Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2015-03-16

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

All times shown according to UTC.

Time Nick Message
00:00 mohawk they are in fact highly related
00:03 opkick [pdl] zmughal force-pushed changesfix from f4198bd to d390896: http://git.io/pxve
00:03 opkick pdl/changesfix 29c0cad Zakariyya Mughal: rename Release_Notes to Changes...
00:03 opkick pdl/changesfix e5f1b58 Zakariyya Mughal: remove install notes and warning about PDL C-API changes
00:03 opkick pdl/changesfix a38e6c4 Zakariyya Mughal: remove PDL v2.007_13 Changes note (not released yet)
00:06 opkick [pdl] mohawk2 comment on pull request #78 d390896: I think if this were `diag` instead of `note` it would show up in the Travis log? http://git.io/hfIh
00:07 mohawk this "checkmanifest.t" thing is already proving its worth
00:08 mohawk although ether's Test::DistManifest is even fancier - it also checks everything is either in MANIFEST or .SKIP, and no overlap
00:09 mohawk it appears that PDL::Transform::Proj4 is done with a .pd, but actually appears to be literally just a .pm file done as a .pd
00:09 mohawk unless i'm very confused
00:23 sivoais mohawk: it autogenerates some of the code
00:23 sivoais see INSERT_FULL_NAME_HERE
00:26 mohawk yeah
00:28 chm joined #pdl
00:29 chm all:  Please reply to the pdl-devel post when we are ready for the final, feature freeze, CPAN developers release.
00:29 mohawk ok
00:29 chm The following release will be RC1.
00:29 chm Good night.
00:30 opkick [pdl] zmughal comment on pull request #78 d390896: It should still show up since I'm using `-v`, but turns out, it isn't running at all?... http://git.io/hfm6
00:30 mohawk chm, night!
00:31 opkick [pdl] mohawk2 comment on pull request #78 d390896: I'll wager test-files might be confusing us. To the docs, Robin! http://git.io/hfmN
00:32 opkick [pdl] mohawk2 comment on pull request #78 d390896: Need AUTHOR_TESTING set to true. http://git.io/hfYq
00:33 mohawk sivoais, where do you feel the test-cleanup branch is?
00:34 opkick [pdl] zmughal comment on pull request #78 d390896: Should default to true, but I'll make it explicit. http://git.io/hfYR
00:34 mohawk sivoais, https://github.com/travis-perl/helpers#control-environment-variables
00:35 mohawk my, it does default to true
00:35 sivoais there's an issue that lists each of the files with a tick
00:35 sivoais for test-cleanup
00:36 mohawk https://github.com/travis-perl/helpers/blob/master/init#L103
00:36 mohawk sivoais, but if we were to merge t-c as is, what effect would that have? are there any parts ready for prime-time?
00:37 sivoais it's all ready
00:37 sivoais that's why I needed Travis in the first place :-P
00:38 mohawk grin
00:38 mohawk well, hold on till i've finished updating t/proj_transform at least :-)
00:38 mohawk let's just have a quick look at increasing coverage
00:39 opkick [pdl] zmughal force-pushed changesfix from d390896 to e75f22d: http://git.io/pxve
00:39 opkick pdl/changesfix e75f22d Zakariyya Mughal: Travis-CI: install CPAN::Changes so that author test runs
00:39 sivoais I'm going to grab dinner and try to finish off test-cleanup
00:39 mohawk great
00:39 mohawk be aware i'm going to push to it
00:40 mohawk bon appetit :-)
00:56 sivoais mohawk: i just realised why it won't run...
00:56 sivoais it's in MANIFEST.SKIP
00:56 sivoais and build-dist is after all that
00:57 mohawk ah, indeed
00:57 mohawk so just explicitly run it yourself?
00:57 sivoais sure
00:57 mohawk (in .t.y)
00:58 * mohawk races on t-c
01:01 sivoais maybe running author tests wouldn't be so bad, but we can discuss that on the list when we have more stuff under xt/
01:01 mohawk uh, not following?
01:01 mohawk this is on a branch in any case
01:02 sivoais oh, I was meaning, take ^xt/ out of MANIFEST.SKIP
01:02 mohawk i don't think it belongs in MANIFEST
01:03 mohawk and we're not doing (nor should we) "make manifest" - am i wrong?
01:04 sivoais I see that some modules like Moo do distribute their xt/ <https://metacpan.org/source/HAARG/Moo-2.000000/xt>
01:05 mohawk hmm
01:05 opkick [pdl] coveralls comment on issue #78:
01:10 sivoais it's just something to think about in the future... for now YAGNI!
01:10 mohawk grin
01:13 mohawk i think i'm going to get a lot more coverage
01:16 opkick [pdl] zmughal force-pushed changesfix from e75f22d to f169ff1: http://git.io/pxve
01:16 opkick pdl/changesfix f169ff1 Zakariyya Mughal: Travis-CI: install CPAN::Changes so that author test runs
01:24 mohawk let the building begin
01:24 opkick [pdl] mohawk2 pushed 3 new commits to test-cleanup: http://git.io/hfC4
01:24 opkick pdl/test-cleanup 091a547 Ed J: Tidy t/proj_transform2.t
01:24 opkick pdl/test-cleanup e2cea92 Ed J: Tidy t/proj_transform.t
01:24 opkick pdl/test-cleanup dfd7d29 Ed J: Exercise the aliases in PDL::Transform::Proj4
01:25 mohawk this ought to lead to a big jump in coverage
01:34 mohawk doh, i want to run a build of t-c but some swine has got in front of me
01:37 mohawk sivoais, i'm getting a number of test fails on test-cleanup
01:37 mohawk investigating
01:37 mohawk (on win32)
01:41 mohawk sivoais, you didn't put t/lib/TestHelper.pm in MANIFEST
01:43 sivoais this is what happens when I don't get Dist::Zilla ;-)
01:43 opkick [pdl] coveralls comment on issue #78:
01:43 * mohawk controls a sense of humour failure ;-)
01:43 mohawk srsly, looks like we need to shift to Test::DistManifest
01:44 mohawk i thought you said test-cleanup was ready for merge
01:45 sivoais it was the last time I checked! I don't know if rebasing changed that >_> :-P
01:45 mohawk the continued existence of  t/lib/TestHelper.pm is a crime against all that is sacred
01:45 mohawk ...as opposed to being ready for merge ;-)
01:46 mohawk i'm going to smash through and undo it
01:46 mohawk and Test::More-ify all the victims
01:51 mohawk it seems as though commits with references to t/l/TH were made, before the file itself was added
01:51 mohawk so some commits are by themselves under-functional?
02:06 mohawk looks like my point above is entirely wrong, sorry about that
02:07 mohawk in my defence, i am an idiot
02:40 mohawk wow, there's a lot of test files to change over
03:00 sivoais mohawk: hahah, yes, the 40 that I changed
03:00 sivoais I just prepped them to get Test::More'd
03:00 sivoais but there are some other cleanup things that need to be done like using lives_ok
03:01 sivoais instead of eval {}; pass if $@;
03:01 sivoais because that doen't give a good indication of what was tested :-P
03:03 sivoais mohawk: are you still working on it? I'm ready to do some work on test-cleanup and don't want to dupe effort
03:20 mohawk i used 'is $@, ""'
03:20 mohawk i am still working on it
03:21 mohawk i decided to purify history by crunching through each commit and annihilating all TestHelper
03:24 preaction joined #pdl
03:24 sivoais joined #pdl
03:24 jberger joined #pdl
03:24 ingy joined #pdl
03:24 mohawk welcome, warrior(s)
03:26 sivoais mohawk: why did ya cancel that test-cleanup job?
03:26 mohawk because it was going to make lots of fail
03:27 sivoais really, I see passing builds
03:27 mohawk and i thought it would be a super quick job to fix up :-)
03:27 mohawk then i saw the state of play with TestHelper, and didn't restart
03:27 sivoais 133.{2,8,9,10}
03:28 sivoais those builds passed
03:28 sivoais <https://travis-ci.org/PDLPorters/pdl/builds/54510756>
03:28 drrho joined #pdl
03:28 shadowpaste joined #pdl
03:28 pdurbin joined #pdl
03:28 robinsmidsrod joined #pdl
03:28 estrabd joined #pdl
03:28 vicash joined #pdl
03:28 Bender1 joined #pdl
03:28 lungching joined #pdl
03:28 Bender2 joined #pdl
03:28 osfameron joined #pdl
03:29 sivoais mohawk: you'll need to op the Bender's ;-)
03:31 sivoais I'm continuing on test-cleanup now
03:32 mohawk uh, not sure that's a great idea
03:33 mohawk i'm still working on it
03:33 mohawk is changesfix completely done?
03:33 sivoais mohawk: which parts?
03:33 sivoais yes
03:33 mohawk i'm going through all the commits
03:33 sivoais I checked the Travis-CI build and the xt/ is running just fine
03:33 mohawk not far to go
03:33 mohawk awesome
03:33 sivoais oh, what are you doing?
03:33 mohawk how's the thesis going? :-)
03:34 mohawk i did git rebase -i and did "edit" on all the commits, just to check through
03:34 sivoais haha, shhhhh...
03:34 mohawk now fighting through with the Test::More stuff
03:34 mohawk can you guide me on what lives_ok does?
03:34 sivoais lives_ok takes a coderef and just makes sure it passes
03:35 sivoais it makes it much clearer what the test is for than just running eval
03:35 sivoais I think some of the evals were originally *string evals*
03:36 sivoais yeah, like in t/argtest.t
03:39 mohawk yes, they were
03:40 mohawk i see you have both a plan, AND done_testing?
03:40 mohawk why?
03:40 mohawk (in a number of places)
03:41 sivoais mohawk: that was to double check that I converted all of them to Test::More since done_testing will check that the plan was correct
03:41 mohawk i'd have thought the absence of it also does that?
03:43 sivoais yes, true, it does, but if you try
03:43 sivoais use Test::More tests => 2; pass; print "ok 2\n";
03:43 mohawk i do wish you'd squashed together all the modifications
03:43 mohawk rather than one commit per file
03:44 sivoais and don't have a done_testing; it gives you a warning, but adding done_testing; gives you a BIGGER warning ;-)
03:44 sivoais oh, I have no problem with squashing it, but it's on a branch, so I can do that later.
03:45 sivoais I was doing it bit by bit so I could have a stopping point
03:45 sivoais :-)
03:47 mohawk it's interesting how similar your ideas and mine were in how to do this
03:57 mohawk ok, the rebase has finished
03:57 mohawk now to clear up any messes i made :-)
04:18 mohawk did i mention building this beast takes far too long? :-)
04:18 mohawk sivoais, where are we at with the milestones and whatnot?
04:18 mohawk i'd like to see an issue proposing the splitting up
04:18 mohawk including that PDL::Core become what's in Basic now
04:19 mohawk and that PDL become just a bunch of dependencies, ie a bundle
04:19 sivoais on't need to see an issue. go to the milestone section on GitHub
04:19 mohawk ok
04:21 sivoais shuffling around tags and milestones doesn't show up in IRC or there would be a ton of notifications
04:21 sivoais :-P
04:22 mohawk ha ha
04:22 mohawk just as well
04:23 mohawk throws_ok { print "oops\n" if $pa } qr/multielement/;
04:23 mohawk is giving:
04:23 mohawk multielement piddle in conditional expression at t/bool.t line 16.
04:23 mohawk help :-)
04:24 mohawk (as in, the test is croaking
04:24 mohawk )
04:25 mohawk oh, somehow Test::Exception managed to get removed
04:25 mohawk duh
04:31 mohawk ok, getting there
04:43 mohawk sivoais, pushing now
04:43 mohawk apologies for any mangling
04:43 opkick [pdl] mohawk2 force-pushed test-cleanup from dfd7d29 to 0a91f23: http://git.io/xXCA
04:43 opkick pdl/test-cleanup 2887450 Ed J: Add Test::Deep, move test dependencies to TEST_REQUIRES
04:43 opkick pdl/test-cleanup 8c73e3e Zakariyya Mughal: convert aaa_load.t to Test::More
04:43 opkick pdl/test-cleanup 26b8a09 Ed J: Convert tests to Test::More
04:44 sivoais heh, I will look over it in about an hour. I have to drive home.
04:44 mohawk now i'm looking forward to seeing the coverage
04:44 mohawk should be ready about time you get home
04:44 mohawk travel safely, amigo
04:49 mohawk this is insanely cool: https://travis-ci.org/PDLPorters/pdl/jobs/54510375#L2845
04:50 mohawk sivoais, it's showing 2.005 as '2000-04-05T22:30:35+00:00' - is that right?
04:54 opkick [pdl] mohawk2 comment on issue #78: Reviewed: looks great! http://git.io/hfba
05:10 opkick [pdl] coveralls comment on issue #53:
05:29 sivoais mohawk: that's what the tag says and MetaCPAN has <https://metacpan.org/release/KGB/PDL-2.005> on "Apr 06, 2000"
05:30 mohawk great
05:30 mohawk as long as it's not an error
05:30 sivoais heh, yeah, if you look at the numbering, 2.005 is followed by 2.1.1
05:30 mohawk let's not go there
05:31 sivoais hahah
05:31 mohawk anyway, i pushed the new t-c
05:31 sivoais ooh, 62% coverage!
05:31 sivoais what did you do!? :-)
05:31 mohawk that was Proj4
05:32 sivoais ah, you went over all the autogen'ed functions
05:32 mohawk that code it generates is just a bunch of aliases t_proj_*, so i saw how it generated them, and checked each doesn't produce undef
05:32 mohawk yes i did
05:32 sivoais awesome!
05:33 mohawk adding 13% = 1/8 = about 2000 lines covered
05:34 mohawk grin
05:34 mohawk next big win = lib/PDL/Graphics/PGPLOT/Window.pm with 1212 lines missed
05:36 sivoais all the PDL::Graphics::* modules need better docs... quite sorry state right now
05:36 sivoais thankfully we can now put images in there that will show up on MetaCPAN
05:49 mohawk yes
06:09 sivoais mohawk: what needs to be done on test-cleanup?
06:10 mohawk that's in your hands now
06:11 mohawk in a perfect world, we'd crunch the commits around a bit to make them pretty
06:11 mohawk but if you're content with the result, then so am i
06:12 sivoais heh, it's cleanup... the end result is pretty but the process of pulling out the weeds is not :-)
06:12 * mohawk frowns
06:12 mohawk ;-)
06:37 mohawk sivoais,as far as i'm concerned both changes and t-c are reviewed
06:38 mohawk naturally we'll hold off on cf until post 2.008
06:38 mohawk but what about t-c?
06:54 sivoais I think we can get that in after 2.008. no need to rush it in. The diff between 2.007 and 2.008 is big enough ;-)
06:54 mohawk indeed
06:54 mohawk ok, up to you
06:55 sivoais also, I need sleep. can't be driving without it. good localtime()! o/
06:56 mohawk grin
06:56 mohawk you were supposed to have gone ages ago
07:15 drrho joined #pdl
07:57 drrho joined #pdl
09:05 drrho joined #pdl
14:19 opkick [pdl] wchristian pushed 1 new commit to master: http://git.io/hTf5
14:19 opkick pdl/master a786102 Chris Marshall: Remove debugging lines from fix to sf.net #256
15:00 travis-ci PDLPorters/pdl#135 (master - a786102 : Chris Marshall): The build passed.
15:00 travis-ci Change view : https://github.com/PDLPorters/pdl/compare/fc32b6ba1c3c...a78610265296
15:00 travis-ci Build details : http://travis-ci.org/PDLPorters/pdl/builds/54575014
19:58 mohawk i'm getting a build failure on linux with gmake -j4 now that i've installed proj4 and am therefore exercising building PDL::GIS::Proj and PDL::Transform::Proj4
19:59 mohawk investigating
20:48 sivoais joined #pdl
20:54 opkick [pdl] wchristian pushed 1 new commit to master: http://git.io/hLWF
20:54 opkick pdl/master 547c66a Derek Lamb: Change printf conversion flags to silence many -Wformat warnings on clang....
20:56 mohawk i think i know why the GIS build is failing
20:58 mohawk it's related to interdependency between the two modules
20:58 mohawk them being separate is the problem
20:58 mohawk i need to smoosh them together
20:59 mohawk ...although that creates a problem with non-xsmulti EUMM
20:59 mohawk i'll figure it
21:00 mohawk this will involve creating (shudder) an SF bug
21:06 opkick [pdl] zmughal opened issue #79: SF#375 Storable files incorrectly processed from older versions http://git.io/hL01
21:06 opkick [pdl] zmughal opened issue #80: SF#376 PDL Segmentation fault working with Storable restored PDL http://git.io/hL0M
21:07 mohawk sivoais, i'm just making another
21:09 sivoais I'll mirror it on GH when I get a chance :-P
21:11 mohawk https://sourceforge.net/p/pdl/bugs/377/
21:14 mohawk sivoais++
21:14 opkick [pdl] zmughal opened issue #81: SF#377 Release-blocker: PDL::Transform::Proj4 not building under latest dev EUMM http://git.io/hLgt
21:19 mohawk sivoais++
21:44 mohawk i think the proj4 solution may be to actually complete and use Alien::Proj4
21:44 mohawk (the one in inc/)
21:44 mohawk how exciting
21:47 opkick [pdl] mohawk2 comment on issue #81: The text of the ticket:... http://git.io/hLDk
22:00 travis-ci PDLPorters/pdl#136 (master - 547c66a : Derek Lamb): The build passed.
22:00 travis-ci Change view : https://github.com/PDLPorters/pdl/compare/a78610265296...547c66aacbed
22:00 travis-ci Build details : http://travis-ci.org/PDLPorters/pdl/builds/54630641
22:42 opkick [pdl] mohawk2 created proj4fix (+1 new commit): http://git.io/htJy
22:42 opkick pdl/proj4fix 38b2105 Ed J: Add Travis case building with dev EUMM
22:44 opkick [pdl] mohawk2 opened pull request #82: Fix Proj4 building on dev EUMM (master...proj4fix) http://git.io/htU2
22:52 opkick [pdl] mohawk2 force-pushed proj4fix from 38b2105 to b9be0ce: http://git.io/htkH
22:52 opkick pdl/proj4fix b9be0ce Ed J: Add Travis case building with dev EUMM
23:12 opkick [pdl] mohawk2 merged proj4fix into master: http://git.io/htYd
23:12 opkick [pdl] mohawk2 closed pull request #82: Fix Proj4 building on dev EUMM (master...proj4fix) http://git.io/htU2
23:12 mohawk that's the opposite of true - it's not fixed
23:12 mohawk but i want the visibility of the EUMM blead on master
23:26 opkick [pdl] mohawk2 pushed 1 new commit to proj4fix: http://git.io/htZP
23:26 opkick pdl/proj4fix 6f22050 Ed J: Have Proj modules use Alien::Proj4 in inc
23:46 opkick [pdl] mohawk2 comment on issue #81: A further issue has come to light: inter-dependency between other modules. PDL::Slatec (in Lib/Slatec) wants to use PDL::MatrixOps to build. This was masked by my having an installed PDL, which slatec.pd was able to load (strictly, the distro MatrixOps.pm had been installed, which DynaLoaded the installed MatrixOps SO). http://git.io/htlh

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