Camelia, the Perl 6 bug

IRC log for #onionsketch, 2012-03-01

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

All times shown according to UTC.

Time Nick Message
03:56 kentnl joined #onionsketch
05:32 TonyC joined #onionsketch
10:18 Zefram joined #onionsketch
10:29 Mithaldu joined #onionsketch
16:26 perigrin joined #onionsketch
17:44 Mithaldu joined #onionsketch
18:17 Corion joined #onionsketch
21:10 leont joined #onionsketch
21:48 Topic for #onionsketch is now The State of the Onion | http://markmail.org/message/e6w4t2umxsnqv6op
21:55 Abigail joined #onionsketch
22:00 rjbs Okay!  Shall we begin?  I don't expect too much trouble today, as we're just starting things off.
22:01 Corion Yeah, why not
22:01 Abigail Ok
22:02 rjbs Great.  So, I posted our mini-agenda to p5p, and there's a link in the topic.  First up: old business.
22:02 rjbs Well, this is our first meeting.  So no real old business.  And that's just fine with me.
22:03 rjbs I don't think there are any open security issues to discuss right now...
22:03 Corion I'm unaware of any
22:03 rjbs I think you've both seen the kerfuffle on p5p about our longstanding :utf8 layer problems.
22:03 Corion Yep
22:03 Abigail Yes
22:04 rjbs Leon T. is going to be pushing up some of his related work soon, and I'm going to make a "[META]" ticket to collect related issues.
22:04 rjbs I still don't expect us to land a fix before 5.16.0, but since people are in motion, hopefully they can be directed toward those two things to offer their assistance.
22:04 * Abigail nods.
22:05 rjbs Corion: 5.15.8 seemed to go pretty smoothly from where I sat watching.
22:05 rjbs Corion: Any snags worth relating?
22:05 Corion Yep - and the snags I hit are mostly patched.
22:05 rjbs What snags remain?
22:05 Corion There is one where I think more patching is needed, but I haven't understood the program yet.
22:06 Corion The corelist generator currently has "use Algorithm::Diff" in it, which is used for the "check" case
22:06 Corion That case seems to be supposed to be run with an outside Perl
22:06 Corion while the recommended usage from the releng guide is to run it with the soon-to-be Perl
22:06 Corion (and the "generate" case, not the "check" case)
22:06 Corion At least, that's the impression I got from a) the documentation b) rafl's use case
22:07 Abigail Is it a problem if something needs to be run with an outside Perl?
22:07 Corion But before patching, I want to look at the code again, and think about this. I accidentially reverted some of rafls changes, not understanding why he had made them.
22:07 rjbs I think it's a-ok for stuff in ./Porting to need an outside perl -- but if the documentation shows using fresh-built ./perl, we'll definitely want to sort that out.
22:08 Corion Abigail: I think that the "generate" use case creates bad data if it's not run with the soon-to-be Perl, because it takes the last two entries in Module::CoreList
22:08 Corion So if you run it with an outside Perl, it picks up different data.
22:08 Abigail Ah, ok, I understand.
22:08 Corion I changed "require Algorithm::Diff" to "require '/outside/perl/lib/Algorithm/Diff.pm'
22:08 Corion just to get the program to compile/run
22:09 Corion I think the fix will be to move the "require" line downwards to where it is actually needed. But before I do that, I want to verify my understanding
22:09 Corion But I think that's it. The other cases I've already patched in the documentation (github access, and something else which I forgot)
22:09 rjbs Cool.  This program is safe to run all the time, so hopefully you can bash on it until it is clear, and then make a fix. :-)
22:10 Corion I'm not yet clear as to what output(s) it is supposed to produce - rafls use case is different from "mine"
22:10 Corion Abigail: If I make any changes, I'll tell you, as you're the next one to really need that program :)
22:10 rjbs Corion: When you make that change, drop me an email.  Otherwise, I'll check in with you before my next meeting with Abigail, and maybe have a go at it myself.
22:10 rjbs Great.
22:11 Corion But yes, the release manager guide seems fairly good, as it took me six hours of mostly waiting to produce a new (point) release of Perl.
22:11 rjbs Worst case, I don't think it should be too tough to do even if the script becomes horribly broken.  It will also help that so few things will change version between .8 and .9!
22:12 rjbs Yeah.
22:12 rjbs Corion: Any other post-5.15.8-release notes?
22:12 Abigail Yeah, I went through the guide a few times, and none of it seemed to be confusing.
22:12 Corion Yes - proof-read the release mail before starting the work
22:13 Corion I had a malformatted link in the mail going out to p5p, but was too tired to proof-read the mail
22:13 rjbs :) I think I did the same thing myself, once.
22:13 rjbs So I agree! :)
22:13 Corion (but I was glad I had prewritten the mail already, as I'd have been too tired to paste the mail together)
22:13 rjbs Abigail: The RMG has gotten better and better, you're lucky!
22:13 rjbs Abigail: Speaking of which:  this is your first release, right?
22:14 Abigail That's why I waited so long to step in ;-)
22:14 Abigail Rjbs: yes
22:14 rjbs Abigail: ...but I see you've got commit access, right?  You've used it recently enough to know your keys is still good ?:)
22:15 Abigail I have commit access. And I know Dennis.
22:15 rjbs Aha, of course.  Excellent.
22:15 Corion Oh - one more thing - the one thing we backed out, Compress::Raw::Zlib had a new upstream (zlib) release, so it should be better now. But I think it would be preferrable if it landed at least two weeks before freeze to give the smokers a better run.
22:15 Abigail I also contacted Andreas and Graham to make sure I have the rights to upload a release.
22:16 rjbs Abigail: Oh, great, that was my next question
22:16 Abigail And they both have replied; I should have that access.
22:16 rjbs Corion: I asked PMQS whether he'd consider making a release with /just/ the warning fix, to keep changes to a minimum...
22:17 rjbs Corion: ...but good to know.  I'll see if we can't get that family of libs updated in a smoke-me branch ASAP.
22:18 Corion rjbs: That would be good for the core, yes - even if it doesn't fit in with the CPAN history, but having 2.0481 with only the warnings fixes would be good for 5.14.x
22:18 rjbs One other "known issue" right now worth mentioning is "the Carp dot."
22:19 rjbs Steffen began a new run of smoking that, yesterday or the day before.  I don't think have any good actions to take beyond "keep an eye on the list and assist people with patches."
22:19 Corion Well - serves people right to use tight regular expressions on Perls error/warning messages. Maybe one day we hand out unique numbers for every string.
22:19 rjbs Abigail: ...but I don't expect you to have to do any of that work.  But it's something we'll potentially see reported to the list.
22:20 rjbs Corion: I sympathize with both sides of the problem. :(
22:20 rjbs Abigail: As for things I *do* hope you will be doing... :)
22:20 Corion Me too - I've relaxed lots of my tests, and luckily I don't have any modules that close to the toolchain
22:20 Corion (except parent.pm , but I'm highly reluctant to make changes there :) )
22:21 rjbs Abigail: Keep an eye on the smoke reports?  Right now, things look about like ususal, but with a bit more black smoke from Win32 than usual, mostly (I think!) due to known problems.
22:21 rjbs So I should mention the other known problems, there: Pod-Html's path normalization.  Craig seemed to have beaten it into submission, but it's acting strangely on at least one Win32 smoker, still.
22:21 rjbs ...and there's a problem with fork emulation on Win32 that we're still not sure what to do with.
22:22 Abigail Yeah, I'm a bit behind on p5p, but except for our Alar game, my weekend, including the Friday night, is free. I should be "current" soon.
22:22 rjbs Great.  Hopefully you can identify which threads you can ignore for the time being early on. :-)
22:22 rjbs Or maybe you're a faster reader than I am and you can enjoy every last paragraph.
22:23 Abigail I'll try.
22:24 rjbs re: other bugs that come up
22:24 Abigail But I expect to have a good idea before the weekend is finished. And I'll keep close tabs on p5p afterwards.
22:24 rjbs I think we're going to get two kinds of bugs:
22:24 rjbs * bugs that are show stoppers
22:24 rjbs * bugs that have to wait
22:25 rjbs I'd rather not see much misc. bug fixing in blead unless it's directly toward blockers or black smoke (which is a blocker).
22:25 Abigail Everything else is for 5.17.1?
22:25 rjbs If you see a ticket land that you think is a blocker, point me at it?  It turns out that this month has become *really* busy, and I'd hate to miss a ticket.
22:26 * Abigail nods.
22:26 rjbs Yeah, I think this is the time to remind all the committers that they can put their bugfixes in topic branches, and we can rebase them once maint-5.16 is branched.
22:26 rjbs (and if they want to keep using topic branches more extensively after 5.17.0 begins, that's great, too)
22:27 Corion Oh - sorry - talking of tickets - the release manager guide does not mention how to access http://rt.perl.org and where to get the credentials
22:27 rjbs I don't want to kill any forward motion, but it doesn't all need to happen in blead.
22:27 rjbs Corion: That's a really good point.
22:27 rjbs Corion: I've put "update the RMG for rt.perl.org access" on my todo from this meeting.
22:27 Corion I didn't need it as DaveM closed my perlbug test-ticket, but ...
22:28 rjbs I don't think anyone here has any code / features in need of review, so I think we can skip that item.
22:28 rjbs I think that's all I have, of substance.
22:28 Abigail Jesse gave me right on rt.perl.org to close tickets, but he did that in 2010, and I don't think I ever tried.
22:28 Corion I have no other questions/remarks
22:28 rjbs Abigail: I'd appreciate if you'd have at least a brief look at the open blockers sometime in the next 2 weeks;  they're all hung off of...
22:29 rjbs 5.16.0 blockers: https://rt.perl.org/rt3//Publ​ic/Bug/Display.html?id=94978
22:29 rjbs I don't expect you to go fix them (but hey, if you want to...) but maybe you will have a flash of insight.  Also, there are only about ten.
22:29 Abigail Will do this weekend.
22:30 rjbs Abigail: I'll check the RT permissions status.  I'm pretty sure Jesse would get it right, though. :)
22:30 Abigail ;-)
22:30 rjbs Abigail: Anything else I can tell you or do for you to make your duty period as enjoyable as possible?
22:31 Abigail I cannot think of anything right away, but I know how to reach you when I hit a wall.
22:31 rjbs Indeed!
22:32 rjbs Okay.  I have our pre-release follow-up scheduled for two weeks from today.  That's March 15th.
22:32 Abigail My rought plans for the next weeks:
22:32 rjbs Same time (for you, Abigail), one hour earlier for me.
22:33 Abigail This weekend, catch up on p5p. Go to the list of blockers. Get an idea of which issues need to be solved before the release, and which ones can be postponed. Write a message to p5p.
22:33 Abigail Next week: keep monitoring p5p closely, specially the smokes.
22:34 Abigail Weekend of 10/11: go through all the motions mentioned in the RMG and produce a release, omitting any uploads or pushes.
22:35 Abigail And then do the real thing on the 20th.
22:35 rjbs Sounds excellent.  And if you hit problems on the 10/11th, you'll have plenty of time, including our meeting on the 15th, to see it sorted out.
22:36 rjbs So: hopefully no last minute drama.
22:36 rjbs Okay, and my to-dos from this meeting are:
22:36 rjbs - zlib updates in smoke-me branch
22:36 rjbs - meta ticket for utf-8 bugs
22:36 rjbs - update the RMG for rt.perl.org access
22:37 Abigail I've changed my work schedule so that I don't have to go to the office on the 20th; I can quit doing $WORK stuff around 16.00, giving me 8 hours to do the real release. (Or 9, and release just before midnight GMT)
22:37 rjbs Excellent.
22:37 rjbs In my experience, the most horrible parts were perldelta and corelist updates.  Both have apparently been greatly improved since I last did a release.
22:37 rjbs So I hope you will enjoy smooth sailing.
22:37 rjbs s/most/only/
22:38 Abigail I spend a few hours earlier this week thinking up an epigraph. I've a short list of 3 possibilities.
22:38 rjbs Yeah, that's the other hard part. :)
22:38 Corion :)
22:38 Abigail At one moment, I thought "maybe I just do 12 releases, so I don't have to choose".
22:39 rjbs Well, I will be happy to schedule you for two more dev releases so you can get all three from that list out there.
22:39 rjbs (The July slot is open, in fact.)
22:39 rjbs but maybe I'll save asking you about that until 5.15.9 is behind you
22:40 Abigail July would not work for 2 of the 3 possibilities.
22:40 rjbs July 20 is OSCON, anwyay
22:40 Abigail Right. And I should be there.
22:40 rjbs so who wants to spend OSCON time releasing perl instead of enjoying Portland?
22:41 Corion Maybe somebody not at OSCON :)
22:41 rjbs Okay.  I think we've covered everything on my list.  If there's nothing else, I will bang my pretend gavel and send us all on our way.
22:41 Corion *ding*
22:41 Abigail Ok, I'm done.
22:41 Abigail Thanks.
22:41 Corion Thanks!
22:41 rjbs Corion: If you want to do July, that can be arranged. ;)
22:42 rjbs Abigail: Thanks very much for your time!
22:42 rjbs Corion: ...and for yours!
22:42 rjbs See at least one of you again, here, in 2wk.
22:42 Corion Oh - I'm doing some meetup in August, I don't think I'll have the mind for July :)
22:42 Abigail Yeah, and I'll talk to you on Saturday.
22:43 rjbs Yup!  Time to go be social with the family.  Tschuß!
22:43 Corion Bye
22:43 Corion left #onionsketch

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