Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2013-02-06

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

All times shown according to UTC.

Time Nick Message
00:15 Molaf_ joined #pdl
02:18 zpmorgan joined #pdl
02:18 jberger joined #pdl
14:49 Mithaldu jberger: your Tie::Array::CSV has borken tests on windows
14:49 Mithaldu i am disappoint
14:50 jberger yeah, I've been meaning to look at that
14:50 jberger shouldn't Text::CSV do the right thing (the problem is line endings right?)
14:50 Mithaldu no idea, forking right now
14:50 Mithaldu btw
14:50 jberger I don't have a windows computer
14:50 Mithaldu your github link is a bit wonky
14:50 jberger please do, I'd be happy for a patch
14:50 jberger thats one of my first releases
14:50 Mithaldu ah, cheers :)
14:51 jberger HAHA its the one that points at issues right
14:51 jberger I remember that
14:51 Mithaldu yes
14:51 jberger its fixed in git IIRC
14:51 Mithaldu it's more useful than the ones linking to the .git dir
14:51 Mithaldu but still odd :P
14:51 jberger I copy/pasted wrong
14:51 Mithaldu git checkout -b win32_fix
14:51 Mithaldu i should make a shortcut for that
14:52 Mithaldu yes, it's fixed in git
14:52 jberger I don't think I do anything is platform specific though, shouldn't Text::CSV handle the line endings?
14:52 Mithaldu haven't looked at the code yet!
14:52 Mithaldu will tell you once i know :)
14:52 jberger haha ok, I'll let you have a look
14:52 jberger it really has been something I have meant to look at
14:52 Mithaldu like, i'm setting up the komodo project at this very time
14:52 jberger I'm happy you are looking
14:54 jberger another thing I would change on those tests is that I rely on the same test case for too long
14:54 jberger by the end its hard to remember what the state of the file/object should be
14:55 jberger I probably should split it into several cases
14:55 Mithaldu ah yeah, that's an issue
14:55 Mithaldu do your tests break if execution order is randomized?
14:56 Mithaldu Tie::File already borks it
14:58 jberger it certainly would
14:58 Mithaldu Tie::File is kind of dumb
14:58 jberger dumb as in naive or dumb as in not good?
14:59 jberger is there something better?
14:59 Mithaldu it assumes the record separator is \r\n
14:59 jberger Tie::File does?
14:59 Mithaldu which is normally hidden away by perl's input output mangling
14:59 Mithaldu yes
14:59 jberger really?
15:00 jberger well, could that be overridden in Tie::Array::CSV's invokation
15:00 Mithaldu yes
15:00 jberger you have my blessing
15:00 * jberger thinks that is dumb
15:01 Mithaldu it is
15:02 Mithaldu well, tests pass now
15:02 Mithaldu patch inc
15:02 jberger thanks!
15:03 jberger I thought it was a clever little module
15:03 Mithaldu if it works right it'll help me SO MUCH
15:03 jberger probably incredibly slow compared to some other things
15:03 Mithaldu because the api of text::csv_xs is so arse
15:04 jberger that why I wrote it
15:04 Mithaldu i'm not concerned about speed atm :)
15:04 jberger (the help, not necessarily the api)
15:04 Mithaldu :)
15:04 jberger yeah, I was surprised that there wasn't something really like it before
15:05 jberger https://metacpan.org/module/Tie::CSV_File but its really odd
15:06 Mithaldu well
15:06 Mithaldu tieing is normally not advisable for things like this due to the performance issues
15:06 Mithaldu pull req sent
15:06 * jberger looks
15:08 jberger merged!
15:08 jberger let me clean up some of the meta stuff and I will push it out soon
15:08 jberger thanks again!
15:08 Mithaldu happy to help :)
15:09 Mithaldu wow
15:09 Mithaldu semi-panic: attempt to dup freed string at C:/Program Files (x86)/ActiveState Komodo IDE 7/lib/support/dbgp/perllib/perl5db.pl line 4573, <$fh> line 50.
15:10 Molaf__ joined #pdl
15:11 jberger in Tie::Array::CSV?
15:11 jberger under the debugger
15:11 Mithaldu yeah, the debugger doesn't like tied stuff much
15:11 jberger much less double tied
15:12 Mithaldu it's cool if i skip over the relevant lines though, so that's fine
15:12 jberger haha, wow, I'm still using dzil on this
15:12 jberger I think I'll take it back to M::B
15:12 Mithaldu why?
15:12 Mithaldu dzil is fine :)
15:13 jberger I don't hate dzil, but I find it hard to make it do what I want sometimes, and I now know M::B really well
15:13 jberger maybe I'll leave it alone
15:13 jberger how do I do meta merge/
15:14 jberger ?
15:14 Mithaldu oh, in that case i can't help, i'd need to look that up with with a flu i'm NOT ready to look at cpan meta stuff
15:15 jberger haha
15:15 jberger nm
15:15 jberger thanks for what you've already done
15:15 Mithaldu and thanks to you for a neat little module tht is quite useful :)
15:16 Mithaldu https://gist.github.com/wchristian/4723193
15:21 jberger thats purdy
15:21 jberger and I was about to ask you what Perl you were using
15:21 jberger I'm thinking of requiring 5.8
15:21 jberger that way Tie::File is core
15:21 jberger and I think everyone is off of 5.6
15:22 jberger oh nm, why not allow 5.6
15:22 jberger ignore me
15:34 Mithaldu jberger: i'm on windows, so normally one of the newer ones
15:34 Mithaldu also, i'd only limit it if the test reports come back blargh
15:36 jberger released
15:36 jberger cheers
15:38 Mithaldu yay :)
15:38 jberger https://metacpan.org/module/JBERGER/Tie-Array-CSV-0.06/lib/Tie/Array/CSV.pm
15:38 jberger go forth and be productive
15:38 Mithaldu btw, the contributor format is from Moo
15:38 Mithaldu i think it's quite useful :D
15:40 jberger indeed
15:40 jberger I think I will use that in the future
15:40 jberger I'm always happy to thank contributors
15:40 jberger open source is mostly a thanks currency after all
15:40 jberger Mithaldu, I do have a question
15:41 jberger what does this do?
15:41 jberger say state $c++ . " / " . ~~ @file;
15:41 Mithaldu hahaha
15:41 Mithaldu prints a progress counter
15:41 Mithaldu ~~ scalarifies @file
15:41 Mithaldu i could probably even skip that
15:41 Mithaldu and state is like my, only it basically declares a global in that specific scope
15:41 jberger I always forget the inchworm
15:42 jberger it looks like a smartmatch to me, but with the left side being an operator I was very confused
15:42 jberger I love state
15:42 jberger that wasn't the issue
15:42 Mithaldu i never use smartmatch
15:42 Mithaldu smartmatch is ew :P
15:42 jberger I wouldn't either
15:43 jberger concat forces scalar context right?
15:44 jberger so save yourself a couple bytes :[)
15:45 run4flat yeah, string concatenation forces scalar context
15:45 Mithaldu i think so
15:45 Mithaldu i kinda just used it out of habit :)
15:45 jberger man, rereading this documentation is hilarious
15:45 jberger I was so proud of my hashref or not arg parsing
15:45 run4flat argh, not-so-good-news: http://m.itworld.com/big-data/340570/python-gets-big-data-boost-darpa?mm_ref=http%3A%2F%2Fdevelopers.slashdot.org%2Fstory%2F13%2F02%2F06%2F0225259%2Fpython-gets-a-big-data-boost-from-darpa%3Futm_source%3Drss1.0moreanon%26utm_medium%3Dfeed
15:46 jberger I probably should pair that down
15:46 jberger s/pair/pare
15:47 jberger harumph on that link
15:47 jberger yet another reason to clean up PDL
15:47 jberger not its gotta compete with mind-readers
15:48 jberger s/not/now/
15:48 run4flat and this troubles me, too: http://blaze.pydata.org/
15:49 run4flat fortunately, PDL already has some of the stuff that Numpy does not, and Blaze perports to address
15:49 run4flat like missing values
15:49 jberger yes, but it'll be new an buggy
15:49 run4flat Blaze's stuff? Yeah
15:49 run4flat the real question is: how do we get a Darpa grant to improve PDL?
15:50 jberger first someone has to be able to hack on it
15:50 jberger for that we need cleanup
15:50 jberger then we need some respect for Perl
15:50 jberger Ovid just blogged about the next Perl 5 version number
15:51 run4flat Here's the DoD XData solicitation: http://www.darpa.mil/Our_Work/I2O/Programs/XDATA.aspx
15:51 jberger personally I think that we need to call it Perl 7 or no external agency gives money to support the thing thats about to be replaced by Perl 6
15:51 Mithaldu the github link for blaze is a 404
15:52 run4flat ha!
15:52 jberger haha
15:52 jberger its probably an organization now
15:52 run4flat https://github.com/ContinuumIO/
15:52 jberger but thats funny
15:53 run4flat they probably meant blaze-core
15:53 * jberger expected bugs, thought they would be contained to the software, was wrong
15:53 * run4flat laughs
15:54 run4flat But you know what? They have a grant.
15:54 * run4flat thinks grants trump bugs
15:54 jberger indeed
15:54 jberger 3 megabucks
15:54 run4flat yeah
15:59 jberger run4flat, look a new release of Tie::Array::CSV : https://metacpan.org/module/JBERGER/Tie-Array-CSV-0.06/lib/Tie/Array/CSV.pm
15:59 jberger and it works on windows! Mithaldu++
16:00 jberger one of my first and favorite modules
16:01 run4flat joined #pdl
16:02 run4flat Yes, I noticed
16:02 run4flat that is, I noticed the discussion. :-)
16:02 jberger k
16:02 jberger just a little excited
16:02 run4flat :-)
16:02 * jberger gets extra open-sourcey when he finds out people actually use his modules
16:03 run4flat This is why open-source is wonderful
16:03 jberger ok back to the real world for a while
16:03 run4flat o/
16:21 opkick joined #pdl
16:21 opkick [pdl01] run4flat pushed 2 new commits to master: http://git.io/s3NZ6Q
16:21 opkick pdl/master c27aa9e Chris Marshall: Fix PDL::Graphics::PGPLOT::Window POD...
16:21 opkick pdl/master c82a5ad Chris Marshall: Merge branch 'master' of ssh://pdl.git.sourceforge.net/gitroot/pdl/pdl
16:21 opkick left #pdl
19:39 sivoais joined #pdl
20:12 sivoais joined #pdl

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