Perl 6 - the future is here, just unevenly distributed

IRC log for #pdl, 2015-02-28

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

All times shown according to UTC.

Time Nick Message
01:31 mohawk so i'm going to put into practice what i said about the topic branch stuff
01:31 mohawk i mentioned i'd rebased LLDF onto current master
01:32 mohawk about halfway through that, the tests all pass
01:32 mohawk i am not going to comb and comb to turn all the commits into coherent works of beauty
01:32 mohawk what i will do is make a branch, lldfp1, and ask for someone to review it
01:33 mohawk once someone has ok-ed it, i will then push it
01:33 mohawk the second half, with PDL_Generic, i will then review more closely
01:33 mohawk there may be some squashing / dicing of commits
01:34 mohawk especially since it doesn't currently pass the tests!
01:35 mohawk (it = the last X commits on the rest of LLDF)
01:48 opkick [pdl] wchristian created lldfp1 (+23 new commits): http://git.io/xYhZ
01:48 opkick pdl/lldfp1 e0a7274 Craig DeForest: remove references to progenitors and mutators.  Update core version number.
01:48 opkick pdl/lldfp1 30b6e35 Craig DeForest: some typos ("propogate"->"propagate" throughout; some typos in CallExt documentation)
01:48 opkick pdl/lldfp1 0203a46 Craig DeForest: more cleanup of pdl.h.PL. Also, cut out foomethod() nonsense since it is not...
02:19 mohawk email sent
02:19 mohawk review sought
02:20 mohawk Mithaldu, feel like taking a look?
02:20 * Mithaldu goes look for it
02:23 Mithaldu man, juicing a game properly is hard!
02:25 Mithaldu the email looks good so far
02:25 Mithaldu though maybe you should wait for chm to voice off for such a major thing
02:26 mohawk it's not that major
02:27 mohawk although in this case, it is clearly a good idea to have chm or maybe craig do the review since it's their stuff, separated by time so they have perspective
02:27 mohawk anyway, not really a rush
03:28 opkick [pdl] wchristian created tempbranch (+1 new commit): http://git.io/xOGl
03:28 opkick pdl/tempbranch 27c804e Ed J: test commit
03:46 mohawk a thing i found when looking at the git post-receive hook - there's a pdl-commit mailing list automatically setup
03:46 mohawk https://sourceforge.net/p/pdl/mailman/pdl-commit/
04:06 sivoais yeah, I'm alread subscribed
04:06 sivoais already
04:14 mohawk nice
05:11 mohawk for the record, this is how to login to sf.net and look at the hooks for pdl:
05:11 mohawk ssh -t YOURUSERNAME,pdl@shell.sourceforge.net create
05:11 mohawk git remote add ghpdlp git@github.com:PDLPorters/pdl.git
05:11 mohawk cd /home/git/p/pdl/code.git/hooks
05:12 mohawk the git remote line isn't necessary, of course - that's just the note i kept in case it wasn't persistent, but it is
05:12 jberger github has a page for that :)
05:13 jberger geee isn't github swell
05:13 mohawk for what?
05:13 jberger hooks
05:13 mohawk github is quite good, yes
05:13 mohawk it does
05:13 mohawk though it only allows triggering of a web request
05:13 mohawk sf for once allows very slightly more
05:13 mohawk albeit slowly and badly
05:14 jberger I'm not sure it's apples to apples, I just like to be snarky on this topic
05:15 mohawk ha ha
05:15 mohawk really?
05:15 mohawk that was not clear
05:20 jberger come back and I will taunt you a second time!
05:21 mohawk jberger++
06:19 sivoais would you all be ok with Travis-CI build notifications in here?
06:19 sivoais so when PDLPorters/pdl gets pushed to, then notifications come in here
06:20 mohawk sounds good to me
06:20 mohawk why not enable it, then turn it off if there are dramas
06:23 sivoais ok, I'm going to test on my own first
06:23 sivoais because it appears that it will mean that all forks that use Travis-CI will also send notifications
06:24 mohawk i think you're right
06:24 sivoais though having many forks with activity would be a good problem to have! :-)
06:25 mohawk indeedily ;-)
06:28 sivoais mohawk: so... I've been seeing build failures on Travis-CI for the past couple days
06:28 sivoais but only on Perl v5.12
06:29 sivoais I figured out why. It's a particular version of ExtUtils::ParseXS
06:29 mohawk ok
06:29 mohawk so we need to specify a suitable version for it?
06:29 sivoais look at <https://travis-ci.org/PDLP​orters/pdl/builds/52259205>
06:29 sivoais it seems so
06:30 mohawk i believe you
06:30 mohawk you have push access
06:30 mohawk try it on a branch :-)
06:30 sivoais I haven't tried to nail the *exact* version, but I know the latest ParseXS works
06:30 mohawk specify the one that comes with 5.14
06:31 sivoais sounds good
06:32 sivoais and that is v2.2210
06:32 mohawk make it so ;-)
17:28 mohawk sivoais, did you have a chance to look at this?
18:06 sivoais only a little here and there, I'll take a closer look in an hour or so when I take a break
18:28 mohawk ok
18:28 opkick [pdl] mohawk2 created prereqdump (+1 new commit): http://git.io/xGha
18:28 opkick pdl/prereqdump e79d804 Ed J: Add test that dumps installed version of required modules
18:29 mohawk let's see what travis makes of that
18:33 opkick [pdl] wchristian deleted prereqdump at e79d804: http://git.io/xZeT
18:42 mohawk wait, what? who deleted it?
18:50 Mithaldu haha
18:51 Mithaldu */5 * * * * cd /root/pdl.git && git update-mirror
18:52 Mithaldu hmmm
18:52 Mithaldu now where is that macro
18:54 mohawk may i suggest replacing that with just "git push --mirror correctremotename"?
18:55 mohawk look in the config under [alias]
18:56 Mithaldu mohawk: https://github.com/PDLPorters/pdl/issues/6
18:57 mohawk great
18:59 Mithaldu give me a concrete request for change in either of those?
19:00 mohawk just pondering
19:00 Mithaldu ah, ok :)
19:00 mohawk i don't know why that would have deleted
19:00 Mithaldu also, i think a look at git config manual is warranted
19:00 Mithaldu i think it might be the mirror option
19:00 mohawk yes, i think so too
19:00 mohawk BUT
19:01 mohawk could you login there and do "git branch" to see if "prereqdump" is there?
19:09 Mithaldu mohawk: looking
19:10 Mithaldu core-cleanup
19:10 Mithaldu cpan_2.006_release
19:10 Mithaldu cpan_2.007_04
19:10 Mithaldu cpan_2.007_release
19:10 Mithaldu lldfp1
19:10 Mithaldu longlong-double-fix
19:10 Mithaldu * master
19:10 Mithaldu release_2.4.10
19:10 Mithaldu tempbranch
19:10 Mithaldu ----
19:11 mohawk so
19:11 mohawk somehow the git fetch didn't work
19:11 mohawk whoa, hang on
19:12 mohawk prereqdump isn't apparently on the sf site either
19:12 mohawk that's weird, let me login and look
19:12 Mithaldu yeah, i'm fairly sure the mirror bit is the key
19:12 mohawk on the face of it, it IS working and something else has happened
19:13 Mithaldu what happened is you pushed to github
19:13 Mithaldu it fetched from sf and synched that to github
19:13 mohawk AND i pushed to sf
19:13 Mithaldu it doesn't synch both ways
19:13 mohawk almost the same time
19:13 Mithaldu oh that's new
19:14 mohawk hold on
19:14 mohawk i think i pushed NOT to sf
19:14 mohawk so the mistake is mine, i THOUGHT i pushed to gh and sf, but i didn't
19:14 mohawk stand down :-)
19:15 mohawk let me push to sf and we'll wait and see
19:17 Mithaldu \o/
19:18 opkick [pdl] wchristian created prereqdump (+1 new commit): http://git.io/xGha
19:18 opkick pdl/prereqdump e79d804 Ed J: Add test that dumps installed version of required modules
19:20 mohawk there ya go
19:21 Mithaldu first step, always, repro :)
19:21 mohawk however, this exercise has shown something useful, which i'm about to email
19:21 Mithaldu now i'm curious
19:33 mohawk about the remaining git branches
19:33 mohawk two of them have no commits added
19:33 mohawk they're just pseudo-tags
19:34 mohawk you'll see in my email in a sec :-)
19:34 mohawk sent
19:38 Mithaldu oh, yeah, branches that should be tags should be fixed
19:42 mohawk yeah
19:42 mohawk i want buy-in for that though
19:42 mohawk and they actually are tags anyway
19:42 mohawk Mithaldu, to answer your "why not force-push" point:
19:42 mohawk $ git push -f upstream
19:42 mohawk Counting objects: 3, done.
19:42 mohawk Delta compression using up to 2 threads.
19:42 mohawk Compressing objects: 100% (3/3), done.
19:42 mohawk Writing objects: 100% (3/3), 417 bytes | 0 bytes/s, done.
19:42 Mithaldu ?!
19:42 mohawk inorite
19:43 Mithaldu i don't know what you're trying to tell me
19:45 mohawk it's rejecting force pushes
19:45 mohawk which is what a non-fast-forward ref is
19:45 mohawk don't you know anything? ;-)
19:46 Mithaldu call me a hazelnut, i have no idea what you're saying
19:46 Mithaldu all i see up there is a successful forcepush
19:46 Mithaldu probably
19:47 Mithaldu usually i do it with tgit, so i have nice red/green result output
19:53 mohawk you didn't see the words at the bottom: remote: error: denying non-fast-forward refs/heads/prereqdump (you should pull first)
19:53 mohawk also, the exit status was 1 (ie failure)
19:53 mohawk it failed because it is currently configured to fail
19:54 Mithaldu mohawk: that didn't get pasted
19:54 mohawk yes it did
19:54 mohawk right above your ?!
19:54 Mithaldu 15-02-28@20:42:32 (mohawk) Writing objects: 100% (3/3), 417 bytes | 0 bytes/s, done.
19:54 Mithaldu 15-02-28@20:42:52 (Mithaldu) ?!
19:54 mohawk on my screen it did
19:54 Mithaldu well your screen is doody then
19:54 Mithaldu clearly!
19:54 mohawk well that's's fascinating
19:54 mohawk anyway, i'm glad that's clarified
19:55 Mithaldu mohawk: fwiw, the network has mechanisms against "spam"
19:55 mohawk i'm going to reconfigure it (if it's possible) to still reject force-push on master, but allow on others
19:55 Mithaldu so the usefullest way to do bigger paste is have a script that delays each line by a second
19:55 mohawk however, if per-branch isn't possible, i'll allow everywhere
19:55 mohawk maybe
19:55 Mithaldu rejecting force-push on master is clever when you're working with literal apes like me
19:56 mohawk it showed you 5 of my lines in a burst, but silently dropped the last 2
19:56 mohawk remember when you were moaning and complaining that i didn't force-push to master? ;-)
19:56 Mithaldu but when working with decent people, depending on the density of your commits over time, and the total time elapsed, you'll end up having to fix your master at least once
19:57 Mithaldu or have a bunch of dumb "revert blah" svn-like commits
19:57 mohawk i don't see that as a catastrophe
19:57 mohawk although i did have a big old history-editing-on-master session on dmake yesterday
19:57 Mithaldu it is :(
19:58 mohawk tell me how it's a catastrophe
20:00 Mithaldu let me try and draw a picture
20:00 Mithaldu you have a house
20:00 mohawk oh god, not proof by bad analogy
20:00 mohawk this is pure think-stuff, not physical stuff
20:00 Mithaldu mohawk: you know i will RUN this shit into the GROUND
20:00 Mithaldu in the basement is the main breaker for that house
20:01 Mithaldu but the room with the main breaker is not in the main basement, but off to the side and its own stair
20:01 Mithaldu people don't often need to go down there, maybe only once or twice a year
20:01 Mithaldu but the railing is missing on that set of stairs and it's narrow so it's easy to fall and hurt yourself
20:02 Mithaldu but nobody cares to fix that, because at that time they didn't want to rebuild the foundation, since it was already in place, and nobody thinks it's important since you only have to be careful
20:03 Mithaldu nevertheless every once or twice in a year someone will DIE there
20:03 Mithaldu so there :|
20:03 opkick [pdl] wchristian force-pushed prereqdump from e79d804 to 7804080: http://git.io/xZ2e
20:03 opkick pdl/prereqdump 7804080 Ed J: Add test that dumps installed version of required modules
20:03 Mithaldu you're advocating broken railings!
20:08 mohawk instead of reading the above (yet), i just learned about triple-dot on commit ranges!
20:08 mohawk that's some good stuff
20:08 mohawk and i've just configured it to accept force-pushes
20:08 mohawk there's no easy way to configure per-branch, and i don't want to make a hook
20:08 mohawk so that will do for now
20:08 Mithaldu triple dot?
20:09 mohawk last section of http://git-scm.com/book/ch7-​1.html#Server-Configuration
20:13 mohawk just read your argument above
20:13 mohawk i think it's a simple non-sequitur
20:14 mohawk there might be further fallacies lurking in there, but i think that one will do
20:16 mohawk actually, i think it might be (from https://en.wikipedia.org/wiki/Lis​t_of_fallacies#Informal_fallacies)
20:17 mohawk both Ignoratio elenchi AND argumentum verbosium
20:17 mohawk that's good solid fallacious work, my friend
20:18 mohawk anyway
20:19 Mithaldu i tried to make it more palatable by making it interesting to read
20:19 Mithaldu i'm a little disappointed now, since the point i'm trying to make is serious
20:20 Mithaldu if you refuse to fix even the worst fuck-ups, someone will stumble across them at the worst point in time, while trying to fix a bug
20:20 Mithaldu this will not be a common occurence, but that person will hate you
20:24 mohawk i don't think a revert commit qualifies as that
20:25 mohawk and there is simply no way to prevent people from missing things or otherwise making mistakes
20:25 mohawk that's another fallacy, the counsel of perfection
20:25 Mithaldu of course, it depends on the magnitude of the fucking
20:25 Mithaldu that's why i said "worst", not "typo"
20:26 Mithaldu *fuck-up
20:26 Mithaldu basically, please try to assume i'm earnestly arguing something that is thought through, and not just a kneejerk reaction :)
20:34 mohawk i am assuming that
20:34 mohawk i just disagree
20:34 Mithaldu then you wouldn't be accusing me of fallacies :)
20:34 mohawk i'm going to send yet another email seeking consensus for forbidding merge commits
20:35 mohawk i am not accusing you of anything
20:35 mohawk i am pointing out flaws in the way you have argued your position
20:35 sivoais joined #pdl
20:35 mohawk if i were accusing you of things instead of addressing your argument, that would itself be the argumentum ad hominem fallacy
20:36 mohawk sent
20:36 Mithaldu ok, i get the difference
20:36 Mithaldu i still think you're fundamentally misunderstanding me, but at the end of the day, i just to have usable PDL
20:37 mohawk i can only go by the things you say, and limited reasonable inferences on what you think
20:37 Mithaldu and trying to keep you from what i consider to be inadvertantly hurting someone also keeps you from working on it, so i'll shut up :)
20:38 mohawk that's a prediction of someone being harmed - if you can show that without bizarre, bad analogies i'll be pleased to listen :-)
20:38 Mithaldu i TRIED
20:38 Mithaldu in this line:
20:38 Mithaldu "if you refuse to fix even the worst fuck-ups, someone will stumble across them at the worst point in time, while trying to fix a bug"
20:39 Mithaldu if, for example, you leave a massive test-breaking commit in the middle of your history, you make the job of anyone trying to find a bug caused before that commit much harder
20:39 Mithaldu as i admitted, that occurence is rare, but the chances of it are decidedly non-zero
20:40 mohawk the "fix" is to just have another commit correcting the problem
20:41 Mithaldu no, that doesn't help
20:41 mohawk down that road lies going back to edit every bug away and erasing it from history
20:41 mohawk well you say it doesn't help
20:41 Mithaldu no there is no slippery slope
20:41 mohawk so that's the point of fact/opinion where we disagree
20:42 Mithaldu see that's what upsets me
20:42 mohawk i know for a fact my position is both widely held, and reasonable
20:42 Mithaldu everytime you disagree with something i say your reaction is to outright stop at that point
20:42 mohawk so i feel regret that you are upset by it, but i won't change it on the evidence you've shown
20:42 Mithaldu instead of...
20:42 Mithaldu goddammit mohawk
20:43 Mithaldu you didn't even bother to wait for me to explain why i'm upset
20:43 Mithaldu that is not how dialogue works between people who respect one another
20:44 Mithaldu if there's disagreement, the most helpful course of action is to ask, ask for elaboration, clarification, more accuracy
20:44 mohawk with respect, you saying "that's what upsets me" when i say "that's the point on which we disagree" doesn't sound like you're going to add anything
20:44 mohawk but if you were going to, then just add it
20:44 mohawk stop being so precious :-)
20:44 mohawk well why don't you just clarify anyway
20:45 mohawk i explicitly said "on the evidence you've shown so far"
20:45 mohawk that is the opposite of "outright stop at that point", which i don't appreciate you claiming of me
20:46 Mithaldu i'm talking about these two lines:
20:46 Mithaldu 15-02-28@21:41:42 (mohawk) well you say it doesn't help
20:46 Mithaldu 15-02-28@21:41:52 (mohawk) so that's the point of fact/opinion where we disagree
20:46 Mithaldu to make an example of how i would've handled that:
20:46 Mithaldu > well you say it doesn't help
20:46 Mithaldu > why do think it doesn't?
20:48 Mithaldu also, to answer the question i posed there: because an additional revert commit does not fix bisect getting a fit, and does not fit blame output being incomprehensible
20:49 mohawk so here's why it's a slippery slope
20:49 mohawk because frequently one makes commits that pass tests locally, but not on all platforms
20:49 mohawk so i can happily bisect away on linux, but someone on win32 can't
20:49 mohawk what then?
20:50 mohawk you have to accept that git bisect has inherent limitations
20:50 mohawk otherwise you're on the counsel of perfection again
20:51 Mithaldu to be honest, that is completely acceptable
20:51 Mithaldu generally for a given project you have roughly one vaguely defined platform where you have stability continuity
20:52 Mithaldu you're correct to bring up counsel of perfection, because i do take that into account
20:52 Mithaldu note that i did not say: eradicate EVERY test-breaking commit
20:53 Mithaldu just, be open to backing out the worst ones when caught early
20:53 mohawk well then we're only arguing about what "early" should mean there
20:53 mohawk i say early = when still on its topic branch
20:53 mohawk and not on master
20:53 mohawk again, i know you disagree
20:55 Mithaldu catching it on the topic branch is optimal
20:55 Mithaldu but i think it's too optimistic to expect that to not fail with a decent probability
20:56 Mithaldu we have branch reviews on various projects i work on and still crap got to master from time to time
20:57 Mithaldu and in those cases, when it was caught on the same day, and was still the top commit, that commit was backed out
20:57 Mithaldu mainly because we can assume that anyone else working on the project was experienced enough to handle that kind of change in stride
20:57 Mithaldu now, if you assume git inexperience to be the norm in your project, forbidding master backouts is acceptable
20:58 Mithaldu as long as you're aware that you're exchanging debugger pain for newbie convenience
21:01 * jberger gets part of the way through the backlog, decides to skip
21:01 Mithaldu jberger: morning o/
21:02 jberger Mithaldu: good afternoon \o
21:11 mohawk jberger, that may be wise
21:12 mohawk Mithaldu, on the project under discussion, we still have git inexperience
21:12 mohawk when that changes, we might revisit
21:13 mohawk sivoais, https://travis-ci.org/PDLPorters/pdl/jobs/52562618 doesn't show the prereqs, but that relies on tests running - we need to add the "prove prereqs.t" first
21:13 mohawk i'll do that now
21:13 Mithaldu mohawk: word, then we are in agreement :)
21:17 mohawk indeed!
21:17 opkick [pdl] mohawk2 force-pushed prereqdump from 7804080 to 0a9f6e3: http://git.io/xZ2e
21:17 opkick pdl/prereqdump 0a9f6e3 Ed J: Add test that dumps installed version of required modules
21:17 mohawk sivoais, let's see if that helps
21:17 mohawk (i put in the "prove" before the "make" but after the cpanm stuff)
21:18 mohawk travis is already running
21:19 mohawk https://travis-ci.org/PDLPor​ters/pdl/jobs/52573278#L2558
21:19 mohawk aww yeah
21:23 Mithaldu pfft, travis
21:23 Mithaldu web 2 O hippies
21:28 opkick [pdl] mohawk2 pushed 1 new commit to prereqdump: http://git.io/xneB
21:28 opkick pdl/prereqdump 01e6577 Ed J: ExtUtils::ParseXS req 3.01, 2.21 known broken on generate_core_flags()
21:29 mohawk sivoais, jberger - can you configure it so i can stop build #11?
21:29 mohawk it's pointless because i just changed things
21:29 jberger you should be able to stop it, no?
21:30 mohawk i don't see the link i was expecting - where would i look?
21:30 jberger https://travis-ci.org/PDLPorters/pdl
21:30 jberger top right corner
21:30 jberger (x)
21:30 mohawk yeah, so it hasn't movd
21:30 jberger I can kill it if you can't, but check to see
21:30 mohawk i have used it before
21:31 mohawk please zap #11
21:31 jberger zapped
21:31 mohawk thanks
21:31 mohawk please also look at settings, i'll google too
21:31 jberger np
21:31 mohawk i'm assuming you're comfortable with me having that control ;-)
21:32 jberger I am, I'm curious why you don't
21:32 jberger *don't have it
21:33 mohawk can you look over the settings?
21:33 mohawk i know i have it for ptg repos
21:33 mohawk as i've used it extensively there ;-)
21:36 jberger I don't see any per-user settings
21:36 mohawk so what is my status on that repo on github side?
21:37 jberger you are listed under owners
21:38 mohawk ok, is that a github thing?
21:38 mohawk let me compare with ptg
21:38 jberger I see no difference in configuration between you and me
21:41 mohawk so it must be something else
21:41 mohawk on ptg i'm a "member", on pdlp i'm an "owner"
21:41 mohawk remind me, what's the text it says when offering to stop a build?
21:41 mohawk that'll help the googling :-)
21:42 mohawk jberger, ^
21:44 jberger "Cancel Buid"
21:45 jberger build
21:56 mohawk great
21:56 mohawk my travis profile shows my other memberships, but not pdlp
21:57 mohawk oh now it shows it
21:57 mohawk after i clicked "sync from github" on my profile
21:57 mohawk and now cancel is available too
21:57 mohawk problem solved
22:00 mohawk and now, back to this parsexs thing
22:05 jberger hunh, ok good to know
22:05 jberger travis keeps some cache of perms
22:11 mohawk exactly
22:11 mohawk and it said the cache was 5 days old
22:14 opkick [pdl] mohawk2 force-pushed prereqdump from 01e6577 to 9a76379: http://git.io/xZ2e
22:14 opkick pdl/prereqdump a9f1eb3 Ed J: Add test that dumps installed version of required modules
22:14 opkick pdl/prereqdump 9367dbe Ed J: ExtUtils::ParseXS req 3.01, 2.21 known broken on generate_core_flags()
22:14 opkick pdl/prereqdump ca7c294 Ed J: Make prereq-dumper include where module is
22:17 sivoais mohawk++
22:19 sivoais mohawk: I think the successful builds are even higher than 3.01 O_o
22:21 sivoais haarg made the Travis Perl builds that we're using and the version of the core modules seem to be different from the versions that are in corelist
22:21 sivoais if I'm looking at my settings correctly <https://travis-ci.org/zmug​hal/p5-PDL/builds/52576837>
22:24 sivoais mohawk: why did you cancel the build?
22:28 opkick [pdl] mohawk2 force-pushed prereqdump from 9a76379 to 13768cc: http://git.io/xZ2e
22:28 opkick pdl/prereqdump 13768cc Ed J: Make prereq-dumper include module filesize and location
22:28 mohawk because i fucked up the test mod :-)
22:29 mohawk this one works now
22:29 sivoais oh, I had started it up again :-P
22:29 mohawk don't do that
22:32 mohawk please let's not second-guess each other
22:32 sivoais ok, I just wanted to see where the problem was :-P
22:32 mohawk next time, ask
22:32 sivoais Umm, so in #14.2 <https://travis-ci.org/PDLPorters/pdl/jobs/52578148>
22:33 sivoais I did, but I wasn't patient enough
22:33 mohawk so it looks like the build process is installing EU::PXS in site
22:33 mohawk but the build process is finding & using one in perl itself
22:33 mohawk viz https://travis-ci.org/PDLPor​ters/pdl/jobs/52578148#L2737
22:34 sivoais yeah
22:35 sivoais mohawk: I have another branch that works if I do : cpanm --force --verbose ExtUtils::ParseXS
22:35 sivoais before the build-dist
22:36 mohawk ah HA
22:37 mohawk even on 5.12?
22:37 mohawk is that recently?
22:37 sivoais yes
22:37 sivoais let me linky
22:37 mohawk please do
22:37 sivoais <https://travis-ci.org/zmug​hal/p5-PDL/builds/52522291>, <https://github.com/PDLPorters/pdl/com​pare/master...zmughal:travis-ci-more>
22:39 sivoais re: the IRC notifications in there, I'll change them to #pdl once it lands in PDLPorters/pdl. We'll see if they are too noisy then.
22:39 mohawk could you do that now?
22:39 sivoais sure
22:39 mohawk remember on #inline, even with the i-m stuff, it was fine?
22:40 mohawk if you do that, we can merge together your branch and mine
22:40 mohawk in fact, i'll just get mine in a good state for that
22:40 mohawk you can rebase it onto yours or i can do it
22:40 sivoais and get the pre-reqs + notifications + parsexs working at once! \o/
22:41 mohawk yup
22:41 mohawk that's the plan
22:41 mohawk THIS is how collaborative stuff with git is done, folks
22:43 sivoais mohawk: rebase now?
22:43 mohawk hold on :-)
22:43 mohawk now
22:43 mohawk 3 commits, last is optional
22:43 opkick [pdl] mohawk2 force-pushed prereqdump from 13768cc to 56093ef: http://git.io/xZ2e
22:43 opkick pdl/prereqdump 34cf96c Ed J: Make prereq-dumper include module filesize and location
22:43 opkick pdl/prereqdump 56093ef Ed J: ExtUtils::ParseXS req 3.01, 2.21 known broken on generate_core_flags()
22:43 mohawk ie, use your best judgement on that
22:44 mohawk 1st 2 are import the test, and my mods to it
22:44 sivoais I think it is good to have the 3.01 version bump to EU::ParseXS. That was when it got refactored.
22:45 sivoais at least bring it to 2011 ;-)
22:45 mohawk sounds good
22:46 mohawk chm just approved branch-zapping, i'll do it now
22:47 opkick [pdl] mohawk2 deleted cpan_2.006_release at 405de0a: http://git.io/xnBB
22:47 opkick [pdl] mohawk2 deleted cpan_2.007_04 at 7f1cdfa: http://git.io/xnBz
22:47 opkick [pdl] mohawk2 deleted cpan_2.007_release at c58b3cf: http://git.io/xnBV
22:49 * sivoais needed this coffee. The caffeine is finally kicking in.
22:50 opkick [pdl] mohawk2 deleted release_2.4.10 at 99c4411: http://git.io/xnRL
22:53 opkick [pdl] wchristian created release_2.4.10 at 99c4411 (+0 new commits): http://git.io/xnRH
22:53 mohawk slight pause on that one because it was apparently ambiguous, had to look up how to specify a branch
22:53 mohawk it was this: git push upstream :refs/heads/release_2.4.10
22:54 mohawk (the ambiguity was using git push --delete, and it's because there's a tag with same name)
22:55 mohawk why are those branches still showing on this? https://github.com/PDLPorters/pdl/branches/all
22:56 sivoais I suppose the script doesn't sync those up
22:56 mohawk we do need to adjust it to use "--mirror"
22:57 mohawk Mithaldu, could you make that happen?
22:57 mohawk i'll manually zap these
22:57 sivoais there's a tag called v2.4.10 as well
22:57 Mithaldu mohawk: not tonight, i'm going to bed atm
22:57 opkick [pdl] mohawk2 deleted cpan_2.006_release at 405de0a: http://git.io/xnBB
22:57 opkick [pdl] mohawk2 deleted release_2.4.10 at 99c4411: http://git.io/xnRL
22:57 mohawk fair enough! night
22:58 mohawk sivoais, how's the travis shtuff going?
22:58 sivoais builds are green so far! Build #90 <https://travis-ci.org/zmug​hal/p5-PDL/builds/52579249>
22:59 sivoais the 5.12 issue is fixed, so it should roll right along
22:59 mohawk good work
22:59 sivoais and we'll see an IRC notification in 20 min or so
22:59 mohawk except... why not go via the sf.net one?
23:00 sivoais ah, this will mean it generates 2 notifications...
23:00 mohawk only if you still use your p5-pdl one :-)
23:02 sivoais I dunno. I like the delay of working on my own fork. It just lets me mess about until I get it right.
23:02 sivoais because when I go, git push upstream, I want to be sure :-P
23:03 sivoais actually, if I worked on the SF, that would still generate 2 notifications. One for the topic branch and another for the master
23:04 mohawk only on master when you merge, surely
23:04 sivoais we'll see if Travis-CI uses the SHA to save time
23:06 mohawk but since you're not going to merge to master till it's gone via the review process, that's a bit moot, surely
23:06 mohawk the only notifications would be re your topic branch
23:06 mohawk help me, i don't understand the problem here
23:09 sivoais Not really a problem unless it gets too noisy. We'll see how it works. I shall try that by getting some of my test cleanup in next.
23:09 sivoais No use guessing how Travis-CI will work since we don't have their source. :-P
23:12 sivoais at a future point, I'd like Windows builds for CI
23:15 mohawk one step at a time
23:16 mohawk that probably won't happen before the freeze for 2.008
23:16 sivoais yeah
23:29 mohawk let's get green lights on linux first
23:29 mohawk and 2.008
23:30 sivoais almost there...
23:31 opkick [pdl] mohawk2 deleted tempbranch at 27c804e: http://git.io/xnrH
23:31 sivoais go, build, go!
23:32 mohawk ok, Mithaldu's mirror script keeps recreating those branches
23:32 mohawk that's irritating
23:33 travis-ci joined #pdl
23:33 travis-ci zmughal/p5-PDL#90 (travis-ci-more - 316dba2 : Zakariyya Mughal): The build passed.
23:33 travis-ci Change view : https://github.com/zmughal/p5-PDL/c​ompare/2abf927182a3...316dba242057
23:33 travis-ci Build details : http://travis-ci.org/zmugh​al/p5-PDL/builds/52579249
23:33 travis-ci left #pdl
23:33 mohawk sivoais++
23:33 opkick [pdl] wchristian created tempbranch (+1 new commit): http://git.io/xOGl
23:33 opkick pdl/tempbranch 27c804e Ed J: test commit
23:33 mohawk damn fine work young man
23:33 sivoais \o/
23:33 mohawk wait, what the script needs is --prune as well
23:33 mohawk i'll comment on the gh issue
23:34 sivoais now I have to eat dinner, but how do I procede now. Send e-mail to pdl-devel and point to branch and Travis-CI?
23:34 sivoais be back in 1hr or so
23:37 mohawk let me look over it first
23:38 mohawk please push it to sf.net for all to review
23:40 mohawk meantime, bon appetit :-)
23:47 jberger yapc::na talks submitted!
23:51 mohawk jberger++
23:51 jberger only one day left
23:51 jberger ::nudge::
23:51 sivoais :-D
23:52 sivoais ok, I pushed it to SF and I'll write up a request for review in a new thread to pdl-devel
23:52 sivoais that might have to wait a bit
23:52 mohawk ok
23:53 opkick [pdl] wchristian created travis-ci-more (+4 new commits): http://git.io/xnPX
23:53 opkick pdl/travis-ci-more 64d73fe Zakariyya Mughal: Travis-CI: no Fortran
23:53 opkick pdl/travis-ci-more bcd2ab2 Zakariyya Mughal: Travis-CI: allow builds with BADVAL disabled to fail...
23:53 opkick pdl/travis-ci-more 749af6b Zakariyya Mughal: use the latest version of ExtUtils::ParseXS
23:53 sivoais coolio!
23:54 sivoais 5 minute delay?
23:55 jberger down from the hour and a half it took to get my mailinglist confirmation
23:55 jberger doing pretty good I guess
23:56 mohawk sivoais, you're not including the prereq-dumper from my branch?
23:56 sivoais I did? Let me check
23:57 mohawk oh, you pushed onto mine
23:57 mohawk that's groovy
23:57 mohawk i just stopped the travis build on mine so yours can run
23:58 mohawk i'll cancel the others too
23:58 mohawk Mithaldu's script made it do a bunch :-)
23:59 mohawk ok, now #22 should be able to start
23:59 mohawk https://travis-ci.org/PDLP​orters/pdl/builds/52583325
23:59 mohawk jberger, those delays really do slow things down, don't they

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