanup
00:51 mohawk ah, i was about to have a go myself
00:51 sivoais mohawk: the non-parallel build is so that we can diff the TAP
00:51 mohawk on t-ci-m
00:52 sivoais mohawk: go ahead!
00:52 mohawk when would we want to diff the TAP?
00:52 mohawk pass/fail, what else matters?
00:52 sivoais I think chm might...
00:53 sivoais having it run fast finish will let it finish as success even if that one build is still running
00:53 sivoais I think
00:53 mohawk since chm isn't actually meaningfully using travis and we are, let's have the actual consumers drive this one
00:53 mohawk what say you?
00:53 mohawk all the builds are taking x amount *times* longer due to non-parallel building and testing
00:54 sivoais sure, makes sense. YAGNI prevails again!
00:54 mohawk i'm not even convinced about having -v on the tests
00:55 sivoais I'd like to keep -v. It doesn't hurt.
00:55 sivoais Because TAP is made to be parsed :-)
00:55 mohawk if it's compatible with parallel, then fine
00:56 sivoais it is if you use that file formatter :-)
00:56 mohawk so, you take a look at the parallel stuff, i'll attack the badval tests?
00:56 mohawk make it so, mr sivoais :-)
00:56 sivoais sure, where are you working?
00:56 mohawk same thing, we'll just fetch & rebase as necessary - coord here
00:56 sivoais sure
01:01 mohawk the next version of EUMM after 7.06 is released is going to have better parallel build support for subdirs
01:02 mohawk which will speed things up for heavy subdir users like PDL ;-)
01:02 sivoais make -j$N should be OK, here, right?
01:02 mohawk i believe so
01:02 mohawk it's linux, it should be GNU make
01:02 mohawk what will you set $N to?
01:03 mohawk houston, we have badval test failure repro here
01:05 sivoais $SYSTEM_CORES
01:05 mohawk make that be that, x 2
01:05 mohawk or hell, say x 3
01:05 mohawk i use x 2
01:10 mohawk that's fixed 1 test
01:10 sivoais ok, I think I have the parallel working and moved the coverage to the beginning
01:10 sivoais where are the BADVAL=0 fixes?
01:10 sivoais I'd like to rebase on that
01:11 mohawk on my hd
01:11 mohawk you go ahead, you need to push
01:11 mohawk i'm not done yet
01:16 mohawk wow, t/transform.t is failing just because Test.pm doesn't allow skipping >1 test
01:16 mohawk and the author never checked
01:16 sivoais there's lots of that kind of thing :-/
01:16 mohawk that's from craig
01:16 mohawk tsk
01:16 sivoais are you change it from Test to Test::More?
01:16 sivoais *changing
01:17 mohawk it's either that or skip for (1..3)
01:17 sivoais yeeeaaahh... like I said
01:17 mohawk yeah, it's T::M
01:18 sivoais I opened the test files and you-make-me-sad-monty-python.gif <http://imgur.com/gallery/k7Oidbu>
01:19 mohawk ha ha
01:19 mohawk there's "use warnings" in this thing, and no "use strict"
01:19 mohawk so i'm fixing that first, THEN t::m
01:20 mohawk i need to listen to some veruca salt
01:20 mohawk actually, some temper trap will hit the spot first :-)
01:28 opkick [pdl] wchristian pushed 2 new commits to travis-ci-more: http://git.io/xnhW
01:28 opkick pdl/travis-ci-more 80f820b Zakariyya Mughal: Travis-CI: move coverage build to start...
01:28 opkick pdl/travis-ci-more 9f898ff Zakariyya Mughal: Travis-CI: parallel builds and tests...
01:34 mohawk i've got the tests looking good
01:34 mohawk now just checking i haven't broken them against BADVAL=1 :-)
01:35 sivoais mohawk: let me bring them to Travis-CI
01:35 sivoais that's what it is for :-)
01:35 mohawk with your crappy non-parallel build/test?
01:35 mohawk who's got that kind of time? :-)
01:36 mohawk hashtag banter
01:36 sivoais it's parallelised now! ^^^
01:36 mohawk yes, i did see that :-)
01:36 mohawk nearly ready
01:37 sivoais shoot. I had a tab
01:37 mohawk not the right context for boasting about drug use, my friend
01:37 mohawk ;-)
01:38 sivoais :-P
01:38 sivoais well, caffeine *is* a drug...
01:39 mohawk ok, cancel that build
01:40 mohawk i'm about to rebase etc
01:40 sivoais no build yet since the GH hasn't been synced yet
01:40 mohawk tsk
01:40 * sivoais made a force push to fix a config file syntax issue
01:41 mohawk i'll push to gh too
01:41 mohawk done
01:41 opkick [pdl] mohawk2 pushed 3 new commits to travis-ci-more: http://git.io/xnjF
01:41 opkick pdl/travis-ci-more 3aa817c Ed J: Fix if WITH_BADVAL=0: t/pdl_from_string.t
01:41 opkick pdl/travis-ci-more f5d64d7 Ed J: Add "use strict" to t/transform.t
01:41 opkick pdl/travis-ci-more 02dd643 Ed J: Rewrite t/transform.t with Test::More so "skip" works right
01:42 mohawk https://travis-ci.org/PDLPorters/pdl
01:42 mohawk you done messed up, son
01:42 sivoais I see that
01:43 mohawk grin
01:43 mohawk anyway, you have the conch shell
01:43 mohawk also, the mirror DID happen, about 15 mins ago - see above
01:44 mohawk [01:28] <opkick> pdl/travis-ci-more 9f898ff Zakariyya Mughal: Travis-CI: parallel builds and tests...
01:44 mohawk e la vita
01:46 mohawk sivoais, while you're fixing that, please move the other COVERAGE ones to top too
01:46 sivoais okie
01:46 mohawk AND move the BADVAL ones out of allow_failure
01:46 mohawk sivoais++
01:57 mohawk sivoais, if you're otherwise engaged, want me to tidy it up? you've done all the hard work
01:57 sivoais almost done
01:57 mohawk fantastic
02:01 opkick [pdl] zmughal force-pushed travis-ci-more from 02dd643 to ea60839: http://git.io/xcfO
02:01 opkick pdl/travis-ci-more 14f1f9a Ed J: Fix if WITH_BADVAL=0: t/pdl_from_string.t
02:01 opkick pdl/travis-ci-more db8fd1c Ed J: Add "use strict" to t/transform.t
02:01 opkick pdl/travis-ci-more 9735378 Ed J: Rewrite t/transform.t with Test::More so "skip" works right
02:01 mohawk aww yeah
02:03 mohawk that is a tight looking batch of 14 commits: https://github.com/PDLPorters/pdl/compare/master...travis-ci-more
02:08 mohawk and the builds are taking 6 mins instead of 8
02:08 mohawk sivoais++ # master of parallelism and CI
02:10 sivoais looks like the brings the build down by 1 minute each
02:11 sivoais compare #22 <https://travis-ci.org/PDLPorters/pdl/builds/52583325> with #25 <https://travis-ci.org/PDLPorters/pdl/builds/52589815>
02:11 mohawk sivoais, what's the value of $(test-jobs)?
02:11 mohawk i make it 2 minutes - 8 vs 6, aka 25%
02:11 sivoais it is SYSTEM_CORES + 1 when there isn't coverage and 1 if COVERAGE=1
02:12 mohawk why + 1
02:12 mohawk ?
02:12 sivoais that's what the travis-perl-helpers do :-)
02:12 mohawk i think x 2 is the normal recommendation
02:12 mohawk hmm, ok
02:13 sivoais I was thinking that I could drop one the first 5.20 test since it is just a dupe of one of the BADVAL tests
02:14 mohawk looks like nearly all our 2 mins came off the parallel testing
02:14 mohawk suits me
02:14 mohawk the make was taking 300s, now about 260
02:14 mohawk better than a sharp stick in the eye
02:15 sivoais I'll comment it out of the build matrix and add a note why I did
02:15 mohawk sivoais, if you're feeling adventurous, could you also add a job with EUMM 7.5.13?
02:15 mohawk i'd like to see if it's much faster building
02:15 sivoais sure
02:15 sivoais I'll do it on another branch
02:16 mohawk so long as it's a branch of this finished one, great ;-)
02:16 travis-ci joined #pdl
02:16 travis-ci zmughal/p5-PDL#91 (travis-ci-more - 346ea22 : Zakariyya Mughal): The build was broken.
02:16 travis-ci Change view : https://github.com/zmughal/p5-PDL/compare/316dba242057...346ea22d162f
02:16 travis-ci Build details : http://travis-ci.org/zmughal/p5-PDL/builds/52588825
02:16 travis-ci left #pdl
02:17 sivoais O-:
02:17 mohawk ha ha, oops
02:17 mohawk i humbly suggest that repo has served its purpose ;-)
02:18 sivoais I'm using it for issues too :-)
02:18 * mohawk twitches
02:19 mohawk shouldn't those be on the pdlp repo?
02:20 sivoais maybe ... :-P
02:22 sivoais right now they are serving as a checklist
02:22 sivoais for my purposes
02:28 sivoais I'll make the move
02:29 travis-ci joined #pdl
02:29 travis-ci PDLPorters/pdl#25 (travis-ci-more - ea60839 : Zakariyya Mughal): The build passed.
02:29 travis-ci Change view : https://github.com/PDLPorters/pdl/compare/02dd6439782c...ea60839b23d7
02:29 travis-ci Build details : http://travis-ci.org/PDLPorters/pdl/builds/52589815
02:29 travis-ci left #pdl
02:30 sivoais \o/
02:30 mohawk now that's a big fucking deal
02:31 sivoais #22 : ran for 2 hrs 29 min 33 sec; #25 : ran for 2 hrs 17 sec
02:35 jberger sivoais++
02:35 jberger man if that doesn't convince chm to go to github ....
02:35 mohawk bear in mind it's lying, that's across all the jobs
02:36 mohawk but the one above did take 37 mins to complete
02:38 mohawk further parallelism improvements:
02:38 mohawk looks like the BADVAL changing ones take longer, we should move them to first too
02:38 mohawk actually, all the 5.20 ones are slower
02:39 mohawk i wonder why
02:40 sivoais that's a good question. Further goal is to use NYTProf on all builds
02:40 mohawk i think that's a bad idea, because it slows things down a lot
02:41 mohawk CI is to make sure things work, i reckon
02:41 mohawk not what's 5% faster on all versions
02:41 sivoais oh, not on the CI, just locally :-)
02:42 sivoais I think that is in one of the TODOs in the repo
02:46 sivoais something like, "make debugging easier"
02:46 mohawk probably more "make" targets are the way forward
02:46 sivoais so purty... <https://github.com/PDLPorters/pdl/issues>
02:46 mohawk i'll comment accordingly
02:46 sivoais mohawk: it was transform.t, right <https://github.com/PDLPorters/pdl/issues/9>
02:48 mohawk correct, already checked :-)
02:48 mohawk i don't see the "make debugging easier" issue?
02:48 sivoais mohawk: that was in the TODO file. You're welcome to make an issue on the GH
02:49 mohawk let's clear the SF and gh bug lists first
02:50 mohawk bear in mind the TODO file is 524 lines
02:50 sivoais yeah
02:50 mohawk that's been a write-only doc for too long
02:50 sivoais and there's <https://github.com/PDLPorters/pdl/wiki/Roadmap>, <https://github.com/PDLPorters/pdl/wiki/PDL-TODO-List>, <https://github.com/PDLPorters/pdl/wiki/GSoC-2012-Ideas> too
02:51 sivoais I've been hangout out with the #openhatch people and seeing how they make onboarding new contributors easier
02:51 sivoais *hanging out
02:51 mohawk nice
02:52 mohawk well, if you have skills in that, i defer to you
02:52 sivoais and I see that making it clear what needs to be done in the bug tracker and which things are "low-hanging fruit" *really* helps gets people to come into the project
02:52 sivoais #ipython does that
02:52 mohawk you're right
02:53 sivoais That was actually what was holding *me* back when I joined this channel years ago
02:53 mohawk actually, the size of TODO is misleading, and i see you've already got lots of it in as issues
02:53 sivoais I had no idea what needed to be done and there just wasn't enough discussion on the mailing list to know where the project was going
02:54 sivoais more discussion was happening here ;-)
02:54 sivoais whe run4flat was around :-)
02:54 mohawk i've chatted with him here
02:55 mohawk i miss hiim
02:55 mohawk i want us to make one Alien::* module using A::B
02:55 mohawk you know, i might just hack up one using c+p out of the PDL code
02:56 mohawk then jberger can convert it to using A::B :-)
02:56 mohawk btw, i do actually want your thoughts on https://github.com/PDLPorters/pdl/issues/12
02:57 mohawk it's not just rhetorical :-)
02:57 sivoais I just commented
02:58 mohawk seen - fair enough
02:59 mohawk i think doing this in a test is fine: like $@, /Distinctive stuff/;
02:59 mohawk but, different strokes
03:00 sivoais but what happens if that exception can happen in multiple places
03:01 mohawk sounds fine to me
03:01 sivoais which is likely if we are dealing with mathematical functions
03:01 sivoais die "division by zero!";
03:01 sivoais die "divided by zero!"
03:02 mohawk i don't think exception classes are the solution
03:02 mohawk make that be a PDL::Config thing, maybe
03:02 sivoais they can still be used as strings if we have them stringify to the same value as before
03:02 mohawk that allows a distant future with i18n
03:02 mohawk for me, this is YAGNI
03:03 sivoais yeah, that's why I put it as an enhancement, not a bug :-)
03:03 mohawk YAGNI,E
03:03 mohawk ever ;-)
03:03 sivoais hahah
03:03 mohawk |                    INSTALLATION                       |
03:03 mohawk Switch from ExtUtils::MakeMaker to Module::Build
03:03 mohawk GOD NO
03:04 mohawk that's literally the official PTG advice, btw
03:04 sivoais ok, I made PDLPorters/Test-PDL even with ebaudrez's version
03:04 mohawk well, the sentiment, maybe a bit less colourful ;-)
03:04 mohawk nice work
03:04 mohawk did you fork his repo into that? seems most idiomatic
03:04 sivoais nope. He forked PDLPorters
03:04 sivoais :-)
03:04 mohawk ha ha
03:05 mohawk did you fetch/push, then?
03:05 sivoais yes
03:05 sivoais and all the tags too
03:05 mohawk you da man
03:05 mohawk suggest comms with him, then
03:06 mohawk Check/set perldl.conf options before the build
03:06 mohawk * Add WITH_CORE_ONLY to perldl.conf
03:06 mohawk a point is missed
03:07 mohawk icecubes are useful
03:07 mohawk icebergs are a danger to shipping
03:08 sivoais mohawk: ebaudrez already said he's ok with getting it into PDL
03:08 mohawk my "vision" is we make more, smaller things
03:08 mohawk under the pdlp umbrella
03:08 mohawk almost a sort of "module-related" programming style
03:08 mohawk i think i just invented a great new paradigm!
03:09 mohawk though i think the name could be catchier
03:09 sivoais which reminds me, have you seen <http://ropensci.org/>
03:09 mohawk i have not
03:09 sivoais all their packages are under one GitHub org <https://github.com/ropensci>
03:10 sivoais jberger: ^ ROpenSci would be a good model for Perl + Science
03:10 mohawk that's a lot of repos
03:10 sivoais yeah, and a lot of contributors
03:11 mohawk i want us to invade that
03:11 sivoais O_o
03:11 mohawk questions?
03:11 sivoais invade...
03:11 mohawk ha ha
03:11 mohawk as in, create mutual awareness
03:12 sivoais aye, good plan
03:12 mohawk a repo there with perl2r stuff, for instance
03:12 sivoais I've been doing that with IPython
03:12 mohawk i think i heard someone had made some sort of "face-inter" between perl and R
03:12 sivoais hehe, that too
03:12 mohawk again, that probably needs a catchier name
03:13 mohawk what if you made a repo there with some of your neat Data::Frame stuff?
03:13 mohawk while i quickly hack up an Alien::* module
03:13 mohawk i'll race you
03:13 sivoais jberger had made <http://perl4science.github.io/> which hasn't gotten enough love lately
03:14 sivoais and I have plans to make a wiki under <https://github.com/eclectic-extractions>
03:15 jberger :(
03:15 jberger actually, that is on my list
03:15 jberger preaction's static site generator is ready for beta
03:15 mohawk hell yeah
03:15 sivoais statocles?
03:16 jberger yeah
03:16 sivoais sweet
03:16 jberger part of the problem with octopress is ... octopress
03:17 jberger doesn't encourage frequent editing
03:17 sivoais do I remember correctly that it runs slowly?
03:17 mohawk any preferences on which one i Alien-ise?
03:17 mohawk i'm thinking about HDF
03:18 sivoais libproj! # actually, I have no preference
03:18 sivoais HDF would be really good
03:18 mohawk whatever i can install on my centos to test against
03:19 mohawk so if you can tell me a package name for it, i'll do it :-)
03:19 sivoais I have the Debian package names in the .travis.yml file
03:20 jberger octopress? isn't especially slow, it is just sucks: old, not maintained, pinned to an old ruby, oh and ruby
03:20 sivoais oh, that old ruby was a pain. Esp. if I have to install it on every machine
03:21 sivoais and then the Gems that it depends on don't install cleanly all the time
03:22 mohawk ah hell, i'll just build it myself
03:25 mohawk building...
03:39 opkick [pdl] zmughal pushed 1 new commit to travis-ci-more: http://git.io/xcGa
03:39 opkick pdl/travis-ci-more aea6254 Zakariyya Mughal: Travis-CI: comment out unnecessary build for Perl v5.20...
03:40 mohawk nice
03:41 mohawk i'm on centos, so debian won't help me, btw ;-)
03:41 mohawk just running HDF's tests
03:44 sivoais added a PR <https://github.com/PDLPorters/pdl/pull/18> and linked it up with the mailing list too
03:50 mohawk damn, nice work!
03:56 mohawk ok, why the shit isn't this HDF stuff being found by PDL?
04:00 mohawk oh dear, does PDL::IO::HDF only do hdf4?
04:00 sivoais mohawk: yes
04:00 mohawk so all that installing was for nothing?
04:01 sivoais heh
04:01 mohawk is it not back-compatible?
04:01 sivoais if you looked at the .travis.yml, you would have seen "libhdf4-alt-dev"
04:01 sivoais ;-)
04:01 mohawk guess i'd better stab someone in the eye
04:01 mohawk probably myself
04:02 sivoais That's something that needs to be dealt with for future work
04:02 sivoais we need milestones
04:02 mohawk first one: mohawk's eye needs stabbing
04:02 mohawk yeah, we still have no idea what actually stands between us and 2.008
04:02 sivoais oh, can we get the GH issues/comments to show up in here too?
04:03 mohawk i know it's possible, from #inline
04:03 mohawk i know there's a bunch of bugs that chm would like to see fixed
04:03 mohawk but some are already fixed, and nothing changes on the bug list
04:03 opkick [pdl] zmughal pushed 1 new commit to travis-ci-more: http://git.io/xcGa
04:03 opkick pdl/travis-ci-more aea6254 Zakariyya Mughal: Travis-CI: comment out unnecessary build for Perl v5.20...
04:03 mohawk and i think some of the pri 5 ones are a big unnecessary
04:03 sivoais that was a test via the settings
04:05 sivoais I have to use the API to enable those
04:05 sivoais looking at <http://akrabat.com/changing-the-github-irc-hooks-notification-events/>
04:06 sivoais odd that there isn't a web interface for that
04:07 mohawk yeah
04:09 mohawk anyway, the alien package will be called Alien::HDF4
04:11 opkick [pdl] zmughal opened issue #19: Update the webhooks for IRC notifications http://git.io/xcC0
04:11 opkick [pdl] zmughal closed issue #19: Update the webhooks for IRC notifications http://git.io/xcC0
04:11 sivoais \o/
04:11 mohawk so i just got the HDF thing found by PDL, which is good
04:12 mohawk sivoais++
04:12 mohawk guess you can close that
04:12 mohawk please add a note with what you did :-)
04:12 sivoais all done already
04:12 mohawk dang
04:20 sivoais mohawk++ <https://github.com/PDLPorters/Alien-HDF4>
04:23 mohawk grin
04:23 mohawk give me time
04:24 mohawk this thing's going to take at least a couple more minutes
04:24 mohawk sivoais, could you set the repo up to notify to here? and to use travis, though i won't put in a .t.y just yet
04:24 mohawk when i do, i'll know what to copy
04:25 mohawk (just stepping away to start some food cooking)
04:26 opkick [pdl] coveralls comment on issue #18:
04:26 opkick [pdl] coveralls comment on issue #18:
04:33 travis-ci joined #pdl
04:33 travis-ci PDLPorters/pdl#27 (travis-ci-more - aea6254 : Zakariyya Mughal): The build passed.
04:33 travis-ci Change view : https://github.com/PDLPorters/pdl/compare/ea60839b23d7...aea62544f94c
04:33 travis-ci Build details : http://travis-ci.org/PDLPorters/pdl/builds/52594218
04:33 travis-ci left #pdl
04:38 mohawk sivoais++
04:41 mohawk that coveralls message was a bit, uh, compact
04:52 mohawk sivoais, nice work on the HDF4 issue
04:58 mohawk sivoais, you're a machine!
05:00 sivoais GET DOWN!
05:01 mohawk and boogie?
05:02 sivoais or should I say... I'll be back
05:05 mohawk ha ha
05:26 sivoais joined #pdl
06:28 travis-ci joined #pdl
06:28 travis-ci zmughal/p5-PDL#94 (test-cleanup-with-travis-ci-more - a471026 : Zakariyya Mughal): The build failed.
06:28 travis-ci Change view : https://github.com/zmughal/p5-PDL/compare/14f1f9a7c401^...a47102637171
06:28 travis-ci Build details : http://travis-ci.org/zmughal/p5-PDL/builds/52599642
06:28 travis-ci left #pdl
06:35 mohawk boooo
06:38 mohawk but that's a huge number of tests changed!
06:51 sivoais yeah
06:52 sivoais I'm just doing things in between letting my research code run :-P
06:52 sivoais and reading papers
06:53 mohawk grin
06:53 mohawk i'm still wrestling again with the madness of the HDF M.PL
06:53 mohawk got it looking much saner now
06:54 mohawk going to use the same pattern as Alien::Gimp
06:54 mohawk it'll be using EU::Depends
06:54 mohawk as in, A::HDF4 will produce
06:54 mohawk and the branch of PDL will consume
06:54 sivoais and Alien::Base?
06:54 mohawk nope
06:55 sivoais oh, is this just the configure param parts or are you working on the build/sharedir too
06:55 mohawk uhh
06:55 mohawk it's not going to get/build
06:55 mohawk it's just the same as there is now, but encapsulated
06:56 mohawk like i said, if jberger wants to convert it, he'll be able to
06:56 sivoais ah, that's a step forward nonetheless
06:56 mohawk and can use the logic i've extracted for what PDL/others will need
06:56 mohawk that was my thought
07:23 mohawk spotted a minor booboo in Alien::Gimp, now sorted
07:42 mohawk and now, a quick test of the supposedly working PDL with Alien::HDF4...
08:12 mohawk only just noticed - the HDF M.PLs use hardcoded .o instead of $(OBJ_EXT)
08:12 mohawk nice work
08:17 opkick [pdl] mohawk2 opened issue #20: IO/HDF/*/Makefile.PL use ".o" not $(OBJ_EXT) http://git.io/xcDU
09:48 opkick [Alien-HDF4] mohawk2 created master (+1 new commit): http://git.io/xCeC
09:48 opkick Alien-HDF4/master d040cdc Ed J: Release v0.01
09:48 opkick [pdl] mohawk2 created alienhdf4 (+1 new commit): http://git.io/xCe7
09:48 opkick pdl/alienhdf4 03a3bc3 Ed J: Use Alien::HDF4 in PDL::IO::HDF build
09:49 mohawk all done
09:51 opkick [pdl] mohawk2 opened pull request #21: Use Alien::HDF4 in PDL::IO::HDF build (master...alienhdf4) http://git.io/xCv0
09:54 mohawk and emailed to pdl-d
09:54 mohawk let us see what happens
10:59 opkick [pdl] coveralls comment on issue #21:
13:25 opkick [pdl] wchristian comment on issue #6: Amended:... http://git.io/xChz
13:28 opkick [pdl] wchristian deleted prereqdump at 56093ef: http://git.io/xCjt
13:28 opkick [pdl] wchristian deleted cpan_2.007_release at c58b3cf: http://git.io/xnBV
16:22 sivoais \o/
16:27 Mithaldu :)
16:27 Mithaldu hope it doesn't break anything
16:28 sivoais I'm trying to bring more people here from #perl
16:28 sivoais :-P
16:28 Mithaldu good :D
16:29 pdurbin sivoais: why? (just curious)
16:31 sivoais pdurbin: because if we can get more people new to PDL to come in, they can help find problems in the documentation
16:32 sivoais and *hopefully*, we can help them contribute the fixes themselves!
16:39 pdurbin good stuff
18:00 mohawk sivoais++
18:00 mohawk Mithaldu++
18:01 mohawk and now we have the github interface available, they'll actually enjoy the contributing process
18:03 drrho joined #pdl
21:09 sivoais I'd like to get Alien-HDF4 to be an actual Alien module (i.e., install libhdf4). No tuits right now, but it should not be hard with Alien::Base.
21:14 sivoais mohawk: if you'd like to take a go at that, I recommend taking a look at other Alien::Base modules. This is one that would probably be the easiest to replicate <https://github.com/rsimoes/alien-libyaml>
21:15 * sivoais & # off to do some reading
21:37 mohawk sivoais, enjoy the reading
21:59 drrho joined #pdl

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