Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2014-08-31

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

All times shown according to UTC.

Time Nick Message
00:36 chm joined #pdl
00:37 chm mohawk:  Could you please send a git format-patch email to me for the PDL-2.007_04.tar.gz devel release.
00:37 chm I would like to catch the sf.net repo up to the new branch.  Thanks, Chris.
00:54 mohawk will do
00:55 mohawk chm, would you like me instead to SF-fork and PR my changes?
01:02 chm Either is fine.  I just want to get your contribution into the official repo.  Thanks.
01:12 mohawk merge requested
01:13 mohawk note that it includes a post-dev release fix as guided by sisyphus
01:13 mohawk if that's a problem, let me know and i'll adjust
01:13 mohawk it also includes the necessary MANIFEST updates
01:14 mohawk you may want to put the _04 tag on commit [37ec5e]
05:23 sivoais from YAPC::EU: Open source deep sky images using perl - StrayTaoist <https://www.youtube.com/watch?v=YjFjSngA_Lo>
05:29 sivoais no code for the above yet. seems like it would be very well-suited to PDL
09:49 osfameron sivoais: ah, was that a good talk?  I missed at yapc (and haven't got around to watching any of the youtube uploads yet)
14:32 vicash hi. quick question. is there a similar function to R's sapply/apply in PDL that I can create a lambda like unit that can be applied to all the data in a PDL
14:36 vicash got it. PDL::Reduce explains
14:40 vicash ok i guess that is wrong. I would like to apply a code-ref that doesn't necessarily reduce the dimension to a single element like sumover
15:39 chm joined #pdl
15:40 chm vicash:  Take a look at thread_define and the cumu* methods
15:47 sivoais osfameron: I'm just watching the YouTube talks now. The talk didn't have enough code examples for my tastes. :-)
15:49 sivoais chm: I was looking at the tests in PDL and noticed that I would have to strip out a lot of lines to get everything over to Test::More.
15:50 sivoais I don't think that will be easy to merge in git. How do you want me to apply my changes? One file per commit?
15:50 sivoais I'll ask on pdl-porters if you like
16:03 mohawk vicash, how would you feel about writing a PDL doc that gives answers to some of the "how do i do in PDL this R thing"?
16:04 mohawk sivoais, the "test::more makes things shorter" is a thing i experienced for other things too - fun, isn't it
16:04 * mohawk does like deleting code
16:16 sivoais hehe
16:21 vicash thanks chm will do. mohawk well i am not a super active user of PDL yet to use a lot of features but yes I can probably give a pull request for PDL::FAQ
16:29 mohawk if you put in a PR with what you found so far and put out a call on pdl-p, it might even revive some PDL activity
16:42 mohawk one commit per file may be tidier, but one giant commit would work just as well
16:42 mohawk i think that any difficulty in merge is moot
16:42 sivoais hhmm, yeah. It just needs to be done.
17:01 chm vicash:  Take a look at PDL::MATLAB as an example.
17:02 drrho joined #pdl
17:03 chm sivoais, mohawk:  As a side note, a large fraction of the test failures in the PDL distribution were bugs in the test code---not PDL.
17:04 chm The clean-up and consolidation with Test::More will definitely improve the ability to test and verify PDL.
17:12 mohawk chm, i hear you
17:13 mohawk i had a test fail on my EU::D changes that was just a slightly broken test
17:13 mohawk on windows, not surprisingly
17:14 mohawk in fact, that was my last test failure on inline-c as well
17:14 mohawk damn windows
17:16 mohawk i infer from the deafening silence to my pdl-p email of yesterday that it's a "no"
17:21 mohawk i see there's a PR from djerius?
17:58 jberger mohawk: the funny thing is that Windows wasn't as much trouble as Mac when developing Alien::Base
18:16 mohawk ha, no kidding!
18:17 mohawk vicash, re https://github.com/Perl5-Alien/Alien-Base/pull/33 - could you add a test please?
18:26 osfameron sivoais: hehe.  actually I think that usually makes for a better talk :-)
18:42 sivoais it depends on the speaker and topic :-P
18:46 mohawk MOAR CODE
18:51 chm mohawk: I don't see a reason to change how PDL developers gain git access which I mentioned yesterday on #pdl
18:53 mohawk indeed
18:54 chm From my point of view, we're going through the stage of contribution to PDL
18:55 chm by new developers in my reponse at 1630 in http://irclog.perlgeek.de/pdl/2014-08-30
18:56 chm Specifically, I would like to see your EUD contribution developed for merge into the PDL master.
18:56 chm I'll be merging in the branch this afternoon and am looking forward to the example/demonstration
18:56 chm of the new feature for PDL.
18:57 chm Given that one benefit is to reduce the amount of boilerplate for PDL based distributions,
18:57 chm I expect that other PDL developers will be interested to see how it works.
18:58 chm Most PDL developers are busy using PDL and turn around time on the mailing list
18:58 chm can be on the order of days/weeks sometimes if they are away or busy.
18:59 mohawk "looking forward to the example/demonstration of the new feature for PDL." - i wrote that yesterday on the ticket
18:59 mohawk did you not see it?
19:00 chm Per my request of yesterday, I'll be taking PDL porters discussions to pdl-porters and no longer address them on #pdl
19:02 chm #pdl is fine for quick questions but some of the issues that you have raised need to be discussed for *real* on pdl-porters
19:03 chm Thanks, o/
19:04 mohawk well that's great
19:05 mohawk i will cut+paste the text i wrote on the request into an email for pdl-p to ignore
19:15 sivoais mohawk: well...there just aren't enough devs and even those don't have lots of free time :-/
19:17 mohawk it's a bit rich to ask someone to write something that you then ignore
19:20 mohawk and then ask them to write something else, which you then also ignore
19:38 sivoais "Image Processing and Data Analysis: The Multiscale Approach" <http://www.multiresolution.com/cupbook.html>
19:39 sivoais Interesting book about image processing. It starts off with wavelets.
19:43 mohawk sivoais, /msg-ed you
19:45 vicash mohawk: working on writing a test right now. will update the pull request when i am ready. also handling the Digest::SHA problem you mentioned
19:46 mohawk vicash, awesome!
19:58 vicash mohawk: is it ok if the test requires a download from the net ?
19:59 mohawk is there a way to fake it by having eg { package LWP::Simple; sub xxx { } }
19:59 mohawk ?
20:00 mohawk a technique i first used to simulate being a mod_perl for testing purposes back in 98 or so
20:00 vicash ok. not sure. i am looking at t/repository.t and trying to add to that itself since it seems like it downloads or fakes a download
20:01 mohawk nothing wrong with adding to/modifying existing tests
20:01 mohawk and if it does an actual download, please change it to also fake that too
20:01 mohawk i don't want fragile tests that depend on "the nets" working
20:02 jberger mohawk: re: "well that's great" cool it
20:03 jberger chm is the reason this project has had ANY momentum for the past many years. PDL is a little slow moving, but it is so because it is a battleship, unlike AB which is slow moving because it is a rowboat
20:03 jberger now PLEASE watch your tone, all of this can be done without it
20:04 mohawk i wish that i believed that
20:05 mohawk if you can guide me on how, i'm all ears
20:07 jberger in the end, its not my or your project, I don't have a PDL commit bit either (I don't think, its been a while since I made any active contributions)
20:08 jberger run4flat resurrected this channel mostly for support, not development, historically the pdl development has been run through the mailing list
20:09 jberger I wanted to move pdl development to github, I made a case, they decided not to, I have to live with it, its not my project
20:09 mohawk i think if they did move to github, development would catch fire
20:09 jberger personally, I think so too
20:09 jberger but its not my call
20:09 mohawk mine was the second ever PR, the first was from a few days before
20:10 jberger to sf.net? that's because that's a brand new thing for them
20:10 mohawk yes, on the sf.net repo
20:10 mohawk https://sourceforge.net/p/pdl/code/merge-requests/
20:11 jberger right, sf.net just converted pdl to the new platform not so long ago, before then it couldn't do PRs at al
20:11 jberger all
20:11 mohawk merge reqs have been on sf for at least a year: https://sourceforge.net/p/forge/site-support/4909/
20:11 jberger but only for projects that had migrated to the new platform
20:12 mohawk wow, 3 years: http://lowendtalk.com/discussion/336/forking-and-merging-over-at-sourceforge
20:12 mohawk well, that's a PDL config choice
20:13 jberger not really, it was a fair amount of work to move, and they finally had to because they were retiring the old platform
20:13 jberger did you do the work of moving it?
20:14 jberger and while we are on the subject, in order to make the case for pdl moving to github, david and I did all the work to do that, including porting the website, rather than just gripe
20:14 mohawk ok
20:15 mohawk so you did a bunch of work for nothing?
20:15 jberger turns out we did, yep, that happens sometimes
20:15 jberger if we didn't try, it definitely wouldn't have happened, though
20:16 mohawk so are you saying the way forward here is for me to do a bunch of work, just on spec?
20:16 jberger and to be honest, it did slow my roll for contributions to PDL, since I understand github, not sf.net
20:16 jberger I'm saying that this is a slower paced project than you might be used to
20:16 jberger but that's no reason for insulting people who have actually been contributing to this project for decades
20:17 jberger (that is not including myself, btw)
20:17 mohawk vicash, nice work but could you rebase?
20:17 vicash i did rebase .. let me try again
20:18 mohawk not sure you did ;-)
20:18 mohawk jberger, please highlight whom i have insulted, and how
20:18 jberger "well that's great, i will cut+paste the text i wrote on the request into an email for pdl-p to ignore"
20:19 mohawk that's moaning
20:19 mohawk where's the insult?
20:20 jberger you and I see it differently then
20:24 mohawk if we unpack what i said, i a) sarcastically said it was great = i don't like it; b) said i would cut+paste (neutral fact); c) said i thought pdl-p would ignore it
20:24 mohawk c) might be seen as insulting, though it's simply a prediction of fact
20:40 jberger on a lighter note, we now know how many angels can dance on the head of a pin!
20:40 jberger http://www.improbable.com/airchives/paperair/volume7/v7i3/angels-7-3.htm
20:41 mohawk nice - a contender for the IgNobel Prize in physics ;-)
20:59 mohawk to finish my point above - the prediction of fact is only insulting in this case if it's true
20:59 mohawk and if so, the problem is the fact, not my saying it
20:59 mohawk ergo, i can't agree that i was insulting above

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