Camelia, the Perl 6 bug

IRC log for #parrot, 2009-02-26

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:00 Tene NotFound: could the makefile just use ../../parrot_config instead?
00:00 * Tene not big on makefile-fu
00:00 NotFound Tene: yes, but I want to let open the possibility to build it in other places
00:01 Tene kk
00:09 AndyA joined #parrot
00:27 dalek parrot: r37010 | jkeenan++ | branches/deprecate_tqueue:
00:27 dalek parrot: Creating deprecate_tqueue in https://svn.parrot.org/parrot//branches
00:27 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37010/
00:31 cotto Do I need to make any Makefile modifications if one PMC depends on another's header?
00:39 chromatic Yes.
00:40 cotto Care to elaborate?
00:42 chromatic It's easy.  Add that header to the rule for the PMC's .o file after the colon.  That's where dependencies go.
00:42 chromatic I'm looking for an example.
00:43 chromatic $(SRC_DIR)/call/pcc$(O) isn't a PMC, but it's an .o file with a PMC header dependency.
00:43 chromatic src/pmc/null$(O) is a better example.  It depends on the Default PMC's header.
00:44 cotto chromatic++ #helping me not break -j
00:49 cotto What needs to be edited to update that part of the Makefile?
00:49 chromatic config/gen/makefiles/root.in
00:49 purl config/gen/makefiles/root.in is what generates parrot/Makefile
00:49 chromatic ... I think.  There may be something in one of the Config libs though.
00:51 cotto It looks like the code is smart enough to look at the .pmc files and dtrt.
00:52 cotto confirmed
00:52 chromatic That's good.
01:12 dukeleto joined #parrot
01:15 ilia joined #parrot
01:16 dalek parrot: r37011 | NotFound++ | trunk/examples/embed/lorito.c:
01:16 dalek parrot: [examples] add -e option to lorito
01:16 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37011/
01:19 ilia so i'm running make spectest on the latest code and i'm getting:
01:19 ilia Scope not found for PAST::Var '@INC'
01:20 ilia is this bad?
01:20 pmichaud it's unexpected.  (I'm assuming you mean in Rakudo.)
01:20 ilia oh yes Rakudo
01:20 pmichaud Someone else reported a similar problem earlier.
01:20 ilia it was working a few commits ago
01:21 pmichaud you might need to update your version of Parrot with --gen-parrot
01:21 pmichaud (assuming that's the way you obtained parrot in the first place)
01:21 ilia i got parrot from svn
01:21 pmichaud sometimes the head version of parrot will break Rakudo.
01:22 pmichaud or perhaps your version of parrot isn't recent enough
01:22 ilia i compiled parrot HEAD
01:22 NotFound pmichaud: there is warning building rakudo on head that looks like a pontential problem
01:23 NotFound And makes it fail buildng with c++
01:23 ilia :(
01:23 pmichaud NotFound: okay.  Rakudo doesn't guarantee that it works against Parrot head.
01:23 pmichaud We only aim to work with whatever version of parrot is in build/PARROT_REVISION
01:24 NotFound I've been unable to find the problem because of pmc2c messing with line numbers.
01:26 ilia ok i will try parrot 0.9.1
01:26 pmichaud ilia:  better is to simply download rakudo, and then run   perl Configure.pl --gen-parrot
01:26 pmichaud that always gets the "correct" version of Parrot for whatever copy of Rakudo you have.
01:26 ilia pmichaud: i can do that for head rakudo?
01:27 pmichaud ilia: yes.
01:27 TiMBuS joined #parrot
01:27 ilia interesting
01:27 ilia is buildbot still running the tests?
01:31 ilia i'm setting up a Hudson job for Rakudo... is that useful to anyone
01:38 dalek rakudo: 2cf2dfe | pmichaud++ | README:
01:38 dalek rakudo: Update README with latest news on building Rakudo.
01:38 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/2​cf2dfeb1812179f7eb4ddcbebb0dfaaed5c0151
01:38 shorten dalek's url is at http://xrl.us/behkez
01:43 ilia dalek: approved :)
01:43 ilia wait, is dalek a bot too?
01:44 pmichaud yes, dalek is a bot
01:44 ilia haah
01:44 pmichaud dalek reports on updates to parrot, rakudo, websites, etc.
01:45 pmichaud afk for a while.
01:46 dalek rakudo: e6b7133 | pmichaud++ |  (3 files):
01:46 dalek rakudo: Update some references to Parrot, Perl, and/or Rakudo.
01:46 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/e​6b71337a1859bdd79df943c138eb3d438c349f1
01:46 shorten dalek's url is at http://xrl.us/behkfw
01:48 ilia i am still getting Scope not found for PAST::Var '@INC'
01:48 ilia using the `perl Configure.pl --gen-parrot` method
01:49 pmichaud ilia: what platform, compiler, etc.?
01:49 ilia os x, perl 5.8.9
01:50 pmichaud what revision number does "parrot/parrot_config revision" say?
01:50 ilia 37000
01:50 pmichaud you might also need to do "make realclean" and run the perl Configure.pl line a second time (it won't rebuild parrot, so it should go quickly)
01:53 pmichaud if that doesn't help, then perhaps report the details to rakudobug@perl.org
01:55 ilia damn still persistent
01:57 pmichaud I'm guessing it must be an osx issue, then.
01:57 pmichaud do you get the error with 'make test'?
01:58 ilia ye[
01:58 ilia yes
01:58 pmichaud very odd.
01:58 ilia i didn't get these errors just a few hours ago
01:58 ilia it started happenning after a git pull
01:59 dalek parrot: r37012 | jkeenan++ | branches/deprecate_tqueue (9 files):
01:59 pmichaud does it happen for all tests or just specific ones?
01:59 dalek parrot: Implement deprecation of tqueue PMC as per https://trac.parrot.org/parrot/ticket/192.
01:59 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37012/
01:59 ilia pmichaud: all
01:59 pmichaud that's very bizarre.
02:00 pmichaud I can't find anywhere that @INC is really being used in a PAST::Var
02:01 pmichaud well, I need to run -- file it as a rakudobug and hopefully someone else can take a closer look at it.
02:03 Theory joined #parrot
02:03 chromatic Maybe there was a conflict, like when you get a .orig and .rej file in a patch fuzz problem.
02:04 ilia git conflict?
02:04 ilia u suggesting i git clone again?
02:05 chromatic Or just look for any local modifications.
02:05 ilia i checked git status, nothing special
02:08 chromatic That was my only other guess.
02:14 chromatic I can't reproduce that, as of a recentish Parrot and current Rakudo.
02:16 ilia i am trying to reproduce again
02:16 ilia from the beginning
02:18 Andy joined #parrot
02:19 ilia ye here it is again
02:20 rg must be an osx thing. i can't reproduce it either with current parrot, current rakudo.
02:21 ilia i sent a bug
02:23 ilia how do i roll back to a revision in git
02:26 ilia git reset?
02:29 chromatic I don't know.
02:29 ilia looks like `git reset --hard f1f722edd9ee6cccfbf7cb0fbbe08188b7b8a382`
03:07 nopaste "rg" at 93.104.124.78 pasted "compare error message to queried message instead of fixed in os.t" (18 lines) at http://nopaste.snit.ch/15722
03:07 pmichaud generally I just do a new clone when I want to make sure I have a clean copy.
03:07 rg chromatic: could you please look at http://nopaste.snit.ch/15722 and possibly commit it. it fixes a test failure on solaris.
03:08 pmichaud for those looking for a rakudo release tonight -- it's likely not going to happen until tomorrow.  My cold is getting steadily worse :-(
03:08 ilia pmichaud: get some sleep :)
03:08 pmichaud rest is becoming more important than release :)
03:08 ilia pmichaud: try oregano oil
03:25 jimmy joined #parrot
03:26 chromatic rg, that test fails that way on Solaris?
03:29 rg it fails because the message on solaris is 'Not owner' (which i find a bit strange, but that's solaris for you).
03:29 rg the errno is the same (EPERM)
03:29 cotto A good hair stylist can help you with that.
03:31 rg i guess it would also fail if you run the test with a different locale on any platform that has localized error messages, which would be fixed by the patch aswell.
03:31 chromatic Okay, good to know.  I *thought* I had the same error text as that from link() in the OS PMC, but that uses strerror now.
03:32 rg that *is* the link test
03:32 chromatic I mean, I read the PMC when I wrote that test.
03:33 chromatic How do we know that the errno for that is *always* 1 across platforms?
03:33 rg oh you mean the message was fixed in the pmc?
03:33 chromatic Yes.
03:33 Theory joined #parrot
03:34 rg we don't know that. if you know how to get to EPERM, that would be preferable.
03:35 chromatic We probably need to generate them as constants somehow.
03:35 janus joined #parrot
03:39 chromatic That sounds like a configure probe.
03:40 rg does perl have a list or errno constants? otherwise you'd have to parse errno.h
03:40 rg sounds almost like it might be easier with c
03:41 chromatic It's almost trivial in C.  Just write out a .pasm file.
03:41 dalek parrot: r37013 | jkeenan++ | trunk/PBC_COMPAT:
03:41 dalek parrot: Wordspaces used where hard tabs were needed to delimit columns; change to hard tabs.
03:41 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37013/
03:46 dalek parrot: r37014 | jkeenan++ | branches/deprecate_tqueue/PBC_COMPAT:
03:46 dalek parrot: As in trunk, change multiple wordspaces to hard tabs to delimit columns.
03:46 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37014/
03:58 chromatic rg, I worked around it in a different way.
03:59 chromatic I'll file a ticket that we need the POSIX errno constants generated somehow.
04:01 dalek parrot: r37015 | chromatic++ | trunk (2 files):
04:01 dalek parrot: [PMC] Improved the "Can't hardlink a directory unless you're root" test of the
04:01 dalek parrot: OS PMC by not matching locale and OS-dependent strerror() output for EPERM
04:01 dalek parrot: (reported by Rolf Grossmann).
04:01 dalek parrot: This test may fail if you run it as root.  Try not to do that.
04:01 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37015/
04:03 rg ah, you coded back a string to match.
04:05 chromatic I figured the error message for that could be clearer.
04:06 rg there is actually an Errno.pm in perl.
04:07 rg could we use that to generate a list of constants or should we already try not to introduce new dependencies on perl.
04:07 rg ?
04:10 chromatic It's probably better not to generate new dependencies on Perl.
04:10 chromatic We want to remove them eventually.
04:15 cotto It's going to be a project to get Perl out of the build process.
04:22 rg ok, i now have a 76 line perl script to generate a 328 c file that will generate a 101 line pasm file with the errno values. any takers for anything?
04:23 rg the fixed os test works. thanks chromatic++
04:30 rg ok, time to get some sleep. i'll check the logs ;)
04:48 jimmy left #parrot
05:22 elmex_ joined #parrot
05:32 dukeleto joined #parrot
05:37 masak joined #parrot
05:44 Theory joined #parrot
05:52 eternaleye joined #parrot
06:04 Tene_ joined #parrot
06:49 cotto seen kj
06:49 purl kj was last seen on #parrot 8 days, 5 hours, 59 minutes and 3 seconds ago, saying: good night all  [Feb 18 00:50:14 2009]
06:52 Fayland_logger joined #parrot
06:56 * cotto crosses fingers that his upcoming huge commit doesn't break anything
06:56 szabgab joined #parrot
07:00 dukeleto joined #parrot
07:01 dalek parrot: r37016 | cotto++ | trunk (32 files):
07:01 dalek parrot: [PMC] replace PMC_struct_val and derivatives with ATTR accessors for the Sub PMC
07:01 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37016/
07:09 dalek parrot: r37017 | cotto++ | trunk/include/parrot/sub.h:
07:10 dalek parrot: [.h] minor tidying
07:10 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37017/
07:20 uniejo joined #parrot
07:52 dukeleto joined #parrot
07:53 dukeleto ilia: get reset HEAD^
07:55 GeJ cotto: well, build and test after a realclean didn't show any sign of b0rkness.
07:58 namenlos joined #parrot
08:08 jimmy joined #parrot
08:15 jimmy cotto++, all passed on mswin32
08:21 mikehh joined #parrot
08:25 cotto That's good to know.  Thanks jimmy.
08:25 cotto I'll sleep well.
08:28 cotto Although I generally sleep well when I break the build, too.
08:33 jimmy good night.
08:40 mikehh Build ok at r37017
08:40 mikehh smoke http://smolder.plusthree.com/app/pu​blic_projects/report_details/18353 reports 100%
08:41 shorten mikehh's url is at http://xrl.us/behmc2
08:41 mikehh however t/pmc/exceptionhandler.t (Wstat: 11 Tests: 6 Failed: 0)
08:42 mikehh Parse errors: Bad plan.  You planned 8 tests but ran 6
08:42 mikehh I have been getting this for a few days
08:43 jimmy yes, it's 6
08:47 jimmy mikehh: please create a ticket for it.
08:49 mikehh ok - got to take the kids to school - will do so whrn I get bavk
08:49 mikehh back
09:21 mikehh t/pmc/exceptionhandler.t  passed with 8 tests at r36930 but only 6 at r36946
09:31 jimmy mikehh: welcome back, could you create the trac ticket?
09:33 mikehh doing that now
09:35 jimmy mikehh++
10:05 dalek rakudo: 2758f79 | (Moritz Lenz)++ | tools/ (2 files):
10:05 dalek rakudo: [tools] use fake executable in autounfudge and test_summary
10:05 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/2​758f79032dc8a55cd98f1d7d6bfd2df0010abfa
10:05 shorten dalek's url is at http://xrl.us/behme6
10:21 dalek rakudo: e4bd268 | (Moritz Lenz)++ | t/spectest.data:
10:21 dalek rakudo: we now pass sub-calls.t, add to spectest.data
10:21 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/e​4bd26861ad165fe00376e25a73c5ca898025309
10:21 shorten dalek's url is at http://xrl.us/behmfk
10:43 jimmy left #parrot
10:55 mikehh joined #parrot
11:44 AndyA joined #parrot
11:45 bsdz joined #parrot
11:47 bsdz an issue buliding parrot from svn with MSVC. after running Configure.pl, nmake gives "fatal error U1086: inference rule cannot have dependents". any one come across something similar?
11:53 jimmy joined #parrot
11:53 jimmy bsdz: could you create a ticket for it?
11:53 bsdz i'll try
11:59 gaz joined #parrot
12:10 magnachef joined #parrot
12:15 PacoLinux joined #parrot
12:25 masak joined #parrot
12:28 PacoLinux joined #parrot
12:32 rg1 joined #parrot
12:34 PacoLinux joined #parrot
12:37 cas joined #parrot
12:42 particle joined #parrot
12:49 PacoLinux joined #parrot
12:55 AndyA joined #parrot
13:02 PacoLinux joined #parrot
13:08 PacoLinux joined #parrot
13:15 PacoLinux joined #parrot
13:18 bsdz left #parrot
13:23 PacoLinux joined #parrot
13:29 PacoLinux joined #parrot
13:41 particle joined #parrot
13:46 gryphon joined #parrot
13:49 PacoLinux joined #parrot
13:57 PacoLinux joined #parrot
14:01 jimmy joined #parrot
14:10 namenlos joined #parrot
14:13 mikehh joined #parrot
14:15 bsdz joined #parrot
14:17 PacoLinux joined #parrot
14:31 PacoLinux joined #parrot
14:44 PacoLinux joined #parrot
14:53 PacoLinux joined #parrot
14:59 jimmy_ joined #parrot
15:12 Andy joined #parrot
15:32 dalek rakudo: d0739a3 | (Patrick R. Michaud)++ | docs/spectest-progress.csv:
15:32 dalek rakudo: spectest-progress.csv update: 314 files, 7041 passing, 0 failing
15:32 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/d​0739a3ec44be9485ae82cfc8d778fc767691252
15:32 shorten dalek's url is at http://xrl.us/behm3b
15:45 dalek rakudo: e9a06c1 | (Andy Lester)++ | README:
15:45 dalek rakudo: Added explanation of how to get Rakudo without git. Minor punctuation fixes.
15:45 dalek rakudo: Signed-off-by: pmichaud <pmichaud@pobox.com>
15:45 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/e​9a06c1f89c8ab7e60be4427e8afa8f0a8bf5ee0
15:45 shorten dalek's url is at http://xrl.us/behm4i
16:16 dalek rakudo: 68b39d2 | pmichaud++ |  (2 files):
16:16 dalek rakudo: Merge branch 'master' of git@github.com:rakudo/rakudo
16:16 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/6​8b39d2a82a5d9de6139efa3b4f86b458029a047
16:16 dalek rakudo: d46df75 | pmichaud++ | build/Makefile.in:
16:16 dalek rakudo: Add some more dependencies into the Makefile.
16:16 shorten dalek's url is at http://xrl.us/behm6v
16:16 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/d​46df759561844219d18a6135c123660c8fb760d
16:16 dalek rakudo: f0a9a5d | pmichaud++ | docs/compiler_overview.pod:
16:16 dalek rakudo: Some minor documentation updates.
16:16 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/f​0a9a5dbf743584c416b059b6d4c5169bb6a1a20
16:16 shorten dalek's url is at http://xrl.us/behm6x
16:16 shorten dalek's url is at http://xrl.us/behm6z
16:16 dalek rakudo: d2cd0fb | (Andy Lester)++ | CREDITS:
16:16 dalek rakudo: updated my info. Normalized fields in some others
16:16 dalek rakudo: Signed-off-by: Moritz Lenz <moritz@faui2k3.org>
16:16 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/d​2cd0fb8185b61b3b094621fdbd92fa9c614efe0
16:16 shorten dalek's url is at http://xrl.us/behm63
16:16 dalek rakudo: 8b9b21e | (Andy Lester)++ | build/Makefile.in:
16:16 dalek rakudo: config/ directory should not be checked with critic
16:16 dalek rakudo: Signed-off-by: Moritz Lenz <moritz@faui2k3.org>
16:16 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/8​b9b21e4a32b97250de0dec2a96537c54878c071
16:16 shorten dalek's url is at http://xrl.us/behm65
16:16 dalek rakudo: c173544 | (Andy Lester)++ | build/gen_junction_pir.pl:
16:16 dalek rakudo: Check return code of close
16:16 dalek rakudo: Signed-off-by: Moritz Lenz <moritz@faui2k3.org>
16:16 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/c​1735441f8b055b12923d30447c121043145da2a
16:16 shorten dalek's url is at http://xrl.us/behm69
16:16 dalek rakudo: af9452e | (Andy Lester)++ | t/harness:
16:16 dalek rakudo: use lexical filehandles, not global; Check for .git dirs as skippable; modernizations
16:16 dalek rakudo: Signed-off-by: Moritz Lenz <moritz@faui2k3.org>
16:16 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/a​f9452ea8c56ddce01364a0b508ad7ed51c898cd
16:16 dalek rakudo: 1951708 | (Moritz Lenz)++ | t/harness:
16:16 shorten dalek's url is at http://xrl.us/behm7b
16:16 dalek rakudo: [t/harness] remove variable $recurse
16:16 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/1​951708d1d6d0e635285457ba96020f6dd5ac6a8
16:16 shorten dalek's url is at http://xrl.us/behm7d
16:16 dalek rakudo: 5e410bb | pmichaud++ |  (4 files):
16:16 dalek rakudo: Merge branch 'master' of git@github.com:rakudo/rakudo
16:16 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/5​e410bbf6ecad37a965b2d49d57c4eae8979fe6a
16:16 shorten dalek's url is at http://xrl.us/behm7f
16:17 pmichaud moritz: you _are_ running tests prior to accepting the commits, yes?
16:17 moritz pmichaud: yes; which is why I added 1951708d1d6d0e635285457ba96020f6dd5ac6a8
16:18 moritz I didn't run full spectests because no actualy compiler changes were made
16:18 moritz but I verified that reconfigure + make + make spectest work
16:18 pmichaud okay.  We particularly need to check that reconfigure with --gen-parrot works, since it's the official build sequence.
16:18 moritz I put them into a branch first (merge_petdance) and tested the bunch of them together
16:21 moritz actually the integration branch feature is quite neat
16:21 pmichaud yes, I'll probably play with that shortly.
16:22 pmichaud when do commits disappear from the fork queue?
16:22 pmichaud when they're applied to a branch, or when they merge back to master?
16:22 moritz the former
16:22 pmichaud so, if you apply a commit to an integration branch, but then abandon the branch, the commit is "lost" from the queue?
16:23 moritz yes
16:23 pmichaud hmmm, that might not be so good.
16:23 moritz not 100% sure though
16:24 pmichaud we might end up deciding not to use the fork queue.
16:31 pmichaud is there a way to get a list of files being managed by git?
16:31 pmichaud i.e., to know which files are part of the repo?
16:32 pmichaud ah, git ls-files
16:36 Tene joined #parrot
16:45 Theory joined #parrot
16:56 khisanth_ joined #parrot
16:56 chromatic joined #parrot
17:03 namenlos joined #parrot
17:16 allison joined #parrot
17:16 * allison wading her way through 700 OSCON proposals
17:19 davidfetter pick mine! ;)
17:23 Andy allison: Are you following the Rakudo naming?
17:23 Andy I'm so excited.
17:23 pmichaud allison probably isn't on #perl6  :-)
17:23 Andy Each code name for each Rakudo name will be a city with a Perl Mongers.
17:23 contingencyplan joined #parrot
17:24 Andy and PM groups get to duke it out who gets naming rights.
17:24 * davidfetter lobbies for Oakland.pm
17:24 Andy davidfetter: You have a while. :-)
17:24 allison Andy: sounds great!
17:24 davidfetter are we talking about the squared circle for this duking business?
17:24 Andy davidfetter: One can only hope.
17:24 Andy Austin and Bratislava are already pegged.
17:24 Andy But C is up for grabs.
17:24 Andy Chicago and Copenhagen can duke it out.
17:24 davidfetter "two mongers enter. one monger leaves!"
17:25 PerlJam sounds like a good thing to "auction" at an appropriate perl conference too.
17:25 Andy OOOH
17:25 pmichaud darnit, merged to the wrong branch.
17:26 pmichaud apparently moritz' changing of the integration branch changes it for everyone.
17:27 moritz that's... unexpected
17:30 ilia joined #parrot
17:31 pmichaud okay, how do I get andy's commit into master now?
17:32 Andy either from the fork queue, or
17:32 pmichaud it's not in the fork queue anymore.
17:32 pmichaud because it got mistakenly applied to a branch.
17:32 pmichaud (i.e., not master)
17:32 Andy oh
17:32 Andy you can cherry-pick
17:32 Andy git cherry-pick c98a526f893eeb7336dac91df66a8cf06e97e8f4
17:32 pmichaud so, I can't do it through github.
17:33 Andy don't think so
17:33 Andy dunno.
17:33 Andy and then do git commit -cc98a526f893eeb7336dac91df66a8cf06e97e8f4
17:33 Andy to use the same commit message
17:33 pmichaud pain.
17:33 pmichaud maybe I'll just re-add.
17:34 pmichaud or maybe you can re-request the pull.
17:34 Andy but I think that may mean dupe changes on different commits
17:34 Andy which means if you do that, it'll be a conflict when I get it.
17:34 pmichaud I haven't committed it to master yet.
17:34 pmichaud so you shouldn't be seeing it yet.
17:34 Andy ok
17:34 pmichaud and git is supposed to be able to manage such things, iirc
17:35 pmichaud how do I drop a branch in github?
17:36 PerlJam git branch -d branchname  # (assuming you mean delete)
17:36 pmichaud that deletes it from github?
17:36 moritz and then git-push origin branchname:
17:36 pmichaud with the colon?
17:37 moritz yes
17:37 Andy pmichaud: Maybe I misunderstand what you meant by "re-add" then.
17:37 pmichaud Andy: I was just thinking about abandoning your push and doing a new one from the source.
17:37 moritz no, it's :branchname
17:37 moritz sorry
17:37 pmichaud just a sec, let me see if I can clean this up.
17:37 moritz http://github.com/guides/remove-a-remote-branch
17:37 Andy pmichaud: But if you do that, then when you push that back, I have a conflict with mine.
17:37 Andy which is fine, I guess.
17:39 Andy I just have to do something to undo my commit.
17:39 Andy and it IS on master right now.
17:39 Andy it's in my FQ at least.
17:39 ilia hey guys, the Scope not found for PAST::Var problem only occurs on the 5.8.9 perl from macports. I see no such problem with apple's stock perl
17:39 ilia which is 5.8.8
17:40 pmichaud ilia:  oh, I see -- it's a Test::Harness issue then.
17:41 pmichaud Andy: when I look at http://github.com/rakudo/rakudo/tree/master and follow the links into docs/  I don't see the file there.  So I think it's not in master yet.
17:41 Andy strange.
17:41 purl But true.
17:41 ilia now how do i update my bug? http://rt.perl.org/rt3/Publi​c/Bug/Display.html?id=63506
17:42 pmichaud ilia: reply to the email you got back from rt
17:42 ilia ok
17:46 Tene pmichaud: so, you said that the AST that should be generated for try {} should be the same as { ...; CATCH {$! = $^ex}} right?
17:47 pmichaud Tene: I think that's what S04 says, I can't be certain of it.
17:47 Tene That's the approach you want me to try?
17:48 pmichaud sure.
17:48 Tene Okay.
17:53 ilia so here's the result of my efforts from yesterday... rakudo spectest suite on Hudson http://twitpic.com/1pgu1
17:54 Tene What's Hudson?
17:54 purl Hudson is, like, better
17:54 Tene Oh.  Helpful.
17:54 Tene purl: go lart yourself.
17:54 purl Tene: sorry...
17:55 ilia i have Hudson set to poll git every hour and build and spectest on any changes
17:55 pmichaud for anyone who wants to try my first cut at a tarball for rakudo:  http://www.pmichaud.com/perl​6/rakudo-200902-test.tar.gz
17:56 * Coke apparently missed a chance for Albany.pm; gypped.
17:57 pmichaud I haven't cut the release yet.
17:57 pmichaud I can go with Albany.
17:57 rdice joined #parrot
17:57 ilia rdice: hi Richard
17:57 rdice ilia, hello.
17:58 ilia rdice: coming to tpm?
17:58 Coke pmichaud: kidding. Albany.pm is all but dead.
17:59 rdice ilia: indeed I am.
17:59 Coke tpm?
17:59 purl tpm is The Perl Machine, a hardware platform designed for compiling and running Perl programs. or Transaction Process Monitoring software or The Phantom Menace. or The Poor Movie or Toronto Perl Mongers or Trusted Platform Modulle, a DRM for MacOS or Tivoli Provisioning Manager (or how to screw a data centre automatically)
17:59 ilia coke: Toronto Perl Mongers
17:59 * Coke wonders if it's worth thinking about going to oscon.
17:59 Coke ilia: I guessed, given what I know of rdice.
18:00 rdice Coke:  I'm sure it's worth thinking about.  Thinking is cheap. :-)
18:00 Coke danke.
18:00 Coke rdice: sadly, nothing else is.
18:00 Coke I'm not even sure I can afford to go to pittsburgh this year.
18:00 rdice Hm.  My sense of self-respect?
18:00 davidfetter oscon is harder for me to get to in san jose than portland
18:00 davidfetter which is a sad commentary on the state of public transit here in the sf bay area :P
18:00 * davidfetter lives in oakland
18:01 Infinoid despite my helping to fund BART with those bridge tolls
18:01 rdice I think it's an interesting change.  I wonder why they did that.  Portland seems to have worked out pretty well the past few years.  Maybe The Great Recession has brought down prices on conference space in San Jose and they figure they'll get more attendance this way.
18:02 rdice davidfetter:  I thought public transit in the bay area was pretty good for american metro areas.
18:02 pmichaud I think the decision for San Jose pre-dates the recession.
18:02 pmichaud I could be wrong about that.
18:02 davidfetter rdice, that caveat leaves a pretty huge hole :P
18:02 rdice pmichaud:  there was a "before the recession"?  I can't remember that far back.
18:02 pmichaud sure, I remember it.  Clinton was president.
18:02 rdice davidfetter:  One I intentionally inserted. :-)
18:03 Infinoid The bay is *big*.  I think San Jose is an hour farther away from me than Oakland is
18:03 pmichaud afk, lunch
18:03 ilia so who is http://twitter.com/0280cbb8710f7b7 ? anyone here
18:04 rdice infinoid:  I know it well. :-)  It's about the same physical scale and population as Toronto.  They're remarkably similar in a lot of ways.
18:05 Infinoid So given that I'd already be driving from the east edge of the state, it turns a 3.5 hour trip into a 4.5 hour one
18:06 Infinoid This is why I don't go to San Jose very often...
18:07 particle joined #parrot
18:07 alvar joined #parrot
18:10 Tene Hey, is anyone here near San Francisco?
18:17 Psyche^ joined #parrot
18:41 Coke f
18:41 Coke b
18:41 Coke b
18:43 Coke wow. that was bizarre.
18:43 Coke (managed to, I think, /background/ my screen session.)
18:47 * ilia slaps Coke around a bit with a large trout... or smth like that
18:50 Theory joined #parrot
19:06 barney joined #parrot
19:29 mberends joined #parrot
19:52 dalek parrot: r37018 | NotFound++ | trunk/examples/embed/lorito.c:
19:52 dalek parrot: [examples] add --warnings option to lorito
19:52 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37018/
19:56 dalek rakudo: f4d2486 | pmichaud++ | README:
19:56 dalek rakudo: More README improvements.
19:56 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/f​4d2486ba121ab204b9bbbad8bd15a75c28e2ad2
19:56 dalek rakudo: 78023fa | pmichaud++ | build/Makefile.in:
19:56 shorten dalek's url is at http://xrl.us/behn2b
19:56 dalek rakudo: Makefile targets for creating MANIFEST and a release tarball.
19:56 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/7​8023fa519072f1e3057eb2ef9bb799c263589ad
19:56 dalek rakudo: b412361 | pmichaud++ | .gitignore:
19:56 shorten dalek's url is at http://xrl.us/behn2d
19:56 dalek rakudo: Ignore any MANIFEST file.
19:56 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/b​41236166d0d1dae758ced0893a267f4b061776d
19:56 shorten dalek's url is at http://xrl.us/behn2f
20:13 ilia what is the process for pushing a branch of rakudo?
20:13 ilia or should i fork
20:26 pmichaud we're still working out the process.
20:26 pmichaud producing diffs and submitting to rakudobug@perl.org   is still valid, though.
20:33 ilia i'm having trouble with producing a diff even
20:33 ilia git diff master shows a change i made but not a new file I added
20:40 gryphon joined #parrot
20:45 ilia see it now
20:45 ilia i had to git rebase origin master
20:52 ilia anyone else getting read errors for github.com
20:59 mberends joined #parrot
21:00 moritz ilia: git-format-patch works well if you did a commit first
21:04 jsut|work wouldn't one easy way be to pull the repo, create a local branch, code away, then use git diff master?
21:05 jsut|work if it's been a long time you can git checkout master; git pull, get rebase master mybranch; to update your branch to current
21:05 ilia nice thanks
21:12 cotto Is there any reason the [GS]ETATTR accessor macros are in curly braces rather than a do{}while(0); ?
21:15 chromatic Doubt it.
21:16 pmichaud chromatic: would appreciate your advice on Rakudo numbering scheme I'm thinking of (more)
21:16 cotto I'll commit the change then if it doesn't cause any test failures.
21:16 pmichaud we were going to use letter-based names:  A B C D ...
21:16 pmichaud Andy suggested that we use .pm cities as release names
21:17 pmichaud I like that, and modified it to be alphabetical city names
21:17 pmichaud but I'm now thinking perhaps we'll just number sequentially, starting with #14
21:17 pmichaud so "Rakudo #14 ('Chicago')"
21:17 pmichaud next month would be #15
21:17 pmichaud etc.
21:17 pmichaud I didn't want to use #1, #2, #3 etc because they're too easily confused with 1.0, 2.0, 3.0
21:18 pmichaud but starting with 14 shouldn't invite that confusion
21:18 purl joined #parrot
21:18 pmichaud it also emphasizes that we've had around 13 releases prior to this one.
21:18 pmichaud thoughts?
21:18 purl Moonlight shines through the dark night / clouds move overhead, casting shadows / dancing in the firelight
21:18 moritz pmichaud: I'm fine with that
21:19 pmichaud that way .pm groups that want to support/associate themselves with Rakudo don't have to wait months or years to do so
21:19 chromatic pmichaud, that seems fine to me.
21:20 pmichaud of course, we can also identify a release by date ("February 2009") -- in fact, I'm thinking that the tarball will be named rakudo-2009-02.tar.gz
21:21 chromatic I like that.  There's no magical thinking involved.
21:22 ilia rakudo-2009-02 Chicagos
21:22 ilia rakudo-2009-02 Chicago Monsters
21:23 szbalint finally a use for Budapest.pm
21:23 szbalint ;]
21:25 Coke chromatic: speaking of magical thinking, are you and allison on the same page regarding our releases yet?
21:25 jonathan szbalint: Sorry, B is taken. :-P
21:25 chromatic No.
21:26 szbalint jonathan: aww :)
21:26 chromatic I don't like calling two releases a year "stable".
21:26 jonathan szbalint: As a condolence, it's by a city just up the river from it. ;-)
21:26 szbalint jonathan: by Bratislava or Boston?
21:26 jonathan Bratislava, of course! :-)
21:26 szbalint ah, so the former then
21:26 chromatic That's like your wife asking "Do you think I'm pretty?" and you saying "Your friend is cute."
21:26 Coke chromatic: Ok. If allison doesn't want to discuss that on list (c.f. "poisonous"), IWBNI the two of you could take it out and at least present a concensus of two.
21:26 chromatic We tried.  There's no consensus.
21:27 chromatic She believes that Ubuntu, for example, will only ever package 1.0 and 1.4 (or 2.0 and 2.6).
21:27 NotFound I'd like to name releases with the unix time in hex X-)
21:28 chromatic I say "If our software is better this month than it was last month (and we have over two years of history demonstrating that we can achieve that), let's release a new version."
21:28 chromatic She says "Normal people won't upgrade to 1.1, 1.2, 1.3, 1.5, etc."
21:28 chromatic I say "They have the option."
21:28 chromatic She says "We'll spend extra time fixing bugs before 1.0, 1.4, 2.0, and 2.6."
21:28 Coke they won't upgrade if we say they aren't stable.
21:28 chromatic I say "How does that make 1.1, 1.2, 1.3, 1.5, etc less stable?"
21:28 Coke ... so it seems like it's kind of self fulfilling.
21:28 NotFound I think "normal people" mean "package managers"
21:29 chromatic What makes them "not stable" though?
21:29 chromatic Do we not run tests for them?
21:29 Coke chromatic: I don't know what the rationale is.
21:29 Coke I'm just trying to have a plan.
21:29 chromatic Do we drop quality for the release right after the one where we spent a lot of time improving quality?
21:29 Coke s/have a/know the/
21:29 chromatic That's what I don't get.
21:29 dalek parrot: r37019 | cotto++ | trunk/lib/Parrot/Pmc2c/Attribute.pm:
21:29 dalek parrot: [PMC2C] wrap GETATTR/SETATTR macros in a do{}while(0); instead of curly braces
21:29 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37019/
21:29 chromatic Why are we calling some releases stable and some releases not stable if we try very, very hard to keep trunk always stable?
21:30 chromatic cotto, someday I'd like to see GETATTR macros usable as rvalues.
21:30 Coke perhaps this would be a good topic for the board to address.
21:30 NotFound Y think that a 'stable' label only makes sense if someone will provide updates with just bug fixes to it.
21:31 cotto If someday were soon, that'd make the conversion to ATTRs a little easier.
21:31 chromatic I agree.  I don't want to do that.  I think it's a great way to waste volunteer resources.
21:31 chromatic cotto, it also helps static analysis tools not see so many "used uninitialized" problems.
21:31 NotFound Well, distro package managers can do, if they see a need.
21:31 Coke chromatic: seems like this falls under the 2nd bullet point at http://www.parrot.org/foundation
21:32 Coke we could even put something up to a vote.
21:32 chromatic NotFound, that was my point.  Ubuntu's update and release and stability policy has absolutely no bearing on me as an upstream developer.  That's pretty much kind of exactly the reason why downstream exists.
21:32 Coke I do think it is is confusing that 1.1 might have /radical/ changes from 1.0
21:33 NotFound Let's radical!
21:33 Coke whereas the 'magical thinking' crowd would expect that to be a minor update.
21:33 chromatic If downstream reports a bug, I'll fix it.
21:33 chromatic I'll tell them which commit fixes it.
21:33 chromatic I'll tell them which new release will contain the fix.
21:34 Coke but just because 1.1 might be a big change from 1.0, that doesn't a priori mean it's unstable.
21:34 NotFound It even can (and even must) be a lot more stable
21:34 chromatic Given that we (by intent and deliberate choice) only merge feature branches to trunk when they're stable, I'm not sure why trunk would ever be unstable.
21:36 Andy I would really like them to be #14, #15
21:36 Andy because then we're not tied to alphabetical
21:36 Andy and we can have more PM involvement.
21:36 Coke I thought our PM quit!
21:36 NotFound Poor Monk?
21:36 cotto I thought "stable" referred to the API, not the code, i.e. there would only be major API changes or deprecations at 1.0, 1.4, etc.
21:37 Infinoid Partly Mortal
21:37 pmichaud Andy: I'm pretty sure we're going with #14, #15
21:37 chromatic "stable" is just too loaded a word.
21:37 Coke cotto: agreeing on terminology would be nice; what does "development" mean, then?
21:38 cotto That's what we do.
21:38 pmichaud Andy: because I agree I don't want to be tied to alphabetical.
21:38 NotFound We can call them "enhanced pro" releases ;)
21:38 Andy pmichaud: YAY!
21:38 Coke cotto: what is a "development release" ?
21:38 purl a "development release" is clearest
21:38 Andy I want to get people fighting over naming!
21:38 Infinoid Yay for PM groups duking it out!
21:38 pmichaud well, I think we've already got Feb and Apr spoken for.
21:38 Andy Chicago!  Boston!  Boca Raton!
21:38 ilia sounds like you guys are going to spend another year over release naming
21:38 Andy ilia: Not at all.
21:38 Coke ilia: parrot or rakudo?
21:39 pmichaud parrot, yes.
21:39 pmichaud rakudo has its house pretty much in order :-)
21:39 szbalint naming-- # let's just use sha256 sums for release names, versions ;-)
21:39 ilia we need perl6 yesterday!
21:39 pmichaud we're all excited about our naming convention :-)
21:39 NotFound NAMECON?
21:41 Coke parrot doesn't care what the name of a release is. We're arguing about something more fundamental.
21:41 chromatic Basically we're arguing over two philosophies.
21:41 cotto We need to do what the College of Cardinals does when electing the Pope: lock all interested parties in a room until a decision is reached.
21:41 Infinoid with a bowl full of candy
21:41 chromatic One is "Users should be able to upgrade at their own pace."
21:41 NotFound cotto: What we will burn for the 'fumata'?
21:42 chromatic The other is "Developers should be able to make changes at their own pace."
21:42 cotto Infinoid, or cookies.
21:42 Coke chromatic: given the volunteer nature of our work, the latter seems more sane.
21:42 Infinoid cotto: I'm openminded to the prospect of cookies.  Definitely pro-cookie here.
21:42 Infinoid Hmm.  Users probably don't care about parrot, per se.  It all comes down to what the HLs depend on
21:43 chromatic Allison cares a lot about downstream packages, and stability there.
21:43 chromatic I don't.
21:43 chromatic Hang 'em.
21:43 NotFound Infinoid: that's a good point. Will be a shame if several languages need different parrot releases
21:43 Infinoid NotFound: it'll be fine as long as we get our installer right
21:43 Coke NotFound: shouldn't be ... what Infinoid said.
21:43 pmichaud (different parrot releases)  I don't expect that to be an issue.
21:43 NotFound Infinoid: not so fine for distro packagers
21:44 pmichaud it should generally be "release xyz or later"
21:44 Coke (distro packages) - I think the HLL packages are more interesting from that standpoint.
21:44 pmichaud and that's part of the point of the longer publication cycle -- we work at 6-month intervals there instead of monthly ones.
21:44 pmichaud in the short term, though, 6-months often isn't frequent enough.
21:45 Infinoid The idea of 6-monthly rakudo releases is a bit shocking to me.  You guys are advancing at a *furious* pace
21:46 chromatic The HLL point didn't come up in our discussion yesterday, actually.
21:46 chromatic That's a good point.
21:46 NotFound "xyz or later" will be good, provided that we take some care to allow it.
21:47 pmichaud yes, the purpose of the 6-month cycle is also so that hll developers can publish HLLs to a known Parrot release point
21:47 pmichaud at least, that's how I understood it.
21:47 pmichaud It's simply that Rakudo, and possibly other languages, likely won't be able to live with Parrot 1.0 for some of their audiences
21:47 chromatic I don't see it as quite that way.
21:47 Coke pmichaud: but all releases are known release points.
21:47 Coke so why can't they target 1.2 if they want?
21:47 chromatic The six month cycle is easier for us to prioritize features.
21:48 Coke chromatic: but our prioritization there has no direct correlation to reality.
21:48 pmichaud Coke: a particular HLL could target 1.2 if they wish, yes.
21:48 pmichaud Coke: but if there are multiple HLLs sharing a common Parrot, that may not work.
21:48 pmichaud Coke: Because 1.2 might not have some features that 1.0 provided.
21:48 chromatic A month is too short, and a year is too long.
21:48 chromatic The real question no one has addressed is "What do we tell people who want bug fixes?"
21:48 Coke pmichaud: we can't really assume how HLLs interacting on parrot is going to look.
21:48 NotFound We can't talk about language interoperability if it becomes normal that each language requires his own release. Well, we can, but nobody will believe
21:49 pmichaud Coke/NotFound:  I agree, which is why I'm not worrying about it until the July release anyway.
21:49 pmichaud interoperability isn't really a goal until then.
21:49 chromatic Coke, it has some degree of correlation with reality.  We haven't figured out our load average so that we can refine our predictions.
21:49 pmichaud by July things may have settled down within Parrot that it's not really an issue.
21:50 chromatic We'll be better at scheduling for 1.4, and better yet for 2.0.
21:50 Coke chromatic: i remain skeptical.
21:50 chromatic Plus we should remember that we don't really have any users right now.
21:50 Coke we've had 8 years so far, and still suck at it.
21:50 Infinoid chromatic: If the HLLs always depend on the once-every-6-months-and-som​ehow-magically-more-stable releases, and those are our deprecation points, and the HLL depends on something that's about to be deprecated, then that means HLL users can *never* upgrade parrot to fix a bug without breaking the HLL at the same time.
21:50 Coke (no users) you /WOUND/ me.
21:50 chromatic How about "no users who aren't also Parrot developers"?
21:50 pmichaud "to fix a bug"  -- those are "bugfix releases"
21:51 chromatic I don't want "bugfix releases".
21:51 Coke but we're not doing bugfix releases in the traditional sense.
21:51 pmichaud chromatic: I can agree to that sentiment.
21:51 pmichaud but I think allison has been aiming in the direction of "bugfix releases"
21:52 pmichaud i.e., we'd issue bugfix releases for the Jul/Jan releases
21:52 chromatic She thinks we'll probably have to have them for "critical bugs" and "security fixes".
21:52 chromatic In other words, if we find a big bug in December, we have to make a 1.4.1 release.
21:52 pmichaud Infinoid: I think the expectation (wrongly or rightly) has been that HLL users would only upgrade parrot at the six month intervals
21:52 chromatic Possibly also a 1.0.1 release, if Ubuntu hasn't updated to 1.4.
21:52 chromatic (To which I said "No.  That's nuts.")
21:53 pmichaud as opposed to HLL developers, which would probably be working with more recent versions of Parrot
21:53 pmichaud (in expectation of preparing for the next Parrot six-month release)
21:53 NotFound We will have a naming problem with bugs found in November X-)
21:53 pmichaud Personally, I'm not advocating any particular model yet, because I think we'll end up making up a bunch of it as we go.
21:54 chromatic I think so too.
21:54 Coke I'd be happy if we had a general plan. Right now we have two plans. :|
21:54 pmichaud to me that argues for a simplistic naming/numbering convention that isn't overloaded with lots of other meanings.
21:55 Coke (other meanings) even our simplistic numbering scheme now implies things are untrue.
21:55 Coke "that are"
21:55 pmichaud our numering scheme now isn't my idea of "simplistic"
21:55 pmichaud depending on which scheme you're referring to.
21:55 chromatic I suggested we refrain from labeling any release as "stable" or "development" or more stable and see what happens.
21:55 chromatic She said "People already call 1.0 and 1.4 stable."
21:55 isop joined #parrot
21:55 chromatic So STOP DOING THAT and I'll give you candy.
21:55 chromatic Or chips.  Or crisps.  Whatever you told Matthew.
21:56 pmichaud cookies won.  :-)
21:56 NotFound Mentos with coke
21:56 pmichaud *diet coke*
21:56 NotFound Ofcoz
21:56 chromatic What does our current numbering scheme imply?
21:56 * pmichaud wonders if Coke has ever had a Mentos, and what the impact was :-)
21:57 NotFound Unstable
21:57 * purl wobbles
21:57 Coke chromatic: 1.1 vs. 1.0 with a 2.0 on the horizon implies that 1.1 is an incremental update to 1.0
21:57 Coke and that 2.0 will be the first release that includes a bunch of new features.
21:58 Coke I know you and I know it doesn't mean that.
21:58 pmichaud Coke: and this is exactly why I've been arguing against the 1.x 2.x scheme.  :-)
21:58 NotFound We can use @ instead of dots, looks more modern ;)
21:59 ilia 2@0 ?
21:59 * pmichaud proposes &circ;
21:59 NotFound Yeah
21:59 chromatic If I had a better numbering scheme, I'd suggest it.
21:59 ilia NotFound: crazy
21:59 NotFound ilia: I am
22:00 NotFound ilia: look at pirric, for example %-)
22:00 Coke if this were a coding project, I'd say lets figure out what problem we're trying to solve, and then a solution will become obvious.
22:00 Coke or at least attainable.
22:00 cotto We could go surrealist.  Instead of 1.0 we could release elephant.flower
22:00 ilia so is there a product manager
22:00 chromatic The problem is "How do we indicate that a release in February is newer than a release in January?"
22:00 Coke ilia: not in any sense that you would recognize from a for-$ project.
22:00 Coke chromatic: that's YOUR problem.
22:00 Coke that's not allison's problem.
22:01 Coke so I think we need to agree on the problem we're trying to solve if we're going to agree ona solution.
22:01 Coke "apparently not alli..."
22:01 ilia Coke: maybe it's time to treat parrot like a commercial product
22:01 NotFound First we need to agree that there is a problem
22:01 ilia parrot and rakudo
22:01 Coke ilia: Ok. give me 3 million dollars.
22:01 pmichaud "Parrot XP" :-)
22:01 pmichaud "Parrot 2007"
22:01 pmichaud "Parrot ME"
22:01 pmichaud "Parrot Vista"
22:01 pmichaud ...that sort of commercial product?  ;-)
22:01 NotFound "Parrot Enhanced Pro for Enterprise Environment"
22:02 Coke or we could use oracle or sun's number schemes.
22:02 chromatic What other kind of information *can* you extract from a version number though?
22:02 chromatic "Should I test it if I upgrade?"
22:02 chromatic "Is it different from the previous version?"
22:02 chromatic "What exactly has changed?"
22:02 ilia pmichaud: how about OS X style productizing
22:02 chromatic "How important are the changes, to me?"
22:03 NotFound I think that the information you must extract for a version number is just what web page to look for inromation
22:03 NotFound information
22:03 purl i guess information is similar, but not the same
22:03 chromatic Yeah, it's just an identifier for a particular release that sorts nicely.
22:04 Coke (that sorts nicely) - do we agree that we're (except in emergencies) going to have a linear set of releases?
22:04 NotFound chromatic: 1 & 2 answers: always yes
22:04 chromatic Yes.
22:05 Coke chromatic: then we don't need the major/minor numbering at all (except for emergencies)
22:05 chromatic Except that Allison says that no one will take us seriously if we jump from Parrot 1 to Parrot 2 too soon.
22:05 chromatic I don't agree, mind you, but she believes that strongly.
22:05 Coke chromatic: if we cared about people taking us seriously...
22:06 Coke no one that I know takes us seriously now.
22:06 chromatic Yeah, I've said what I've had to say about only caring about grown-up opinions.
22:06 NotFound Foolest proposal of the day: Complex version numbers: Parrot 1.4 + 3.5i
22:06 Tene We could just drop numbers completely!  "Parrot Awesome!"  "Parrot ITSREALLYGOODIPROMISE"
22:07 Tene "Parrot DontUseThisOne"
22:07 mberends what is wrong with something like 2009.02 like Gentoo and Ubuntu?
22:07 chromatic Predictable version numbers *do* have meaning, though.  They tell you what's newer and what's older.
22:07 chromatic mberends, nothing's wrong with that, as far as I can see -- but people who want biannual Super Stable releases don't like that.
22:07 Tene mberends: Then the version information doesn't also carry metadata about the release itself.
22:08 pmichaud Tene: I'm not sure the currently adopted system carries metadata either.
22:08 Tene Doesn't distinguish, in the version number itself, between "REALLY STABLE" and "Kinda sorta stable-ish"
22:08 Coke the only metadata that we're willing to give (it seems) is whether or not it is a "milestone" release.
22:08 Coke and right now, we have to do that by marking them somewhere else.
22:09 Coke but the ONLY thing we're using that to track are items in DEPRECATION.pod
22:09 pmichaud and "milestone release" is defined by a date, not by a set of features.
22:09 chromatic But what's the value in exposing milestone releases to end users?
22:09 Coke so let's just track that in DEPRECATION.pod
22:09 Coke "eligible for removal after the 2009.07 release".
22:10 chromatic In other words, the milestone releases are merely deprecation points?
22:10 * pmichaud is certain he's heard that somewhere before.  :-)
22:10 pmichaud (the "eligible for removal..." part)
22:10 Coke chromatic: based on my reading of you, yes.
22:10 Coke chromatic: based on my reading of allison, no.
22:11 Coke but I like that plan.
22:11 chromatic I like that plan too.
22:11 chromatic It removes a lot of magic from the plan.
22:11 Coke magic--
22:12 chromatic pmichaud?
22:12 purl i think pmichaud is http://www.pmichaud.com/ or "Patrick R. Michaud" <mailto:pmichaud@pobox.com> or in charge of toaster experiments
22:12 Tene But... but maybe I want a feeling of pride every time I see someone using my proposed numbering scheme!
22:12 Coke gotta run, but will try to review.
22:12 * Tene stops trolling #parrot to go troll his students instead.
22:12 pmichaud since I was arguing for "identify releases by date" at PDS, I think I'm the choir here.  :-)
22:13 chromatic Just wanted to get you on record again.
22:13 mikehh If the   product passes the test suite - surely that is stable?
22:13 chromatic mikehh, that's right.
22:14 NotFound Note that, as in TAP, we can have a plan but lots of skip
22:14 chromatic NotFound, I'm trying to remove as many skips as possible... would appreciate assistance from anyone.
22:15 mikehh but to be stable it must pass the full test suite
22:15 NotFound chromatic: I unTO an old TODO yesterday
22:15 chromatic Great!
22:15 Whiteknight joined #parrot
22:16 chromatic prove -r --directives t/ 2> directives.log is a big help
22:22 NotFound That TODO was just that someone just don't knowed how to build a key
22:24 mikehh that starts out with the banchmark tests
22:34 dalek rakudo: cdcb92c | pmichaud++ | docs/NEWS:
22:34 dalek rakudo: Initial version of NEWS for Rakudo.
22:34 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/c​dcb92cc5cf87dd0fa231561b762bbfe14386421
22:34 shorten dalek's url is at http://xrl.us/behonh
22:47 Theory joined #parrot
22:52 dalek rakudo: f45384a | pmichaud++ | docs/release_guide.pod:
22:52 dalek rakudo: First version of release_guide.pod, based on a proposed version
22:52 dalek rakudo: from (Andy Lester)++ .
22:52 dalek rakudo: review: http://github.com/rakudo/rakudo/commit/f​45384ae464a271a4a81c1098c9bb67d555cd9d1
22:52 shorten dalek's url is at http://xrl.us/behorb
22:55 Andy pmichaud: My first overhual is on my branch
22:55 Andy s/branch/fork/
22:56 pmichaud checking.
23:02 rurban joined #parrot
23:02 pmichaud Andy: did you do a pull request?
23:03 Andy no
23:03 pmichaud ah.
23:03 rurban hi folks. I believe pbc is now platform compatible, but confirmation for 64bit appreciated: https://trac.parrot.org/parrot/attachm​ent/ticket/254/tt254-64bit-pbc.tar.gz
23:03 shorten rurban's url is at http://xrl.us/behosi
23:03 Andy it's 343c28e05a36d42b2a6de4248b4fb74c6a78a35f
23:03 rurban seen rg
23:03 purl rg was last seen on #parrot 18 hours, 33 minutes and 15 seconds ago, saying: ok, time to get some sleep. i'll check the logs ;)
23:03 pmichaud hmmm... that's a bit longer than I was planning... certainly longer than what we've done in Parrot's NEWS
23:04 rg i'm here
23:04 pmichaud we'll go with your version for now, though.
23:04 rurban msg rg please test http://xrl.us/behosi
23:04 purl Message for rg stored.
23:04 rurban rg: I think I've found the 64bit compat problems
23:05 rg so much that we won't even need -xmemalign ?
23:05 rurban rg: But I have no 64bit machine to test.
23:06 rurban rg: no sorry. -xmemalign will go to the new hints file, not yet done. I fixed the more general pbc compat problem
23:06 Andy ok, now I did.
23:06 rg i can test it on amd64 pretty quit
23:06 Andy pmichaud: It's longe,r but it's more readable.
23:06 Andy Ther'es more to see, rather than a block of text.
23:06 pmichaud yes, I agree.
23:07 Infinoid rurban: testing now
23:07 pmichaud We can shorten later if we decide it's too long.  For now it's good as longer, readable text.
23:07 Andy One of the problems with the regular releases is that they're not interesting in themselves.
23:07 rg but i guess you mean big endian
23:07 Andy So when we put out Parrot, who cares?  It's just another release.
23:07 rurban struct ptr_alignment on sparc is not possible to solve for now.
23:07 Andy WHY is it interesting?
23:07 Andy Same thing with Vienna14.
23:07 chromatic One of the features of regular releases is that they're just not interesting in themselves.
23:07 rurban rg: no be<>le is no problem at all. 64<->32 bit was the big problem
23:07 pmichaud well, part of the reason for the shorter version of NEWS is that it's easier to cut-and-paste into a release announcement.
23:08 Andy chromatic: I understand that, so we need to pick out the good stuff to tell people.
23:08 pmichaud We don't want the release managers to have to go through a lot of editing work to cut a release.
23:08 Andy pmichaud: The way around that is for someone else to write the release announcement
23:08 Andy such as we had happen here.
23:08 pmichaud "good stuff to tell people" belongs in the release announcement, not in NEWS
23:08 chromatic "Here is a cool feature AND YOU CAN USE IT TODAY RIGHT NOW IN A NEW STABLE RELEASE!"
23:08 Andy well, I guess I don't know what's this, what's that.
23:08 Andy I don't know what you see as release announcement vs. NEWS
23:09 pmichaud release announcement is what gets sent to use.perl, rakudo.org, mailing lists, etc.
23:09 Andy chromatic: I'm not dissing reuglar releases at all.
23:09 Andy And you don't see NEWS as an aggregation of those announcements?
23:09 pmichaud an aggregation of the highlights, sure.
23:09 pmichaud more like an rss feed than a document.
23:10 Infinoid rurban: the patch doesn't apply to latest HEAD, and the pbc tests don't seem to do anything with the unpacked pbc files
23:10 Andy I guess I don't see that  NEWS is too heavy by aggregating the release announcmenets.
23:10 Andy that's what I thought I was doing.
23:10 Andy chromatic: the downside of Just Another Parrot Release is that nobody cares unless we tell them why to care
23:10 rurban Infinoid: yes, I was away from the net for a few days. Tommorrow I'll update and make a clean patch to apply to latest.
23:11 Andy I'm not going to run an announcement on Perlbuzz about Parrot #97 just because it comes out.
23:11 Infinoid rurban: Cool.  Looking forward to giving you a pass result on gentoo/amd64 :)
23:11 rurban Infinoid: But the new pbc tests (sparc 64bit) are handled as _6 in integer.t and number.t
23:12 chromatic Would you run an announcement about an awesome new feature?
23:12 Andy Yup
23:13 Andy And so if someone hands me a release announcement that says "This has this new feature", then I'll run that.
23:13 rurban Infinoid: I'll also finish the missing floatval converters in the meantime. This should be done quick. Maybe even 4-byte single float, since its so easy to add
23:13 Andy So when I declined to run a 0.9.1 announcement, and I explained that it wasn't interesting, the reply came "I didn't realize it needed to be interesting."
23:13 Andy Which is one of the big problems that people too close to the project can have.
23:14 Andy It's interesting to US, but what about to THEM?
23:14 Infinoid Again, for most people, parrot is probably just an HLL dependency
23:14 rurban Infinoid: I had to write a written permission request to enable the VMWARE Virtualization support BIOS setting in my laptop for 64bit testing. No answer yet.
23:14 Andy All of this goes back to my ongoing frustration that FOSS people don't think about marketing their projects.
23:15 rurban I'm leaving bye, thanks!
23:15 Infinoid rurban: Well, I'm always happy to test :)
23:15 Infinoid bye rurban
23:17 bsdz_ joined #parrot
23:18 Andy ok, goin' home.  New rakudo.org up tonight.
23:19 bsdz_ i'm trying to dlfunc an exported. when i try "func = dlfunc lib, "dgesvd_", "itt33p3pp3p3p33"" i get "Parrot VM: PANIC: itt33p3pp3p3p33 is an unknown signature type.". anyone know if that looks wrong?
23:20 isop joined #parrot
23:29 NotFound bsdz_: Are you randomly testing ot that is a real signature of some function?
23:29 bsdz_ it's a real signature from clapack
23:30 bsdz_ i noticed a load of hardcoded sigs in nci.c
23:30 PerlJam bsdz_: add it to src/call_list.txt then
23:30 PerlJam (and recompile)
23:31 NotFound I hope to never have to work in a project with such functions
23:31 bsdz_ ah okay thanks.
23:32 bsdz_ does this mean we can't use any sig without recomp parrot core?
23:32 NotFound bsdz_: if you don't have a nci-jitable platform, yes
23:32 GeJ Good morning everyone
23:32 bsdz_ i recall a colleague once solving a similar problem (dyn sigs) using C++ templates. note sure how though
23:33 bsdz_ is windows a non nci-jitable plat?
23:33 bsdz_ (morning gej)
23:33 PerlJam bsdz_: I don't think we even have an NCI-compiler yet, so no platform is jitable in that regard.
23:34 bsdz_ ha ha
23:34 bsdz_ okay thanks PerlJam
23:34 NotFound PerlJam: we have
23:34 PerlJam NotFound: excellent!
23:34 * purl plays air guitar
23:34 NotFound Linux i386 at least
23:35 PerlJam heh, I was about to ask if it was x86 only  :)
23:35 bacek_ joined #parrot
23:36 bsdz_ what generates call_list.txt? should i edit the progenitor?
23:38 japhb NotFound: NCI JIT is broken, even on Linux i386.  It claims to work, and then does the wrong thing
23:38 TiMBuS joined #parrot
23:38 bsdz_ no probs i'll take a look in makefile
23:38 japhb bsdz_: config/gen/call_list
23:38 japhb (.pm, and the dir)
23:38 bsdz_ thanks japhb
23:39 japhb bsdz_: no problem -- the first version of that was mine, to work around the fact that my OpenGL bindings were needing a BUNCH of NCI sigs that are only discovered at Configure time
23:40 NotFound japhb: last time I checked it was enabled
23:40 rdice joined #parrot
23:40 japhb dunno how much it has been hacked since then -- $day_job has taken me largely out of Parrot work except for answering questions and such
23:41 japhb NotFound: yup.  It's a horrible, horrible lie.  I rant about it semi-regularly, and no one seems to care.
23:41 bsdz_ i'll see if i need to add a load of sigs
23:42 japhb (No, I haven't created new Trac tickets for it.  I haven't had time to get over my frustration with the infrastructure migration.)
23:42 dalek parrot: r37020 | rurban++ | trunk/src/packfile/pf_items.c:
23:42 dalek parrot: pbc-compat only TT #254: add missing fetch_iv inits
23:42 dalek parrot: review: https://trac.parrot.org/parrot/changeset/37020/
23:42 NotFound japhb: all my test programs works fine
23:43 japhb NotFound: run any OpenGL example with and without JIT.
23:43 skv joined #parrot
23:43 japhb Unless major improvements have happened in the last month, they will work perfectly with --jitcapable=0 ... and give blank windows with NCI JIT on
23:47 NotFound Don't remeber if that problem was solved, will check when/if I have time
23:48 japhb NotFound: Seriously, thank you.  I've been feeling really ignored for months on this.  I went through special effort to write variants of my examples to help narrow down the problem, and just couldn't get anyone who understood the internals to care.
23:49 NotFound japhb: maybe will be easier to test and diagnose now that we have some clean embedding examples.
23:50 japhb nod.
23:50 * japhb rebases to look for said embedding examples ...
23:50 NotFound japhb: t/src/emebd.t and examples/embed
23:51 japhb NotFound: roger that
23:51 NotFound Well, they are clean to me, because I wrote them ;)
23:52 rdice_ joined #parrot
23:53 japhb NotFound: heh
23:57 skv joined #parrot
23:58 ewilhelm joined #parrot

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

Parrot | source cross referenced