Perl 6 - the future is here, just unevenly distributed

IRC log for #perl6-soc, 2008-06-04

| Channels | #perl6-soc index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
17:46 pmichaud joined #perl6-soc
19:30 [particle] hi-ho everybody
19:30 moritz_ ping
19:31 Auzon hi
19:31 pmichaud pong
19:32 moritz_ ok, let's start
19:32 moritz_ Auzon: you first please
19:32 Auzon - Currently working on S29. Spent time yesterday reorganizing a few tests and fighting with SVN (it doesn.
19:32 Auzon - Added 25 sprintf tests today. Might add more, but some of them are too system depedent to test or be usable.
19:32 Auzon - S29 lacks a lot of system functionality tests (like signals and concurrency). I think I can do some of them, but I want help testing on Win32 and OS X. Are there any known pitfalls in testing this type of functionality?
19:32 Auzon - Wasn't able to work much yesterday or at all on Sunday/Monday. Friday is also looking bad.
19:32 Auzon EOR
19:32 moritz_ [particle]: you next
19:33 [particle] nothing to report here
19:33 moritz_ pmichaud: you then
19:33 pmichaud Mostly been working on rakudo this week.  Updated the test suite a bit.  I'm thinking of writing a "guide to the test suite" unless someone beats me to it.
19:34 pmichaud the guide would explain the conventions for tests in the suite, as well as the #?impl skip and #?impl todo, etc. syntaxes
19:34 pmichaud end.
19:34 moritz_ * fixed some tests (S02-magical/*, still need more work) * added some tests * improved rakudo's testing infrastructure a bit * watched Auzon++s commits * wrote t/deprecated-syntax.pod * bit of evalbot hackery * thought about whether we can somehow turn commands to evalbot into test (or preliminary tests). Had no good concept, so abandoned the idea
19:34 moritz_ * don't know i I'll be available for our meeting next w eek, $real_life is likely to interfere
19:35 moritz_ queuing 2 questions
19:35 moritz_ Auzon: you asked about portability of I/O and IPC tests
19:35 moritz_ Auzon: I think a good start is mostly ignore platform differences
19:36 moritz_ Auzon: and if you know of them, add skips for them
19:36 moritz_ perl 5's perlipc man page might be a good start
19:36 moritz_ (perl 5 tries hard to overcome as many limits as possible, so it will be hard for Perl 6 to *really* do better)
19:37 moritz_ does that sound sensible?
19:38 Auzon Yes. Is there any good way to get a handle on platform-specific failures, or will that wait until Perl 6 is more implemented?
19:38 [particle] the perl 6 compiler(s) should provide abstraction, so the tests don't need to worry about it (to a point)
19:39 [particle] the spec tests shouldn't need to worry about platform-specific failures, except for native things like 'int' as opposed to 'Int'
19:39 moritz_ aye
19:39 [particle] so for now, use fudge where you see it's needed, and put in good todo/skip comments
19:40 pmichaud I made a remark earlier that I'd like to see skip comments for rakudo also have an associated RT ticket.  doable?
19:40 [particle] auzon: i expect you'll run into situations where no implementations implement what you're testing
19:40 Auzon I already have :)
19:40 pmichaud or that none of the implementations implement it correctly.
19:40 [particle] ...you'll just have to do your best in writing those tests, and hope they're future-proof
19:41 moritz_ pmichaud: does it really make sense to have $N ($N > 50) tickets with "$xy not implemented"?
19:41 [particle] pmichaud: it's totally doable, but i don't want auzon to shoulder that load
19:42 pmichaud moritz_: it might make sense to do that, yes.  In particular, it tells the rakudo developers which features are being fudged.
19:42 [particle] it's easy enough to write a script that looks for /#\?rakudo \d+? skip .*#\d+/
19:42 pmichaud there is that.
19:42 moritz_ pmichaud: ok, I'll have a look at it
19:42 pmichaud never mind then, I retract the RT comment.
19:42 pmichaud (for now, anyway)
19:42 moritz_ ok, two more questions
19:43 pmichaud (I still stand by my comment that I'd like to see "CATCH not implemented" instead of "parsefail", though.  )
19:43 moritz_ actually just one ;)
19:43 moritz_ Auzon: do you expect to work more on the project in the following week?
19:43 [particle] pmichaud: agreed. i'm guilty of many 'parsefail' comments
19:44 [particle] ...mainly because there were so many parse fails on any given line, it couldn't be limited to one thing
19:44 moritz_ I just have no felling how much a gsoc student is expected to work
19:44 [particle] moritz_: figure it as a full-time job for the course of the summer
19:44 pmichaud moritz_: if we assume that GSOC is approximately equivalent to a grad student (and that's what it pays), then standard would be 20 hrs/wk
19:44 [particle] that's full-time for a grad student :)
19:45 moritz_ ;)
19:45 pmichaud what does GSOC pay?  $1500/month?
19:45 moritz_ because currently I have the feeling that we're lagging behind a bit
19:45 [particle] pmichaud: basically, yes
19:45 Auzon I definitely plan to get 20 hours a week, probably more. I have a few days here and there where I can't do anything (plus weekends), but other than that, I plan ~5 hours a day.
19:46 [particle] auzon: do you feel you're on track?
19:46 pmichaud at the univ I was at, $1500/month would be roughly 30 hrs/week.
19:46 pmichaud (might be 25 hrs/week by now.)
19:47 Auzon [particle]: Nearly. Monday wasn't used at all, so other than that, I think I'm fine.
19:47 [particle] is your test-counting utility working well enough?
19:48 Auzon If you are referring to the one I mentioned in last week's report, no. It doesn't tell the difference between reorganized tests and tests I added (and it only includes completely new files).
19:48 [particle] ok
19:49 Auzon I'm thinking of just counting them in commit messages.
19:49 [particle] yes, i think that's a good idea
19:49 moritz_ I think that's sane
19:49 moritz_ especially since you have "plan $count" on top anway
19:49 moritz_ *anyway
19:49 [particle] it can be simply scripted
19:49 moritz_ one more general question: how do we handle requests for tests?
19:50 moritz_ [particle] proposed a ticket system last week
19:50 moritz_ any idea which one to use? or a wiki?
19:50 [particle] pugs uses trac, right?
19:50 moritz_ think so, yes
19:50 [particle] http://trac.pugscode.org
19:51 [particle] we can use http://dev.pugscode.org/newticket
19:51 moritz_ ok
19:51 moritz_ all questions answered for me
19:51 [particle] and add 'Spec Tests' to the component list with trac-admin
19:51 Auzon I used Hiveminder to some success for tracking tasks in a personal project, and you can send requests via email. I haven't used other programs, but I'm willing to give it a try.
19:52 [particle] i can do hiveminder, too (i use it every day)
19:52 Auzon [particle]: That page already has "Perl 6 test suite" on it
19:52 moritz_ does it have some kind of web interface?
19:52 [particle] Auzon++ # i looked too quickly
19:53 [particle] ok, it's set up, so let's use trac
19:53 * moritz_ doesn't know hiveminder at all
19:53 moritz_ ok, trac then
19:53 moritz_ anything else that needs clarification?
19:53 [particle] we'll need a blog post about it
19:53 Auzon Hiveminder is neat, but Trac is already established, so that's preferable.
19:54 moritz_ [particle]: can you do the blog?
19:54 Auzon I have a SVN question
19:54 [particle] sure, tonight or tomorrow
19:54 [particle] ask away, auzon
19:55 TimToady .oO(Perl 6 is neat, but Perl 5 is already established, so that's preferable... :)
19:55 moritz_ TimToady: back to your lurking silence if you please ;)
19:55 Auzon Sometimes I want to modify a file and then move it. SVN complains about this, and I either have to commit it first or use --force. I assume --force does something bad to the history.
19:55 Auzon Is there a way to avoid this?
19:56 moritz_ just use --force, I think
19:56 [particle] use the --force
19:56 Auzon Alright. That's all I have.
19:56 [particle] do you use svn mv?
19:56 moritz_ Auzon: btw intermediate commits aren't that bad either
19:57 Auzon svn mv, yes.
19:57 moritz_ ok, let's call it a week then
19:57 moritz_ thanks everybody for attending, and for the good work
19:57 [particle] ~~
19:58 moritz_ ~. actually ;)
20:03 Auzon hm. Where is the View Tickets  link that the Trac guide talks about?
20:19 [particle] auzon: seem you need to login, and login is broken :(
20:20 Auzon It says I'm logged in though :-/
20:24 [particle] hrmm, i can't find http://trac.pugscode.org/report/
20:24 [particle] (that should be the url)
20:27 Auzon So no one is using trac.pugscode.org for bug reports, correct?
20:28 moritz_ right
20:29 Auzon Are there any other bug trackers for Perl 6?
20:32 TimToady historically, the t/ directory was considered the bug tracker :)
20:32 Auzon Hm, yeah, but that doesn't help much for requesting tests. ;)
20:33 TimToady sure, just put a failing file with a comment out there
20:34 moritz_ TimToady: since so many tests fails that's not really visible
20:36 TimToady white patches in the smartlinked specs were also construed as requests for patches...
20:36 TimToady s:2nd/patches/tests/
21:28 [particle]1 joined #perl6-soc
22:23 Auzon Regarding requesting tests for me to make: How about adding them to t/spec/REQUESTED (or maybe TODO)?
22:23 moritz_ probably a good solution until we have a working RT
22:25 Auzon Of course, we already have t/TASKS...
22:26 moritz_ we could reuse that if somebody updates it first
22:26 moritz_ bed &
22:27 Auzon see you.

| Channels | #perl6-soc index | Today | | Search | Google Search | Plain-Text | summary