Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2015-08-09

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

All times shown according to UTC.

Time Nick Message
14:18 mohawk 2.012_01 has been out since the 1st - 8 days
14:18 mohawk 81 passes, 4 fails
14:18 mohawk chm, time to release .013
14:22 mohawk the 4 fails are all in opengl.t
14:23 mohawk that does not belong in the PDL distro but in a separate distro
14:23 mohawk it is time to release 013 and then make the main effort the splitting
14:52 mohawk now i'm going to quickly switch our CI from legacy to container-based
14:52 mohawk time estimated: 6 minutes
15:05 sivoais mohawk: cool
15:05 sivoais umm, yeah, I'll see if I can reach out to chm and get this going
15:07 mohawk fingers crossed
15:08 mohawk this container thing may cause a problem with conditionally installing gfortran
15:08 mohawk i'll disable that for first cut and see what happens
15:09 sivoais yeah, that was my concern
15:10 sivoais we might need to dig into the detection machinery to find an alternative
15:11 mohawk or use the build & cache the migration page mentions
15:12 mohawk then it would just be conditional setting of PATH
15:12 mohawk i'm trying to find what to add to find these packages: pgplot5 libvpx-dev libxpm-dev
15:14 mohawk sivoais, they're not on the whitelist so they ain't getting in
15:14 mohawk ...for now
15:15 sivoais can we request them?
15:15 mohawk we surely can
15:15 mohawk they may have been req'ed already
15:16 mohawk care to have a look? https://github.com/travis-ci/apt-package-whitelist/issues
15:16 mohawk i'm just about to PR the travis changes
15:17 opkick [pdl] mohawk2 created containerci (+1 new commit): http://git.io/v3IZz
15:17 opkick pdl/containerci e6aba6c Ed J: Use container CI
15:17 sivoais reading the instructions
15:17 mohawk awesome
15:19 opkick [pdl] mohawk2 opened pull request #133: Use container CI (master...containerci) http://git.io/v3IZA
15:22 mohawk ruh roh, that failed :-)
15:25 opkick [pdl] mohawk2 force-pushed containerci from e6aba6c to 9cf3b50: http://git.io/v3IcD
15:25 opkick pdl/containerci 9cf3b50 Ed J: Use container CI
15:25 chm joined #pdl
15:26 mohawk sivoais, please feel free to kill the appveyor builds for these PRs
15:26 mohawk i don't know if there's any cost or jobs/day limit?
15:26 mohawk chm, hi
15:26 chm mohawk: the opengl.t fails are tester configuration problems, so not an issue
15:26 chm mohawk: I would like to fix the massive number of solaris fails for 2.013
15:27 chm The problem seems to be some type we are defining one way and solaris another...
15:27 mohawk chm, i'm not seeing that on cpantesters as shown on metacpan, could you link me?
15:27 chm ... in quiet_nan.c
15:28 chm http://www.cpantesters.org/cpan/report/9cef1120-3d7d-11e5-a85a-e28fcaadd3a7
15:28 chm With that fixed, then the dev release is essentially at 100%-ish
15:29 mohawk seems like that line just needs the right #ifdef, then?
15:29 mohawk or really #ifndef
15:29 mohawk chm, i'm gathering you've been busy recently?
15:29 * sivoais looks
15:30 chm That would be my guess.
15:30 sivoais there already is an ifdef
15:30 sivoais in Basic/Math/mconf.h line 85
15:31 mohawk chm, what's your view on giving sivoais co-maint on PDL?
15:43 sivoais it looks like mconf.h came from the Cephes library
15:45 sivoais but it seems to have been modified quite a bit from the one in <http://www.netlib.org/cephes/eval.tgz>
15:45 mohawk sivoais, why are we including libvpx-dev?
15:46 mohawk also, libxpm-dev? i know about pgplot5 ;-)
15:47 sivoais mohawk: I got them from the apt-get build-dep pdl
15:48 sivoais I know pgplot5 is for plotting with PDL::Graphics::PGPLOT
15:48 mohawk as do i
15:48 sivoais I suppose the other two are subdeps of something like libgd
15:48 mohawk sivoais, could you have a look through https://travis-ci.org/PDLPorters/pdl/jobs/74802910 and see what we're not building/testing?
15:49 mohawk i know about fortran and pgplot, can you find others?
15:49 mohawk i'm going to have a go at a gfortran workaround
15:54 sivoais mohawk: opengl is not enabled, but that makes sense here
15:55 sivoais PDL::Slatec doesn't compile , but that's due to not having gfortran
15:55 mohawk so opengl is because of no xpm / vpx?
15:57 sivoais mohawk: no, it's probably because it doesn't have an X display
15:58 sivoais I think libxpm and libvpx are coming from something else, but I need to check the dep graph
15:59 sivoais mohawk: I found the source of those lib{xpm,vpx}-dev packages
15:59 sivoais it is libgd3 <https://packages.debian.org/jessie/libgd3>
16:01 sivoais "  The  long  parameters n to quiet_nan(n) and signaling_nan(n)
16:01 sivoais are presently unused but are  reserved  for  future  use  to
16:01 sivoais specify the significand of the returned NaN."
16:01 sivoais >_>
16:02 mohawk sivoais, is there a way to test the relevant opengl stuff without the 2 libs, maybe an alternate package?
16:02 mohawk chm's gone all quiet
16:03 sivoais mohawk: opengl is separate from those libs. We were never testing it.
16:03 mohawk ok
16:03 sivoais we can test it if we install Xvfb
16:03 sivoais but that is for another time :-)
16:03 mohawk aha
16:03 mohawk i want to get to testing all the PDL stuff with packages available in https://github.com/travis-ci/apt-package-whitelist/blob/master/ubuntu-precise
16:04 sivoais I'll add an issue for that when we split off TriD
16:04 mohawk ok
16:04 sivoais still working out the quiet_nan issue
16:04 mohawk no problem
16:05 mohawk i'm going to install gfortran, but then optionally insert in ~/bin a gfortan that will fail
16:05 mohawk (which will obviously be first in $PATH)
16:07 sivoais seems reasonable but hacky :-P
16:07 mohawk your point being? ;-)
16:07 sivoais no point, just an observation :-D
16:14 sivoais weird, so ... where is MY_QUIET_NAN even being set
16:17 mohawk looks like it's not
16:17 mohawk what does git blame say?
16:20 sivoais looking at the log on mconf.h, I see that the change was made ages ago O_o 1999
16:21 mohawk 11 jan 1999 by KGB
16:22 mohawk BUT
16:22 mohawk since that's ifdef something we think's not used...?
16:22 sivoais looking some more
16:23 sivoais for the callers
16:23 mohawk https://www.google.co.uk/?gws_rd=ssl#q=MY_QUIET_NAN&amp;nfpr=1
16:23 mohawk pdl 2.4.3
16:23 sivoais yeah, it's been in there for a bit
16:23 mohawk yes
16:24 sivoais don't use MY_* as the prefix! Use PDL_*  :-)
16:24 mohawk i agree
16:24 mohawk so are we saying it's also used in sun's math stuff?
16:24 sivoais stil looking
16:26 mohawk i want to make a vagrant vm farm that can test pdl (and other things) across a range of OSes
16:26 mohawk illumos is available linked on vagrantbox.es
16:26 mohawk that would cover the solaris issue
16:26 sivoais true, true
16:27 mohawk it also has modern.ie windows boxes
16:28 sivoais but there is something that is troubling me in that the functions that use quiet_nan are ones that should be in C89
16:28 sivoais asinh()
16:28 sivoais actually, just acosh(), atanh()
16:29 sivoais so why exactly is this code even there?
16:29 mohawk that's a darn good question
16:29 mohawk we could ask KGB?
16:30 sivoais right
16:31 mohawk great
16:31 sivoais $ git log Basic/Math # currently looking through this
16:31 mohawk it seems to me most probable that this solaris breakage is not new
16:31 mohawk chm, is that correct?
16:33 sivoais since nobody has touched that part in a while, perhaps
16:33 sivoais I'll write an e-mail to the list and KGB
16:33 mohawk so there's no possible reason to not release .013 right now
16:36 mohawk seems to me the only reasonable policy is "it's no worse - release it"
16:36 mohawk because without that, it never gets better at all
16:37 mohawk viz 2.008
16:38 opkick [pdl] mohawk2 force-pushed containerci from 9cf3b50 to 953cb78: http://git.io/v3IcD
16:38 opkick pdl/containerci 93896d6 Ed J: Sort PDL::Config dump output
16:38 opkick pdl/containerci 4ec45bb Ed J: Tests non-verbose unless TEST_VERBOSE == 1
16:38 opkick pdl/containerci 953cb78 Ed J: Use container CI
16:40 sivoais right, I have a fix for this, but I don't feel justified in putting the fix in a) not having tested under Solaris/Illumos, b) not understanding why the callers are there
16:42 mohawk http://www.vagrantbox.es/
16:42 mohawk solaris 11.1 box right there
16:43 sivoais I looked at past times that PDL was tested on Solaris
16:44 sivoais back then, the compiler was probably different
16:44 mohawk yeah
16:44 mohawk by all means, let's fix it
16:44 mohawk in 013_01
16:44 sivoais what I see in the output that is a little surprising is /opt/solarisstudio12.3/bin/cc
16:45 mohawk yeah
16:46 sivoais I have no tuits right now :-/ I have to get back to $WORK
16:46 mohawk when i said chm had gone all quiet, i missed that he ping timeout-ed like an hour ago
16:50 opkick [pdl] zmughal commented on issue #133: Looking good! And it seems to run each build about 25% faster individually.... http://git.io/v3I98
16:52 opkick [pdl] mohawk2 commented on issue #133: Yes it does. In the meantime, I'm going to have it built and installed per the migration instructions. http://git.io/v3I9K
16:53 opkick [pdl] zmughal opened issue #134: Solaris `quiet_nan()` takes a `long` argument http://git.io/v3I99
16:55 opkick [pdl] zmughal commented on issue #131: See <https://github.com/PDLPorters/pdl/pull/133>. http://git.io/v3IHl
16:59 opkick [pdl] zmughal commented on issue #133: OK, made a request for pgplot5 at <https://github.com/travis-ci/apt-package-whitelist/issues/759>. Do we want the others? http://git.io/v3IQ3
17:30 opkick [pdl] mohawk2 commented on issue #133: @zmughal I'd say yes we do. http://git.io/v3IxN
17:33 mohawk the gfortran-disabling seems effective, which is good
17:33 mohawk t/linfit.t runs ok in https://travis-ci.org/PDLPorters/pdl/jobs/74809447 but skipped on https://travis-ci.org/PDLPorters/pdl/jobs/74809452
17:34 mohawk a previous build took a report 2h15, this one took 1h29
17:34 mohawk admittedly, still not running pgplot
17:37 opkick [pdl] mohawk2 pushed 1 new commit to containerci: http://git.io/v3IjI
17:37 opkick pdl/containerci 7e18d69 Ed J: Build pgplot 5.3.0 in cached dir, use
17:39 mohawk that went a bit wrong
17:39 opkick [pdl] zmughal commented on issue #133: OK, I linked back from the rest of the deps here:... http://git.io/v3IjF
17:40 opkick [pdl] mohawk2 comment on issue #133: Great! http://git.io/v3LeT
17:44 opkick [pdl] mohawk2 force-pushed containerci from 7e18d69 to ef82025: http://git.io/v3IcD
17:44 opkick pdl/containerci ef82025 Ed J: Build pgplot 5.3.0 in cached dir, use
17:56 opkick [pdl] mohawk2 force-pushed containerci from ef82025 to 42f974c: http://git.io/v3IcD
17:56 opkick pdl/containerci 42f974c Ed J: Build pgplot 5.3.0 in cached dir, use
18:02 opkick [pdl] mohawk2 force-pushed containerci from 42f974c to 63f7aef: http://git.io/v3IcD
18:02 opkick pdl/containerci 63f7aef Ed J: Build pgplot 5.3.0 in cached dir, use
18:05 opkick [pdl] mohawk2 force-pushed containerci from 63f7aef to 6f47647: http://git.io/v3IcD
18:05 opkick pdl/containerci 6f47647 Ed J: Build pgplot 5.3.0 in cached dir, use
18:12 opkick [pdl] mohawk2 force-pushed containerci from 6f47647 to 26178c6: http://git.io/v3IcD
18:12 opkick pdl/containerci 26178c6 Ed J: Build pgplot 5.3.0 in cached dir, use
18:20 opkick [pdl] mohawk2 force-pushed containerci from 26178c6 to c48e4ca: http://git.io/v3IcD
18:20 opkick pdl/containerci c48e4ca Ed J: Build pgplot 5.3.0 in cached dir, use
18:24 opkick [pdl] mohawk2 force-pushed containerci from c48e4ca to 529bbc9: http://git.io/v3IcD
18:24 opkick pdl/containerci 529bbc9 Ed J: Build pgplot 5.3.0 in cached dir, use
18:28 opkick [pdl] mohawk2 force-pushed containerci from 529bbc9 to d130977: http://git.io/v3IcD
18:28 opkick pdl/containerci d130977 Ed J: Build pgplot 5.3.0 in cached dir, use
18:35 opkick [pdl] mohawk2 force-pushed containerci from d130977 to dde4027: http://git.io/v3IcD
18:35 opkick pdl/containerci dde4027 Ed J: Build pgplot 5.3.0 in cached dir, use
18:46 chm joined #pdl
18:47 chm sivoais, mohawk: sorry about that, cable internet down---again.
18:47 chm sivoais: re solaris, I think that PDL hasn't been tested there in a while
18:48 chm maybe because of a low memory VM setup.  It used to pass there as well.
18:50 sivoais ouch.
18:50 sivoais chm: why don't we get v2.013 out now and work on figuring out the proper quiet_nan() approach in the next release?
18:51 sivoais that way Marek can get a fix and one platform+compiler doesn't hold up all the others
18:52 sivoais I think that failure might be compiler specific, but I need to ask KGB about the define that is there right now
18:52 sivoais because the define is never set
18:52 opkick [pdl] mohawk2 force-pushed containerci from dde4027 to e9a764e: http://git.io/v3IcD
18:52 opkick pdl/containerci e9a764e Ed J: Build pgplot 5.3.0 in cached dir, use
18:58 opkick [pdl] mohawk2 force-pushed containerci from e9a764e to d1510a6: http://git.io/v3IcD
18:58 opkick pdl/containerci d1510a6 Ed J: Build pgplot 5.3.0 in cached dir, use
18:58 mohawk chm, glad you're back
18:59 mohawk chm, i asked earlier: chm, what's your view on giving sivoais co-maint on PDL?
19:07 opkick [pdl] mohawk2 force-pushed containerci from d1510a6 to 1aba12a: http://git.io/v3IcD
19:07 opkick pdl/containerci 1aba12a Ed J: Build pgplot 5.3.0 in cached dir, use
19:14 opkick [pdl] mohawk2 force-pushed containerci from 1aba12a to 294f96e: http://git.io/v3IcD
19:14 opkick pdl/containerci 294f96e Ed J: Build pgplot 5.3.0 in cached dir, use
19:24 chm mohawk, sivoais: I'm testing a fix for the quiet_nan.c problem, change line 3 from double quiet_nan(void) to double quiet_nan(NANARG)
19:25 chm So far all tests are passing... do we go through a branch merge cycle or just apply to master and test for release?
19:25 opkick [pdl] mohawk2 force-pushed containerci from 294f96e to 048ddab: http://git.io/v3IcD
19:25 opkick pdl/containerci dc3595d Ed J: Latest stable is now 5.22
19:25 opkick pdl/containerci 08878d7 Ed J: Save CI time by not testing 5.12-5.20
19:25 opkick pdl/containerci 1b00160 Ed J: Use container CI
19:25 mohawk chm, please put on a branch for a quick look over
19:26 mohawk i'm about to merge a CI update, so i can do both at once and save some CI time
19:26 chm ok, will do.  I haven't heard back from Marek yet re the fix....ping'd him again...
19:26 mohawk chm, i'd love to hear your thoughts on the co-maint question above
19:30 opkick [pdl] mohawk2 force-pushed containerci from 048ddab to 6fee91d: http://git.io/v3IcD
19:30 opkick pdl/containerci bd41648 Ed J: Save CI time by not testing 5.12-5.18...
19:30 opkick pdl/containerci fe08e9e Ed J: Use container CI
19:30 opkick pdl/containerci 6fee91d Ed J: Build pgplot 5.3.0 in cached dir, use
19:35 opkick [pdl] wchristian pushed 1 new commit to master: http://git.io/v3LoT
19:35 opkick pdl/master 59d2a4c Chris Marshall: Update MANIFEST with new test file
19:36 mohawk ah ha, that looks like sivoais managed to fail to add his test to manifest? ;-)
19:36 sivoais eep!
19:36 sivoais that should have been caught in the code review! :-P
19:37 mohawk which one's worse? ;-)
19:37 sivoais but I have no excuse either way! :-P
19:37 mohawk i see
19:37 mohawk just trying to drag me down with you, eh
19:37 mohawk tsk
19:37 sivoais hahaha!
19:37 mohawk looks like PR #133 is behaving now
19:38 mohawk i'm going to stop the master CI build
19:39 mohawk what would be super-handy right now is if we had IM with chm
19:39 mohawk sivoais, please review #133
19:39 mohawk i've been bold and removed 5.12-18 since we've never caught problems there
19:40 mohawk the container CI seems to be 20% or so quicker, which is nice
19:40 mohawk and the 3 hours i spent getting round the lack of pgplot5 package was worth it, i'm sure ;-)
19:50 opkick [pdl] wchristian created solaris-quiet-nan-conflict (+1 new commit): http://git.io/v3LXP
19:50 opkick pdl/solaris-quiet-nan-conflict e9636f3 Chris Marshall: Fix to solaris build problem re quiet_nan...
19:51 opkick [pdl] devel-chm opened pull request #135: Fix to solaris build problem re quiet_nan (master...solaris-quiet-nan-conflict) http://git.io/v3LXx
19:51 opkick [pdl] zmughal commented on issue #133: Wow, that was quite a bit of work on pgplot5! It's looking good.... http://git.io/v3L1O
19:53 chm mohawk, sivoais: please review #135
19:53 sivoais I'm taking a look
19:54 sivoais chm: did you get a chance to test on a Solaris system?
19:54 sivoais mohawk: ooops, looks like #133 failed on the pgplot test <https://travis-ci.org/PDLPorters/pdl/jobs/74823557>
19:56 chm sivoais: I don't have a solaris system.  If there had been solaris testers, we would have caught this already.
19:56 sivoais yep. mohawk says there is a vagrant box for solaris . I'll try it there now
19:57 sivoais $ vagrant  box add solaris-11.2 http://www.benden.us/vagrant/solaris-11.2.box
19:58 mohawk sivoais, i'll look at 135
19:59 mohawk oops, 133 failed on disable fortran
19:59 mohawk sivoais, i'll dig into that
19:59 mohawk please could you review 135 and hopefully merge/release as 013?
20:00 sivoais I'm going to attempt to test on Solaris. Hopefully the sunmath library is still around. I've never used Solaris for development
20:01 mohawk sivoais, PLEASE capture all you do so we can record in PDLP/devops
20:03 sivoais mohawk: good idea
20:03 sivoais I'll record it into an issue for now until it can be automated
20:04 mohawk great
20:04 mohawk chm, are you happy that if it works on sivoais's solaris vm, we can release as .013?
20:05 opkick [devops] zmughal opened issue #1: Solaris testing procedure http://git.io/v3L9v
20:06 chm mohawk: I was more concerned that non-solaris platforms don't break.  As is, solaris is broken either way if the fix doesn't
20:06 sivoais true, true
20:07 mohawk well, let's let the CI do its thang
20:08 opkick [pdl] mohawk2 commented on issue #135: Looks good to me, so if it works on Solaris, merge away! (and release) http://git.io/v3L9X
20:08 mohawk chm, while we wait for sivoais, please could you actually state your position on the co-maint question
20:08 mohawk that's the third time of asking
20:08 chm I'd like to give Marek until tomorrow to respond and then release 2.013 tomorrow evening.
20:09 mohawk that sounds good
20:09 chm Sorry, trying to get work done first.
20:09 chm I thought there was an email to the list requesting co-maint which would be the logical place to reply
20:09 chm Can't find it.
20:09 opkick [pdl] zmughal commented on issue #134: See <https://github.com/PDLPorters/pdl/pull/135> http://git.io/v3L9p
20:14 opkick [pdl] mohawk2 force-pushed containerci from 6fee91d to 7489420: http://git.io/v3IcD
20:14 opkick pdl/containerci f84b7c1 Ed J: No need to ls -lR in CI
20:14 opkick pdl/containerci 4a873f9 Ed J: Temporary disable of all but 2 jobs
20:14 opkick pdl/containerci 1d976ac Ed J: Use container CI
20:15 chm mohawk: found it, will reply there...
20:15 mohawk i'm foreseeing that the #135 will pass the CI
20:15 mohawk if it does, let's merge it
20:16 mohawk actually, better: let's pass merge control of it to sivoais to merge if he's also happy it passes on solaris
20:16 mohawk in a perfect world, you could also try without the #135 change to check you've found the breakage
20:17 sivoais GitHub has this nice little feature where you can assign issues/PRs to a person
20:17 sivoais which makes it easy to delegate in large projects :-)
20:19 chm sivoais: is that the Assignee?
20:21 sivoais chm: yes! :-)
20:21 mohawk not a substitute for communication in this kind of tricky coordination situation
20:22 mohawk but a useful tool, for sure
20:24 sivoais true. But if you're facing a list of issues, then knowing what you have to do can be useful. We're probably not there yet. :-) I've seen it used by other projects.
20:26 mohawk let's crack the chunk size and continuous delivery problems first
20:26 mohawk ;-)
20:29 chm sivoais, mohawk:  Please feel free to update any of the release notes, changes, known problems for 2.013
20:32 mohawk ok
20:33 mohawk i don't see the need except maybe to say "CI improvements" - sivoais, i'll leave that with you
20:34 mohawk and obviously "fixed problem in Solaris" ;-)
20:34 sivoais ok, will do
20:35 mohawk hurry up travis on chm's PR, i want you to run mine!
20:36 mohawk i am so looking forward to splitting this thing up
20:36 mohawk CI will take no time at all
20:36 mohawk dev will become actually enjoyable
20:37 mohawk #135 is DONE
20:38 mohawk wait, it's not
20:38 mohawk but it starts the next job if there are spare CPUs, which is cool
20:39 sivoais ugh. I need to create an Oracle account
20:39 mohawk ha ha, oh dear
20:39 sivoais just to get at this SolarisStudio 12.3
20:39 mohawk hope this is all getting recorded
20:39 sivoais yes, recording as I go :-)
20:39 mohawk while i'm waiting, i'm going to make PIG use containers too
20:41 opkick [pdl-io-gd] mohawk2 deleted github-sprucing at ea69aca: http://git.io/v3LAr
20:43 mohawk chm, may i ask why you won't release PDL::IO::GD out of its new location?
20:44 mohawk and/or give co-maint on it to me and sivoais?
20:46 opkick [pdl-io-gd] mohawk2 created containerci (+1 new commit): http://git.io/v3LxD
20:46 opkick pdl-io-gd/containerci 1edf783 Ed J: Use container Travis build
20:46 opkick [pdl-io-gd] mohawk2 opened pull request #5: Use container Travis build (master...containerci) http://git.io/v3Lx9
20:46 chm mohawk: don't know what you mean here.
20:46 mohawk chm, sivoais and i have extracted PDL::IO::GD into a separate repo on github
20:46 mohawk its version history got extracted too
20:47 mohawk https://github.com/PDLPorters/pdl-io-gd
20:47 mohawk it's ready to be released
20:48 sivoais I might have to create a set of PDL-specific Vagrant boxes :-)
20:49 sivoais because having to install SolarisStudio can't be automated
20:50 chm sivoais: It might be easier to just push to master as long as all other systems/platforms pass with the change
20:50 mohawk chm, https://sourceforge.net/p/pdl/mailman/message/34186603/
20:50 mohawk from june
20:50 mohawk sivoais, that'll do ;-)
20:50 chm mohawk, sivoais: Thanks for the pointer.  The main sticking point for much of this is how to not break PDL for users.
20:51 mohawk chm, respectfully, that's what dev releases are for
20:51 chm One idea I had was to make a PDL bundle that users could use to get a specific, working PDL configuration
20:51 chm (in the midst of all the coming changes)
20:51 mohawk that's what i've said "PDL" will continue to be
20:52 mohawk you did actually read sivoais's and my emails on the subject, right?
20:52 chm that's what I've sed I'm trying to ensure
20:52 mohawk ok
20:52 mohawk it seems as though you're saying you think sivoais and i aren't also aiming for that
20:53 chm The new development model doesn't stability since master==CPAN official release.
20:53 sivoais PDL will install everything it already does. But will do it by depending on PDL::Core + PDL::* as separate dists. So there will be no discontinuity for people that want PDL
20:54 mohawk chm, your statement there is incorrect
20:54 chm mohawk: which statement?
20:55 sivoais I'm not sure I understand the statement about stability
20:55 mohawk "<chm> The new development model doesn't stability since master==CPAN official release"
20:56 chm My understanding is that the current rebase model forces master to be the CPAN official release.
20:56 mohawk i have no idea where you got that from
20:57 mohawk and separately, since it does have a distant relationship with reality, what problem are you perceiving that to create?
20:58 chm That is my understanding.  And the idea came from you and sivoais and the various release discussions and issues
20:58 mohawk if it were true, what problem would that create?
20:58 sivoais I need clarification, because I'm lost at where the problem is
20:59 chm mohawk, sivoais:  I'll try to write up things more coherently for the list discussion.
20:59 chm At the moment, dinner calls, o/
20:59 sivoais later then o/
21:05 opkick [pdl-io-gd] mohawk2 closed pull request #5: Use container Travis build (master...containerci) http://git.io/v3Lx9
21:06 opkick [pdl-io-gd] mohawk2 deleted containerci at 1edf783: http://git.io/v3tfJ
21:08 mohawk sivoais, where are we with the solaris task?
21:08 mohawk (btw, you know it's possible to freeze a vm as a box for sharing and whatnot?)
21:09 mohawk i can see us making a dir in PDLP/devops with our boxes
21:09 mohawk "pdl::whatever is breaking on cygwin"
21:09 mohawk "hold on, i'll spin up my pdl-cygwin box"
21:11 sivoais it may be too large of a file. GitHub has a 100 MB limit
21:11 sivoais but I can host on my server
21:13 mohawk aha, good point
21:13 mohawk in any case, we should make these things
21:14 mohawk infra as code!
21:16 opkick [pdl] mohawk2 force-pushed containerci from 7489420 to c15a413: http://git.io/v3IcD
21:16 opkick pdl/containerci f9c57f3 Ed J: Temporary disable of all but 2 jobs
21:16 opkick pdl/containerci e00b047 Ed J: Use container CI
21:16 opkick pdl/containerci c15a413 Ed J: Build pgplot 5.3.0 in cached dir, use
21:23 sivoais shoot
21:23 sivoais so
21:23 sivoais Solaris comes with Perl, but not CPAN
21:23 sivoais cpan(1)
21:24 sivoais ah, perl -MCPAN -eshell
21:31 mohawk App::cpanminus has a url you can pipe into a shell
21:32 mohawk still recording this stuff? :-)
21:33 sivoais yep, yep
21:40 sivoais mohawk: I'm going to have to step away as well. I'll report my results on the #135 PR when this finishes installing all the deps
21:41 mohawk great
21:51 opkick [pdl] mohawk2 force-pushed containerci from c15a413 to f72e96d: http://git.io/v3IcD
21:51 opkick pdl/containerci 4f5b9fd Ed J: Temporary disable of all but 2 jobs
21:51 opkick pdl/containerci 82ac33a Ed J: Use container CI
21:51 opkick pdl/containerci 36476fa Ed J: Build pgplot 5.3.0 in cached dir, use
21:59 opkick [pdl] mohawk2 force-pushed containerci from f72e96d to 4217c98: http://git.io/v3IcD
21:59 opkick pdl/containerci 7047293 Ed J: Add "env" command to CI
21:59 opkick pdl/containerci 028dfd4 Ed J: Temporary disable of all but 2 jobs
21:59 opkick pdl/containerci 033970b Ed J: Use container CI
22:13 opkick [pdl] mohawk2 force-pushed containerci from 4217c98 to 81f48ff: http://git.io/v3IcD
22:13 opkick pdl/containerci 9d15504 Ed J: Add "env" command to CI
22:13 opkick pdl/containerci 8a3eba4 Ed J: No need to test EU::PXS
22:13 opkick pdl/containerci 7e8fe35 Ed J: Temporary disable of all but 2 jobs
22:28 opkick [pdl] mohawk2 force-pushed containerci from 81f48ff to 2f97f5f: http://git.io/v3IcD
22:28 opkick pdl/containerci 2f97f5f Ed J: p

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