Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2015-02-18

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

All times shown according to UTC.

Time Nick Message
00:44 opkick [pdl] wchristian pushed 21 new commits to master: http://git.io/Nji3
00:44 opkick pdl/master a541c04 Ed J: Clean more generated files
00:44 opkick pdl/master f8ad0a8 Ed J: No no_index for FFTW since is now separate distro
00:44 opkick pdl/master 83d00ff Ed J: pdl.c no need to be .PL
01:46 mohawk chm++
06:27 mohawk just MRed the attached patch for bug 370
06:56 mohawk and i commented on 346
07:03 mohawk sivoais, if you're feeling brave, you can go into PDL::PP and fix this: https://sourceforge.net/p/pdl/bugs/363/
07:06 mohawk should just be a simple RE in the OtherPars handler
07:06 mohawk don't forget to add a test first :-)
14:28 lungching joined #pdl
17:43 sivoais mohawk: taking a look
17:44 sivoais the images display  now! <https://metacpan.org/pod/release/ZMUGHAL/Devel-IPerl-0.002/example/20150209_IPerl_display_demo.pod>
17:46 mohawk sivoais++
18:07 mohawk sivoais, any joy on PDL::PP?
18:12 sivoais mohawk: something came up, but I can look now
18:12 mohawk you rock
18:13 sivoais I had an idea for using Convert::Binary::C with PDL for reading in binary structs quickly. I'll take a note to do something with that later
18:14 mohawk when you do, could you compare that with Inline::Struct?
18:19 sivoais good idea
18:23 jberger isn't it kinda perfect that ::PP is now defacto "pure-perl" ie "easy to install" and yet in PDL, for historical reasons, it means "the hardest bit of PDL to understand"
18:24 mohawk ha ha ha
18:24 mohawk jberger++
18:24 mohawk of course, you're making a rookie mistake in implying it's humanly possible to understand at all
18:25 jberger I'm assuming that a human wrote it, which almost implies that someone understood it at some point
18:25 mohawk that may be a bridge too far :-)
18:25 mohawk i have written code that by the end of it, i no longer fully understood
18:26 jberger true enough
18:28 estrabd is there going to be another scientific computing panel/BoF at YAPC this year?
18:28 mohawk jberger, got any views on show-stopper PDL bugs?
18:28 mohawk in fact, anyone got any?
18:30 jberger mohawk: I'm not gonna lie, I really haven't been keeping that close to the PDL specifics these days
18:30 mohawk i hear you
18:30 jberger plus the ticket tracker is on sf.net
18:31 mohawk i am not unaware
18:31 mohawk but despite that, i'm asking you to hold your nose and look :-)
18:31 sivoais so out of those two build bugs, <http://sourceforge.net/p/pdl/code/merge-requests/21/> fixes <http://sourceforge.net/p/pdl/bugs/370/>
18:32 sivoais and I'm looking into the other one <http://sourceforge.net/p/pdl/bugs/346/>
18:33 sivoais :-( I have to click 3 times to open a merge request. Poor UI design right there
18:35 sivoais oh, and I need to also look at <http://sourceforge.net/p/pdl/bugs/363/>
18:35 sivoais I think the 346 bug (g77 issue) can be addressed by just running Travis-CI
18:36 sivoais mohawk: you were saying 363 is just a regex?
18:37 mohawk i think 346, the "do the param names clash with PDL ones", could just be checked using an RE
18:38 mohawk wait, let me check that
18:38 sivoais 363 is the param names one. 346 is the fortran one
18:38 mohawk so it is
18:38 * sivoais dredges up SF account cred
18:39 mohawk yes, 363 seems like a simple RE
18:43 sivoais I'm going to look into the Fortran first
18:43 mohawk ok
18:43 mohawk what would you like me to do?
18:44 sivoais maybe look for things that need to be closed. I saw this <https://sourceforge.net/p/pdl/bugs/342/>
18:44 sivoais that looks fixed
18:48 mohawk sounds like a quick email to chm is called for
18:49 mohawk he was very responsive to my MR yesterday
19:06 sivoais who deleted pdldoc?
19:07 sivoais weird
19:07 mohawk i did [1]
19:07 mohawk [1] your old makefile thought of it as generated, which it was
19:07 sivoais mohawk: what's going on here <https://travis-ci.org/zmughal/p5-PDL/jobs/51270811>?
19:07 mohawk [2] now it's not generated
19:08 sivoais i saw that in the git log
19:08 sivoais make: *** No rule to make target `pdldoc', needed by `pdldoc.pod'.  Stop.
19:09 mohawk yes, i see that
19:09 mohawk is this definitely getting all the right files, in a clean dir? git clone --depth=50 --branch=travis-ci git://github.com/zmughal/p5-PDL.git zmughal/p5-PDL
19:11 sivoais ok, it looks like the build works locally O_O
19:11 sivoais no dep issues coming up from make
19:11 sivoais let's try rebuilding the Travis-CI jobs
19:14 sivoais mohawk: there's another issue with a patch from Petr at <https://sourceforge.net/p/pdl/bugs/350/>
19:16 mohawk ok, seen
19:16 mohawk i'll make an MR with that as well
19:18 vicash left #pdl
19:23 sivoais the pdldoc thing shouldn't be a problem...
19:24 sivoais I suspect the make might be buggy?
19:24 mohawk if you can repro it, i'll agree
19:24 mohawk chm had the same initial problem
19:25 mohawk but then he said:
19:25 mohawk "I figured out the problem:  make clean
19:25 mohawk deletes the new non-PL files...
19:25 mohawk "
19:26 mohawk so i think this is just because the git clone isn't wiping out old files (or not running perl Makefile.PL first, which will render it moot)
19:26 mohawk ok, now running test on 350 fix...
19:29 sivoais trying it out with : git clean -x -f
19:30 sivoais it builds fine
19:30 mohawk there ya go
19:31 sivoais so something weird is happening on travis-ci. the version there should be make 3.81
19:34 mohawk ok, that passed
19:37 mohawk https://sourceforge.net/p/pdl/code/merge-requests/22/
19:43 jberger the documentation indexing thing is a mess
20:07 sivoais I've been thinking about comparing packages along these lines: Matlab == Linear algebra, Numpy == Vector calculus, PDL == Tensor calculus
20:08 mohawk i hope you're also contemplating some PDL embrace and extend
20:10 sivoais oh yes... embracing and extending Numpy is next on my list. Inline::Python makes it easy
20:10 sivoais then comes Octave
20:11 sivoais then Matlab if anyone wants it at that point :-D
20:11 mohawk damn, dog
20:11 sivoais R is pretty much done. Just some more integration with plotting and everything will be good :-)
20:12 sivoais I'll need either Inline::CPP or FFI w/ C++ for Octave
20:12 mohawk Inline::CPP, i'm here to help
20:15 sivoais jberger: it looks like the GitHub site doesn't work with Google indexing because of the JS <https://encrypted.google.com/search?hl=en&amp;q=site:http://pdlporters.github.io/>
20:17 jberger certainly true for the documentation at least
20:18 jberger we could probably "static-ize" some more of the site to make it more indexable
20:18 mohawk we really should
20:18 sivoais yeah, that should be *easy*. There's no reason to use PHP for something like this
20:20 jberger the problem is that doing partial includes in a static site means you either do it live in js or you do it at deploy time with some manager
20:20 jberger which would be sufficient
20:21 jberger preaction's statocles is now the manager for our chicago.pm.org

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