Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2015-03-13

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

All times shown according to UTC.

Time Nick Message
15:19 mohawk i'd say the closest P6 either should or even can get to XS is either FFI or JNI-style
15:20 mohawk (due to the multi VM issue)
15:22 mohawk if i'm wrong, i'd love to learn how :-)
15:25 mohawk folks, (esp chm and sivoais) the current MANIFEST on PDL master is slightly wrong - contains 2 generated files (Basic/Core/pdlbadvalinit.c and Basic/Core/pdldataswitch.c) and a no-longer-existent t/issue-sf-368.t
15:26 mohawk i'm going to make a PR that fixes these (including a MANIFEST.SKIP which i'm sure was the source of this), plus includes a test for existence of all MANIFEST entries
15:26 mohawk mea culpa, i failed to add my t/00-report-prereqs.t to MANIFEST
15:27 mohawk so we have a minor process problem which i will technologically fix
15:31 sivoais :-)
15:32 sivoais link me it when you're done
15:32 sivoais I think I need to add some of the other tests I created to the MANIFEST too.
15:57 mohawk not according to my "make realclean ; perl Makefile.PL" quick check
15:58 mohawk anyway, i'm trying to help Mithaldu on #inline, one thing at a time :-)
16:30 mohawk so i'm going to temporarily abandon the inline::c dramas i'm finding and fix up lldp1 at least
16:43 opkick [pdl] mohawk2 force-pushed lldfp1 from ac82533 to 911ad92: http://git.io/x2va
16:43 opkick pdl/lldfp1 b229309 Craig DeForest: remove references to progenitors and mutators.  Update core version number.
16:43 opkick pdl/lldfp1 9b7ee23 Craig DeForest: some typos ("propogate"->"propagate" throughout; some typos in CallExt documentation)
16:43 opkick pdl/lldfp1 d89abb8 Craig DeForest: more cleanup of pdl.h.PL. Also, cut out foomethod() nonsense since it is not...
16:44 mohawk sivoais, please review
16:45 mohawk ...while we wait for travis to do its "thang"
16:49 opkick [pdl] mohawk2 deleted pdl-eq-with-string-warning at 4b77820: http://git.io/pQt6
16:54 opkick [pdl] zmughal comment on issue #52: It looks good to me. Just waiting on Travis. http://git.io/pQmj
16:54 mohawk sivoais, thanks
16:54 mohawk sivoais, current master is failing 1/2 on travis
16:55 mohawk PR "fails", push succeeds
16:55 mohawk i thought we prevented that?
16:55 sivoais hmm, where does it show that?
16:55 mohawk https://github.com/PDLPorters/pdl/branches/all
16:56 sivoais mohawk: oh, that's because Travis grabbed the PR ref *after* I merged
16:57 sivoais if you look at the lines in the logs, you'll see
16:57 mohawk possible process improvement?
16:57 mohawk don't worry, i believe you :-)
16:58 mohawk may i suggest a note in DEV*?
16:59 sivoais it's caused by doing the rebase, merge --no-ff from the command line. Travis isn't very smart
17:00 mohawk i don't see how that's different from just pushing new stuff?
17:00 mohawk (direct to master)
17:01 mohawk is that because master can't be merged?
17:01 mohawk why was it even trying to merge pull/65?
17:03 sivoais After rebasing, I pushed the rebased changes back to the topic branch. Then I switched over to master to do the merge and pushed that.
17:04 sivoais GitHub detects the merge, but I'm guessing GitHub removes the refs immediately after merging
17:05 sivoais the +ref/65/merge ref
17:05 sivoais also, wow, we are at 65 issues already :-)
17:08 mohawk ha ha
17:08 mohawk but to our shame we still have 30-ish open
17:08 mohawk let's fix that
17:24 opkick [pdl] mohawk2 created testmanifest (+3 new commits): http://git.io/pQRd
17:24 opkick pdl/testmanifest e94ce6e Ed J: Basic/Core/pdl{badvalinit,dataswitch}.c MANIFEST -> MANIFEST.SKIP
17:24 opkick pdl/testmanifest f0500ea Ed J: Remove t/issue-sf-368.t from MANIFEST
17:24 opkick pdl/testmanifest a8017fe Ed J: Add manifest-checking test
17:25 opkick [pdl] mohawk2 opened pull request #66: Test manifest (master...testmanifest) http://git.io/pQ0w
17:25 mohawk wow, the tools we are making are kind of kicking entropy's pasty white ass
17:26 sivoais hehe, looking at the PR now
17:27 mohawk lldfp1 is all green so far on travis
17:29 opkick [pdl] zmughal comment on issue #66: Nice! This will be very useful. http://git.io/pQud
17:46 mohawk sivoais, once lldfp1 has a green light, are you happy for me to merge?
17:46 mohawk i'm stopping build 90 because i'll then rebase and repush
17:47 mohawk green light
17:47 opkick [pdl] zmughal opened issue #67: Create a hand-written Changes file http://git.io/pQiD
17:48 opkick [pdl] coveralls comment on issue #52:
17:48 mohawk sivoais, on your command, i will unleash lldfp1
17:48 sivoais mohawk: unleash the kraken^W lldfp1
17:48 sivoais !!
17:49 mohawk done
17:49 mohawk now to rebase/push testmanifest
17:50 opkick [pdl] mohawk2 closed pull request #52: longlong-double-fix resurrection part 1 (master...lldfp1) http://git.io/xgaV
17:50 opkick [pdl] mohawk2 merged lldfp1 into master: http://git.io/pQXO
17:51 opkick [pdl] mohawk2 force-pushed testmanifest from a8017fe to b8c25c5: http://git.io/pQ1m
17:51 opkick pdl/testmanifest 19f60b5 Ed J: Basic/Core/pdl{badvalinit,dataswitch}.c MANIFEST -> MANIFEST.SKIP
17:51 opkick pdl/testmanifest 77076b6 Ed J: Remove t/issue-sf-368.t from MANIFEST
17:51 opkick pdl/testmanifest b8c25c5 Ed J: Add manifest-checking test
17:52 opkick [pdl] mohawk2 deleted lldfp1 at 911ad92: http://git.io/pQ1o
17:54 mohawk sivoais, i want to stop the build on master since by definition we know the results, so it can build testmanifest
17:54 mohawk make sense to you?
17:54 sivoais yes, sure
17:58 mohawk ok
17:58 mohawk so now travis is just sitting there
17:58 mohawk that doesn't seem ideal :-)
17:59 sivoais you get what you pay for! :-P
18:00 sivoais I was looking at Tapper as a CI solution <http://tapper.github.io/Tapper/>
18:00 mohawk wake up, travis, dangit
18:00 mohawk go on?
18:00 sivoais frere Travis, frere Travis, domez vous...
18:00 sivoais *dormez
18:00 mohawk dormez
18:00 mohawk grin
18:01 sivoais :-P
18:01 mohawk sivoais++
18:01 mohawk me-- # too slow on the pedantry
18:01 sivoais so, Tapper seems to be able to launch virtual machines
18:02 sivoais if needed, so I'm thinking we could get Windows setup and actually have a way of making sure everybody tests locally under the same setup
18:02 sivoais plus, it can run benchmarks so we can look at performance regressions
18:03 sivoais and it looks like AMD used it for exactly the type of things that PDL will have to move towards in the future: GPU and OpenCL things
18:03 sivoais e.g., <https://github.com/tapper/Tapp​er-TestSuite-OpenCL-Mandelbrot>
18:07 mohawk great, now i have "frere jacques" stuck in my head
18:07 sivoais oh, and there's another thing we can go towards. I was talking about it with pdurbin
18:07 sivoais hahaha
18:07 mohawk on the plus side, that's reminded me to email the nice french lady i met yesterday
18:08 mohawk go on re pdurbin?
18:08 sivoais it's a little on the academic CS side of things, but generative testing looks like a neat way to test math libraries
18:08 sivoais talked about it at <http://irclog.greptilian.com/sourcefu/2015-03-11>
18:09 mohawk ok
18:09 sivoais so we can talk about the properties of functions and generate test cases until they fail
18:10 sivoais so either our properties are wrong or we found a bug! :-)
18:10 mohawk sivoais, looks good
18:10 mohawk bring the guy in
18:11 mohawk with gh we have an attractive modern environment
18:11 mohawk let's test/prove that assertion!
18:11 pdurbin oh man, you guys are gonna steal him away ;)
18:11 mohawk that's the plan
18:11 sivoais hahah
18:11 pdurbin finders keepers I say >:(
18:11 mohawk you snooze...
18:12 pdurbin touche
18:12 sivoais :-P
18:12 sivoais I haven't seen generative testing for numerical software yet, but I haven't looked very hard
18:12 mohawk sivoais, are you on that?
18:13 mohawk i think that's the 3rd time i've cancelled build 91
18:14 mohawk is someone restarting it?
18:14 * mohawk squints at sivoais
18:14 sivoais I pinged the person on Twitter, but haven't gotten a response yet :-P
18:14 sivoais nope, not me...
18:14 mohawk ha ha, ok
18:16 sivoais this is the closest paper I've seen so far: " Self-Testing/Correcting with Applications to Numerical Problems" <http://www1.icsi.berkeley.edu/~luby/self.html>
18:19 mohawk sivoais, travis seems to have ground to a halt
18:20 mohawk given the modest nature of "testmanifest" (no substantive code changes, distro files / non-PDL test changes only) what do you think of just merging it?
18:20 mohawk oh hang on, travis actually started
18:20 sivoais \o/ crisis averted!
18:20 mohawk ha ha
18:20 mohawk indeed
18:21 mohawk although weirdly, 92.1 (the coverage job) stopped itself
18:21 mohawk then restarted itself
18:21 mohawk now it's just sitting there
18:21 mohawk i guess it's waiting for a slot as there are 4 others going
18:21 mohawk ok, crisis is suspended
18:21 mohawk time to change the lightbulbs back
18:24 mohawk https://youtu.be/Qa_gZ_7sdZg?t=1m25s
18:39 mohawk sivoais, green lights so far on testmanifest
18:39 mohawk are you happy for me to merge it once it's in?
18:55 mohawk i'll take that as a yes
18:56 mohawk if the coverage job ever finishes (17 mins and counting)
19:08 sivoais heh. yeah, go ahead once it finishes
19:10 opkick [pdl] coveralls comment on issue #66:
19:13 opkick [pdl] mohawk2 merged testmanifest into master: http://git.io/p70y
19:14 mohawk i cancelled #93 because again, exactly the same as tm
19:15 opkick [pdl] mohawk2 deleted testmanifest at b8c25c5: http://git.io/p7Ea
19:33 travis-ci PDLPorters/pdl#93 (master - b8c25c5 : Ed J): The build was canceled.
19:33 travis-ci Change view : https://github.com/PDLPorters/pdl/c​ompare/911ad924fd7b...b8c25c5cfdcc
19:33 travis-ci Build details : http://travis-ci.org/PDLPo​rters/pdl/builds/54290830
19:33 mohawk wow, that's rapidly responsive
19:33 * mohawk eye-rolls
19:54 mohawk sivoais, i want to cherry-pick the t/ops.t -> Test::More commit out of the "hainest-assgn_badflag" branch
19:54 mohawk thoughts?
19:57 sivoais mohawk: can it be turned into a TODO test and merged?
19:58 mohawk uh, it's a rewrite of t/ops.t to use T::M
19:58 mohawk so, not really
19:58 sivoais cool, that's much more useful. go for it!
19:59 mohawk great
20:01 opkick [pdl] mohawk2 pushed 1 new commit to master: http://git.io/p7QH
20:01 opkick pdl/master d751ac7 Ed J: Rewrite t/ops.t with Test::More
20:03 opkick [pdl] mohawk2 force-pushed hainest-assgn_badflag from 6f8d393 to 2f47a52: http://git.io/xgxg
20:03 opkick pdl/hainest-assgn_badflag 2f47a52 Tim Haines: Add propagation of badflag with .= (Ops::assgn) for sf.net bug 3543056.
20:05 mohawk sivoais, did you see my ideas on #toolchain about assert_blib?
20:05 mohawk might be a useful thing for your test-cleanup
20:07 mohawk sivoais, i'm just rebasing your test-cleanup - this num_ok thing you've done, i strongly disagree with
20:08 sivoais I said it was a WIP! I am going to squash those out!
20:08 mohawk grin
20:09 sivoais didn't see them. I will take a look once i am not afk
20:11 mohawk no problem
20:11 mohawk i'll c+p here
20:12 mohawk i'm imagining something like: assert_blib 'PDL'; # fails if $INC{'PDL.pm'} !~ m:blib/lib/PDL.pm:
20:14 mohawk sivoais, i'm kind of assuming you'll go back and rearrange the commits so the num_ok isn't in the history?
20:15 sivoais yep! If you could leave a comment on the PR, that would help
20:15 mohawk wilco
20:15 mohawk just about to force-push it
20:16 mohawk hey, look at this bit from TODO
20:16 mohawk Need to test PDL build from scratch both WITH_BADVAL and without
20:16 mohawk * Find a way to automate this process for release testing
20:17 sivoais haha!
20:17 mohawk that whole paragraph - isn't that just cpan smoke?
20:19 opkick [pdl] wchristian force-pushed test-cleanup from eb07f84 to f458fa1: http://git.io/xXCA
20:19 opkick pdl/test-cleanup c27ec1f Zakariyya Mughal: convert aaa_load.t to Test::More
20:19 opkick pdl/test-cleanup dec5c01 Zakariyya Mughal: consolidate test helper function of form `ok( $no, $result )`...
20:19 opkick pdl/test-cleanup d9ff4eb Zakariyya Mughal: done_testing not needed
20:21 opkick [pdl] mohawk2 comment on issue #53: Per @zmughal request, I am adding to here my request that before this can be merged, the `num_ok` changes be edited out of the history. http://git.io/p7py
20:22 opkick [pdl] mohawk2 force-pushed doc/vsearch-example from b635451 to a96743c: http://git.io/p7hw
20:22 opkick pdl/doc/vsearch-example 1cfc461 Zakariyya Mughal: add vsearch example...
20:22 opkick pdl/doc/vsearch-example a96743c Zakariyya Mughal: add links to other vsearch_* functions
20:24 opkick [pdl] mohawk2 force-pushed longlong-double-fix from bbc3962 to 59bb36a: http://git.io/p7j1
20:24 opkick pdl/longlong-double-fix 69f160d Chris Marshall: Implement PDL_Generic for PDL_Anyval type
20:24 opkick pdl/longlong-double-fix c4a6d0a Chris Marshall: Add 8byte array as first member of PDL_Union...
20:24 opkick pdl/longlong-double-fix 266a16a Chris Marshall: Add ANYVAL_IS_EQ() macro to replace PDL_Anyval == PDL_Anyval constructs
20:44 travis-ci PDLPorters/pdl#94 (master - d751ac7 : Ed J): The build passed.
20:44 travis-ci Change view : https://github.com/PDLPorters/pdl/c​ompare/b8c25c5cfdcc...d751ac7eb426
20:44 travis-ci Build details : http://travis-ci.org/PDLPo​rters/pdl/builds/54296942
21:32 opkick [pdl] coveralls comment on issue #51:
22:04 opkick [pdl] coveralls comment on issue #53:
22:35 opkick [pdl] coveralls comment on issue #58:
23:19 sivoais nice work, mohawk!
23:21 mohawk sivoais, thanks? ;-)
23:22 sivoais maybe I can work a little more on test-cleanup... I think I can finish it off with a couple hours of work
23:26 mohawk sivoais++

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