Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2016-06-04

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

All times shown according to UTC.

Time Nick Message
07:38 ribasushi joined #pdl
10:07 plicease joined #pdl
10:23 drrho joined #pdl
13:12 mohawk i would ask: why would we need to have a mailing list, and source control, from the same provider?
13:58 mohawk more elegantly put: do those two things actually need to be from same provider?
14:04 chm joined #pdl
14:05 chm jberger, mohawk, vicash, pdurbin: I see there was some git-hosting discussion recently.
14:06 chm mohawk: sorry about the surprise with merging into the github vs sf repo and the resulting overwrite
14:07 chm mohawk: I ended up recovering manually then saw there was a "recover this branch" option on github
14:08 chm Regarding the sf/github host bikeshedding, here is my understanding and approach:
14:09 chm PDL-2.x dist is in maintenance mode with git repo on sf.net etc.
14:09 chm New PDL development, PDL Next Gen stuff is on github.
14:10 chm Eventually, I could see reversing the auto-pull and have sf.net repo only image the github one[s]
14:11 chm In general, folks can use what works best for them independent of my approach.
14:11 chm It would be nice if more energy could be devoted to PDL development rather than development hosting.  :-)
14:12 jberger chm: it hasn't been mentioned in quite some time and indeed only was because the split had caused a problem
14:12 jberger I don't see it as a major energy sink
14:13 jberger I can tell that the issue frustrates you though
14:13 jberger not sure what to say about that
14:14 jberger I've said my piece a long time ago and I really don't even see myself having a vote anymore since I haven't used PDL in years
14:14 jberger (owing to not being in the scientific field anymore)
14:14 chm mohawk, jberger: I guess it is frustrating when the PDL progress I see on #pdl is hosting discussions
14:14 jberger but I still have a soft place in my heart for it
14:16 chm FWIW, I've been working on some technologies towards my ideas for PDL::NextGen
14:16 chm I'm working on getting tcc to run on cygwin.  I compiles but no luck with the includes
14:17 chm I've ported COS to cygwin (https://github.com/CObjectSystem/COS)
14:18 chm COS is a C Object system that could provide a C FFI OO capability for the PDL::NG computation core.
14:19 chm jberger: I think a next step might be to implement a stub PDLNG in Moo with the goal of making PDL-2.x a role to apply
14:19 chm That might allow us to use existing PDL-2.x and implement test PDLNG alongside.
14:19 chm jberger: I believe you took a poke as that a bit previously.
14:20 chm Now that PDL-2.016 is out, I'm focusing on OpenGL support for modern API
14:20 jberger if I did, it didn't go very far :s
14:21 chm I think it was just about enabling a Moo PDL with the PDL attribute.
14:21 mohawk chm, nothing for you to apologise for! i looked in my command history and saw where i went wrong
14:23 chm It might also be possible with a PDL-2.x role to fold in the perl6 support a la sisyphus on the mailing list
14:23 mohawk i think the way to get us off SF is for me to finish up PDLA such that people can use it instead of PDL
14:24 chm mohawk: PDL-2.x is in maintenance mode on SF.  If we have new development, I think github is the place.
14:24 mohawk chm, and good work in recovering from my minor snafu ;-) the work was still in my local repo (on master) so it wasn't truly lost, in any case
14:25 mohawk chm, i'm not sure what you mean by that?
14:25 chm mohawk: ...and yes, I don't have the time to remerge the 64bit into PDLA but moving the split along is desired
14:25 mohawk since the current system, using Mithaldu's excellent mirror, goes from SF -> GH, not the other way round
14:25 mohawk if you want to switch that, it's straightforward
14:26 mohawk chm, i'll take responsibility for such merging
14:26 chm mohawk: with the OpenGL work I will see about pulling the TriD out into its own package.
14:26 mohawk there will come a moment when the direction of travel changes, and you're changing PDLA and we back-port to PDL :-)
14:26 mohawk chm, fantastic
14:27 mohawk chm, if you're saying you'd like to reverse the flow for PDL-2 between SF and GH, we'd need a transition process: an email where it's announced, then we ask no-one to push when we flip the switch, then flip switch, then all is good
14:27 mohawk and we'd need to coordinate with Mithaldu
14:28 mohawk it's pretty straightforward, we'd just need to make the decision and then action it :-)
14:28 chm mohawk: whoa!  The reverse doesn't even make sense until there is a viable PDLA out.
14:29 chm As mentioned, I plan to keep PDL-2.x in maintenance mode at sf.net.
14:29 chm The only issue is if someone doing "development" on github and wants to merge into PDL-2.x maintenance master.
14:30 chm This is a non-issue once there is live PDLA and PDLNG coming directly from github.
14:30 chm And even in that case, what might make sense is to have a github -> sf repo mirror for PDLA, PDLNG
14:32 chm At that point PDLNG or PDLA would be the PDL3 master and the PDL-2.x should be superceded by the new tech
15:22 pdurbin joined #pdl
15:34 drrho joined #pdl
15:52 mohawk chm, i don't think i understand what you mean by maintenance mode, then?
17:15 mohawk chm, i think the idea of two repositories (GH and SF) being out of sync is a terrible one, i assume you're not advocating that?
19:34 mohawk just by-the-by, re the earlier mention of FFI, this is a perl 5 module: https://metacpan.org/release/NativeCall

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