Camelia, the Perl 6 bug

IRC log for #onionsketch, 2013-06-19

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

All times shown according to UTC.

Time Nick Message
00:23 khw joined #onionsketch
00:50 khw joined #onionsketch
04:23 khw joined #onionsketch
05:05 ether_ joined #onionsketch
13:30 xdg joined #onionsketch
13:30 rjbs left #onionsketch
13:30 rjbs joined #onionsketch
13:30 rjbs WTF...
13:31 rjbs I was getting a bizarre error.
13:31 rjbs I couldn't send to channel despite being voiced and an op.
13:31 rjbs Well, all's well that ends well.
13:31 xdg blame mst :-)
13:31 rjbs And good morning!
13:31 rjbs The strategy of first resort!
13:31 xdg Yes, to you too!
13:32 rjbs I can't believe June is almost over. :-/
13:32 xdg I know
13:32 rjbs So, last time I walked away with three to-do items.
13:32 xdg I mean, technically, we're not far past halfway, but it's rushing forward
13:33 rjbs One was #118297, which is now in maint-5.18.  Done.
13:33 rjbs One was to create a META for 5.18.1 blockers, done.
13:33 rjbs The other was about 02packages.
13:33 rjbs https://github.com/andk/pause/issues/57
13:33 rjbs I've made that and emailed Andreas about it, but no reply.  I will email again.
13:33 rjbs This isn't likely to really affect you, since blead releases are skipped anyway.
13:34 xdg low bus number bites us
13:34 rjbs Yeah.  We can talk about getting some more access to the db directly.  (A vision of RJBS forgetting a WHERE on a DELETE...)
13:35 rjbs but the impact of this problem is, fortunately, pretty low for now
13:35 rjbs and the case-insens patches are live in pause, so new things should not be cropping up
13:35 xdg we're only talking about it because we're the ones hacking on PAUSE
13:35 xdg Let's move on
13:35 rjbs Yup.  5.19.1 seems "normal" from here.  How's it look from over there?
13:36 xdg No red flags.
13:36 xdg No responses to my email on the list and smokes seem generally OK
13:36 xdg (fsvo OK)
13:36 rjbs (Oh, I see, the PAUSE thing was in there as an "old business" from 5.18.0, which was affected.)
13:36 rjbs Yeah.
13:36 rjbs How's the delta?
13:36 xdg I need to see if some perldelta things landed from davem and steve hay
13:37 xdg people have been pretty good about updating it, but sadly not in a way that's easy to track
13:37 rjbs What would make it easier to track?
13:37 rjbs (other than putting work and delta on a merged branch :)
13:38 xdg problem is: commit, commit, commit, commit, perldelta commit
13:38 xdg merged branches might help
13:38 rjbs Yeah.  Well, hey, I am in favor of more branches and merges, for sure.
13:38 xdg but I'm not sure p5p is ready for git flow style
13:38 rjbs Heh.  I don't think we even need that.
13:39 xdg well, more no-ff merges anyway.
13:39 xdg but i'm not sure how much process we need to enforce for the sake of perldelta
13:39 rjbs I need to find someone to rewrite the "rebase-then-merge" procedure, because it seems proven that I can't explain it sufficiently well to a non-githead.
13:40 xdg ask auggy?
13:40 rjbs Yeah, for ten-commits-then-delta, a merge seems useful.
13:40 rjbs xdg: Can try!
13:40 xdg she's lookign for way to get involved and has technical writing experience
13:40 rjbs Well, anyway: if you have any good ideas on how folks can easily indicate how many commits they just covered with their delta patch, let me know. :)
13:40 rjbs (well, let everybody else know, too)
13:41 xdg The other thing that would help is if people said "no perldelta needed" more often in commits
13:41 rjbs Smokes: decent. Delta: decent.
13:41 rjbs Oh, that's a really good idea, yes.
13:41 xdg e.g. "I tweaked this internal thing.  No delta entry needed."
13:41 xdg there were a few like that
13:41 xdg Or "will add perldelta later" if it's not on a branch
13:41 rjbs Yes, that would be a big help, I know I've sent my share of emails saying, "Does this need a delta at all?"
13:42 xdg I don't know how many other RMs use git-deltatool.  But I'm planning to tweak it to collapse branch merges.
13:42 xdg (Proper merging is needed though, or I have to do a lot of logic about it)
13:42 rjbs I have not used it, which is mostly a result of me mostly working on branches that need nearly no delta, or a huge delta compilation.
13:43 rjbs Maybe I should do a "normal" release later this year and give it a go.
13:43 rjbs I don't think we have any gonna-break-the-world changes in 19.1 so far, do we?
13:43 xdg I don't think so.
13:43 xdg Or at least, nothing on the list seems to suggest it.
13:44 rjbs Right.
13:44 xdg can't prove a negative and all that
13:44 rjbs heh
13:44 rjbs Okay.  So it sounds like smooth sailing, unless something horrible comes up in the next 36 hours.
13:44 rjbs and you already said "don't do anything horrible"
13:44 xdg better commit messages in general would help, but I don't know how to get that happening
13:45 xdg The big unknown I have is Module::CoreList.
13:45 rjbs Oh, cripes, right.
13:45 xdg Not a problem for release, but then afterwards...
13:45 rjbs So there are at least two changes with that:
13:45 rjbs * regen after release
13:45 rjbs * deprecated is now a delta
13:45 rjbs I've tested the second one, pretending to bump to 5.19.2;  seemed okay.
13:46 xdg I'm going to do a dry run through the RMG today and may try to tackle M::CL in a branch
13:46 rjbs Now, it looked like we're fixing two problems with that first point:
13:46 rjbs 1. make sure $] is in the core list as early as possible
13:46 rjbs 2. keep the removed/deprecated status up to date during blead for andk's smokers (and others)
13:46 rjbs regen after release solves #1
13:47 rjbs Are we going to need to regen immediately after cpan imports, etc., too?
13:47 xdg It would be great if we can get there.
13:47 rjbs The last time I looked, doing the corelist diff twice would create two sections for $] rather than update one in place.
13:47 rjbs Know whether that's still the case?
13:47 xdg I need to see how deltas were implemented.
13:47 xdg I don't know.
13:47 rjbs Okay.
13:48 xdg At one time, we'd mooted separate files per release.
13:48 rjbs I feel like we're going to end up wanting to become even more datafile-driven.
13:48 xdg That would allow blowing away and regenerating.
13:48 xdg Yeah.
13:48 xdg I took a cut at it last year (?) but I think BinGOs implemented something else.
13:48 rjbs Hm.  Yeah, there were a few solutions to different problems going on at the time.
13:49 xdg Yeah.
13:49 rjbs Well, we have some ideas for ways forward, then, if things get hairy.  So, good.
13:49 xdg Regardless, it doesn't block the release.
13:49 rjbs Yes.
13:49 xdg If I see an easy fix, then great.  Otherwise, we'll muddle along until someone finds enough tuits for a fix.
13:49 rjbs todo for me: review andk's original error report to see what it implies about regen frequency
13:49 rjbs right
13:50 xdg Ideally, we'd have a porting test to check M::CL against what we have.
13:50 xdg (much like version bump check)
13:50 rjbs yes
13:50 xdg That should go on a todo list somewhere
13:50 rjbs I think we've covered everything.  You implied you have an epigraph, so that's covered. :)
13:50 xdg yes
13:50 rjbs I'll put it on my todo to put on a todo ticket.
13:51 xdg can we create tickets in a browser yet?
13:51 xdg (or from a CLI?)
13:51 xdg (without email)
13:51 rjbs from a CLI, I think
13:51 rjbs This is the week I get back to bugging BPS about RT upgrade, though.
13:51 rjbs which leads to many other things
13:51 xdg woo
13:51 rjbs (our worker there was on leave until just recently)
13:51 rjbs getting youa link...
13:51 xdg I noticed that rt.cpan.org now flags if a distribution has a non RT tracker, which is awesome
13:52 rjbs https://rt.perl.org/rt-cli.html
13:52 rjbs Yes, that is awesome!
13:52 rjbs Maybe the CLI doesn't do creation.
13:52 xdg (sigh)  "Download ... and put it in your $PATH
13:52 rjbs At any rate, getting tickets via the web is on my agenda.
13:52 rjbs I did a survey of other language projects a few months ago.
13:52 xdg They know we have this thing called CPAN for distributing code, right?
13:53 rjbs The great majority of them all agree: "verified email on signup, then use account to submit bug"
13:53 rjbs Heh.  Well, if the RT-CLI is useful, we can put it on CPAN for easy install.
13:55 rjbs Okay.  I think we've covered the gamut.  Anything else to cover or get on the log before calling it done?
13:55 xdg I don't think so.
13:56 rjbs Okay!  Well, you know where to find me.  I look forward to upgrading to 19.1!
13:56 xdg woo!
13:57 * rjbs will now go back to idling here for a few weeks.
13:57 rjbs Oh: I meant to get Aristotle here today, and did not.  He owes me a bunch of pre-release setup data.
13:57 rjbs I'll let you know when our first-half-July #onionsketch is, and if you can make it to catch him up on any useful things, super.  If not, no big deal.
13:58 xdg OK
14:26 khw joined #onionsketch
22:25 khw joined #onionsketch

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