Camelia, the Perl 6 bug

IRC log for #parrot, 2011-09-19

Parrot | source cross referenced

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

All times shown according to UTC.

Time Nick Message
00:02 preflex_ joined #parrot
00:09 NotFound ~~
00:10 soh_cah_toa whiteknight: i'm getting a failure for test 5 in t/manifest/03-regenerate-skip.t but i can't quite figure it out. it's the one with the description "No need to regenerate MANIFEST.SKIP"
00:10 soh_cah_toa whiteknight: think you could take a look?
00:14 NotFound soh_cah_toa: I must update the winxed snapshot to give it a release number before freeze.
00:16 soh_cah_toa NotFound: that's fine, just make sure it doesn't bork any tests (you already know that though) :)
00:16 NotFound Ok
00:34 gbacon joined #parrot
00:45 dalek parrot: 698fe27 | jkeenan++ | MANIFEST.SKIP:
00:45 dalek parrot: Rerun mk_manifest_and_skip.pl
00:45 dalek parrot: review: https://github.com/parrot/parrot/commit/698fe275b9
00:47 dalek parrot: d0dcdea | soh_cah_toa++ | tools/dev/resubmit_smolder.pl:
00:47 dalek parrot: Added coda to the end of tools/dev/resubmit_smolder.pl so that t/codingstd/perlcritic.t passes.
00:47 dalek parrot: review: https://github.com/parrot/parrot/commit/d0dcdea00c
01:37 dalek rakudo/nom: 35ec9c7 | Coke++ | t/spectest.data:
01:37 dalek rakudo/nom: track failures, run fudged tests.
01:37 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/35ec9c7286
02:08 soh_cah_toa NotFound: ping
02:10 bluescreen joined #parrot
02:32 soh_cah_toa msg dukeleto you've got mail...read it! :)
02:32 aloha OK. I'll deliver the message.
02:46 soh_cah_toa msg NotFound i'm not trying to rush you or anything but i'd try and update the winxed snapshot soon since the release is this tuesday
02:46 aloha OK. I'll deliver the message.
03:08 dukeleto soh_cah_toa: ?
03:09 soh_cah_toa dukeleto: i sent you an email
03:14 cotto ~~
03:14 cotto soh_cah_toa, I agree with your msg.
03:15 soh_cah_toa cotto: great :)
03:15 cotto good catch.  That's what happens when you throw something together quickly.
03:15 soh_cah_toa lucky for you, i have a patch all ready to push ;)
03:16 cotto soh_cah_toa++
03:19 dalek parrot: 716f204 | soh_cah_toa++ | / (3 files):
03:19 dalek parrot: Changed 'allhlltest' Makefile target to 'all_hll_test' since a) that's the name of the script and b) it's easier on the eyes.
03:19 dalek parrot: review: https://github.com/parrot/parrot/commit/716f204cac
03:25 soh_cah_toa cotto: while on the subject of nitpicky spellings...some targets separate words w/ hypens and others (most) use underscores. how would you feel about changing some of the hyphens to underscores so that everything is consistently separated by the same character? that way i don't have to think "oh darn, do i need a - or _ for this one? let me go check and waste a bunch of time"
03:26 soh_cah_toa not atm, post-release
03:29 cotto soh_cah_toa, my lazy solution is to let ubuntu's tab completion figure it out.
03:29 cotto It's funny what you find when you get in the habit of pressing tab frequently.
03:30 soh_cah_toa indeed
03:30 cotto soh_cah_toa, I can get you access to the ftp server.  What's your public key?
03:30 soh_cah_toa cotto: i'll email it to you, h/o
03:31 cotto or not
03:31 cotto when you get access, let me know. ;)
03:32 soh_cah_toa wait, you are or you aren't able to add my public key?
03:33 cotto I can't get to ftp-osl.osuosl.org
03:33 cotto er, into
03:33 soh_cah_toa that's odd
03:33 soh_cah_toa maybe it's not just my key then
03:33 cotto They probably clobbered the authorized_hosts file.
03:34 soh_cah_toa at such a convenient time too
03:48 sorear cotto: you know osuosl was massively hacked earlier this month and is mostly or completely offline for forensics?
03:49 sorear ah
03:52 cotto sorear, I knew they were hacked.  I didn't know that they're mostly offline.
03:53 benabik Does osuosl host kernel.org?
03:55 sorear Yes.
03:56 benabik Ahhhhhh....
03:58 soh_cah_toa that's so lame
03:59 soh_cah_toa attacking an advocate of foss
04:00 soh_cah_toa was probably a stupid 5ri1pt k1d33 who learned how to use metasploit
04:17 nbrown joined #parrot
04:20 soh_cah_toa msg dukeleto no more worries. i talked to Alan_S and he took care of the issue w/ my ssh keys
04:20 aloha OK. I'll deliver the message.
04:21 rfw joined #parrot
04:21 cotto soh_cah_toa, you have ssh access now?
04:22 soh_cah_toa cotto: yeah
04:22 cotto great.  Can you add my key?
04:29 woosley joined #parrot
04:29 soh_cah_toa cotto: alright, i added it. give it a minute
04:33 soh_cah_toa and the verdict is...?
04:33 cotto trying now
04:33 cotto d'oh.  wrong key
04:33 soh_cah_toa :)
04:35 cotto worked!
04:35 soh_cah_toa so should i add the second key too?
04:35 cotto no need
04:35 soh_cah_toa alright
04:37 soh_cah_toa now all i need for the release are our credentials for wikipedia and c2
04:38 cotto neither need credentials
04:39 soh_cah_toa ok
04:40 * moritz tests rakudo on latest master
04:40 moritz *latest parrot master
04:41 soh_cah_toa damn, i still can't login to soh_cah_toa@parrot.org though
04:47 soh_cah_toa cotto: are you able to login to parrot.org?
04:54 cotto soh_cah_toa, let me check
04:54 cotto soh_cah_toa, the drupal site looks fine
04:54 soh_cah_toa cotto: i mean through ssh
04:54 cotto soh_cah_toa, that one worked fine
04:54 soh_cah_toa :(
04:55 fperrad joined #parrot
04:55 cotto soh_cah_toa, you have to request access to that on an individual basis
04:56 soh_cah_toa cotto: so i need a separate support ticket for ftp-osl.osuosl.org (done) and parrot.org?
04:57 cotto yup
04:57 soh_cah_toa crap, i hope they can add me in time
05:02 soh_cah_toa i guess i'll take care of that tomorrow then
05:10 JimmyZ joined #parrot
05:13 * JimmyZ would like to suggest stopping monthly release, there is no features that's worth with release
05:21 cotto JimmyZ, we'll see how that goes.
05:22 JimmyZ hello cotto, how is M0 ongoing?
05:23 moritz rakudo passes spectests on parrot/master
05:26 plobsing JimmyZ: we do have some changes. prt0, while not a terribly user-visible change, is a change nonetheless
05:27 moritz speaking of M0, I have a rather general question
05:28 moritz it seems to me that M0 is rather slim and minimal
05:28 dalek parrot-libgit2: e1bbadf | dukeleto++ | / (2 files):
05:28 dalek parrot-libgit2: Get closer to working winxed tests; still incurring the wrath of: Null PMC access in invoke
05:28 dalek parrot-libgit2: review: https://github.com/letolabs/pa​rrot-libgit2/commit/e1bbadfe8b
05:28 moritz but users still want a "fat" system somewhere (like with all the built-in ops like split and find_cclass_...)
05:29 moritz where exactly will that "fat" be? in which subsystem?
05:33 cotto JimmyZ, I'm trying to figure a coherent plan in light of the deprecation policy changes and to see how M0 would fit into such a plan.
05:33 plobsing moritz: the idea is that most of parrot's functionality will run on top of m0. (AFAIK) the jury is still out on whether users will have direct access to m0. even if they do, they'll still have access to the layers above.
05:34 cotto I'm also looking closely at other JIT libraries (libjit, lightning, nanojit) to try and learn lessons from them.
05:34 JimmyZ plobsing: I don't think it's woth with monthly  release
05:34 JimmyZ s/woth/worth/
05:35 moritz plobsing: ok, but "on top of m0" doesn't tell me *where*
05:35 plobsing JimmyZ: monthly is a "developper release"
05:35 moritz plobsing: will there be a new, library-like subsytem that implements most of our current features on top of m0?
05:37 plobsing moritz: as much as possible, current abstraction layers will continue to work. winxed, nqp, pir, etc... if you are using those, things should continue to mostly be the same.
05:37 JimmyZ plobsing: and seems that the last week in monthly release is  a pain to developer who wants to commit something to master
05:37 plobsing JimmyZ: you have a change you want to make?
05:37 cotto nqp and winxed are worth preserving
05:38 JimmyZ plobsing: that means, every month, parrot have a pain week
05:38 cotto JimmyZ, branching and merging is cheap in git.
05:38 cotto the freeze (or slush) is only a couple days
05:39 JimmyZ well, I just give a advice
05:39 cotto JimmyZ, I appreciate it.
06:27 SHODAN joined #parrot
06:58 mj41 joined #parrot
07:18 contingencyplan joined #parrot
07:31 woosley1 joined #parrot
07:32 woosley1 left #parrot
07:40 schmooster joined #parrot
07:46 preflex joined #parrot
08:15 lucian_ joined #parrot
08:24 moritz http://blog.llvm.org/2011/09/greedy-​register-allocation-in-llvm-30.html
08:40 dalek Heuristic branch merge: pushed 484 commits to rakudo/nom-exceptions by moritz
08:41 dalek rakudo: d7f45ff | moritz++ | src/core/Cool.pm:
08:41 dalek rakudo: Cool.eval
08:41 dalek rakudo: review: https://github.com/rakudo/rakudo/commit/d7f45ff8cf
09:04 moritz rakudo: say now.perl
09:04 p6eval rakudo 35ec9c: OUTPUT«Instant.new(x => 2008861638745/1526)␤»
09:04 moritz rakudo: say now.perl.eval
09:04 p6eval rakudo 35ec9c: OUTPUT«Method 'eval' not found for invocant of class 'Str'␤  in <anon> at /tmp/xhT_GL7bZb:1␤  in <anon> at /tmp/xhT_GL7bZb:1␤␤»
09:04 woosley joined #parrot
09:32 dalek parrot/mls/sub-profiler: d049fd1 | (Michael Schroeder)++ | frontend/parrot2/main.c:
09:32 dalek parrot/mls/sub-profiler: destroy the interpreter also in the failure case so that profiling data gets dumped
09:32 dalek parrot/mls/sub-profiler: review: https://github.com/parrot/parrot/commit/d049fd18cd
09:32 dalek parrot/mls/sub-profiler: c28d72b | (Michael Schroeder)++ | src/runcore/subprof.c:
09:32 dalek parrot/mls/sub-profiler: hash with the seg+start_offs pointer so that cloned subs just get one entry
09:32 dalek parrot/mls/sub-profiler: review: https://github.com/parrot/parrot/commit/c28d72bcde
09:32 dalek parrot/mls/sub-profiler: 796ddae | (Michael Schroeder)++ | / (3 files):
09:32 dalek parrot/mls/sub-profiler: use Parrot_pmc_gc_register instead of a custom mark function
09:32 dalek parrot/mls/sub-profiler: review: https://github.com/parrot/parrot/commit/796ddaea6c
09:32 dalek parrot/mls/sub-profiler: b715a02 | (Michael Schroeder)++ | / (11 files):
09:32 dalek parrot/mls/sub-profiler: Merge branch 'master' into mls/sub-profiler
09:32 dalek parrot/mls/sub-profiler: review: https://github.com/parrot/parrot/commit/b715a021b3
09:53 JimmyZ joined #parrot
10:01 woosley left #parrot
11:36 Psyche^ joined #parrot
11:55 ligne joined #parrot
12:10 mtk joined #parrot
12:32 JimmyZ joined #parrot
12:38 atrodo =~
12:43 redicaps joined #parrot
12:47 bluescreen joined #parrot
12:50 lucian joined #parrot
12:52 lucian joined #parrot
12:56 particle joined #parrot
13:12 lucian joined #parrot
13:35 whiteknight joined #parrot
13:44 whiteknight good morning, #parrot
13:44 bubaflub morning whiteknight
13:47 ligne hello
13:49 whiteknight good morning bubaflub, ligne. How are you two doing today?
13:52 mls hi whiteknight!
13:52 whiteknight hi mls! I've been looking at your last code push
13:53 mls is that a good or a bad sign?
13:53 whiteknight good. I've been looking, but haven't said any curse words yet!
13:53 PacoLinux joined #parrot
13:53 mls ;)
13:54 moritz 0 WTFs/minute
13:54 mls with the last change, the last ugly part is gone (the mark function)
13:55 atrodo quite fine code, indeed.  we should all be lucky to get 0 WTFs/minute
13:55 ligne whiteknight> good thanks!  as good as any monday is likely to be, anyway :-)
14:02 SHODAN joined #parrot
14:03 dalek winxed: 4816221 | NotFound++ | NEWS:
14:03 dalek winxed: update NEWS
14:03 dalek winxed: review: https://github.com/NotFoun​d/winxed/commit/4816221ab1
14:06 Coke ... didn't I delete NEWS!??!
14:06 moritz in winxed?
14:06 Coke oh, different project. ;)
14:07 Coke Sorry, still have nightmares about it. ;)\
14:07 dalek winxed/version_1_2: 565319d | NotFound++ | winxedst1.winxed:
14:07 dalek winxed/version_1_2: set version number to 1.2.0
14:07 dalek winxed/version_1_2: review: https://github.com/NotFoun​d/winxed/commit/565319d24f
14:08 moritz where do news for the release announcements go now?
14:08 * Coke wonders trying to push his department internally to use git instead of svn.
14:08 Coke moritz: NEWS is dead, long live ChangeLog
14:09 Coke aloha, NEWS?
14:09 aloha Coke: NEWS is good NEWS
14:09 Coke aloha, no NEWS is dead, long live ChangeLog
14:09 aloha Coke: Okay.
14:09 atrodo aloha> NEWS?
14:09 dalek winxed/version_1_2: ab841f1 | NotFound++ | pir/winxed_compiler.pir:
14:09 dalek winxed/version_1_2: update installable compiler to 1.2.0
14:09 aloha atrodo: No clue. Sorry.
14:09 dalek winxed/version_1_2: review: https://github.com/NotFoun​d/winxed/commit/ab841f1205
14:09 atrodo aloha: NEWS?
14:09 aloha atrodo: NEWS is dead, long live ChangeLog
14:10 Coke atrodo: Did I not just say that?
14:10 atrodo Coke> Slow morning.  Sorry.
14:25 atrodo_ joined #parrot
14:28 atrodo joined #parrot
14:32 dalek parrot: 07300ac | NotFound++ | / (2 files):
14:32 dalek parrot: update winxed snapshot to ab841f1205
14:32 dalek parrot: - Operator ~
14:32 dalek parrot: - Tagged RELEASE_1_2_0
14:32 dalek parrot: - Add winxed news to ChangeLog
14:32 dalek parrot: review: https://github.com/parrot/parrot/commit/07300ac896
14:37 NotFound msg soh_cah_toa winxed snapshot in master is ready for the release
14:37 aloha OK. I'll deliver the message.
14:41 dalek winxed: 6c2f71b | NotFound++ | / (3 files):
14:41 dalek winxed: set master to 1.3.devel
14:41 dalek winxed: review: https://github.com/NotFoun​d/winxed/commit/6c2f71b706
14:42 lucian joined #parrot
14:44 particle joined #parrot
14:57 particle joined #parrot
15:07 dmalcolm joined #parrot
15:17 nbrown joined #parrot
15:40 JimmyZ 7812146
15:45 lucian joined #parrot
15:46 davidfetter_ joined #parrot
15:53 mtk joined #parrot
16:40 pjcj joined #parrot
16:50 dalek winxed: eed90d7 | NotFound++ | winxedst1.winxed:
16:50 dalek winxed: rearrange a bit builtin argument emision and get rid of the remainings of the old "predefined functions" naming
16:50 dalek winxed: in comments and identifiers
16:50 dalek winxed: review: https://github.com/NotFoun​d/winxed/commit/eed90d7df2
17:04 pjcj joined #parrot
17:11 plobsing joined #parrot
17:19 SHODAN joined #parrot
17:27 mj41 joined #parrot
17:56 cotto_work ~~
18:08 dukeleto ~~
18:09 benabik joined #parrot
18:13 dukeleto mls: congrats on getting your commit bit
18:13 dukeleto cotto_work: 3.9 is coming fast. And the m0 C implementation still needs a lot of work...
18:14 dukeleto Also, it is pretty cool that the maintainer of git has humble suggestions for our git workflow doc and has forked parrot.git on github.
18:15 cotto_work dukeleto: I'm glad you saw those.
18:17 cotto_work dukeleto: (M0) I'm glad that CM0 has gotten some attention.  I want to rethink a couple aspects of it and have been carefully looking at and comparing other jit libraries to try to learn from them.
18:20 dukeleto cotto_work: good to hear
18:21 dukeleto cotto_work: i haven't had time to look deeply, but thinking about how m0 could interact with something like nanojit could help us
18:21 dukeleto cotto_work: which jit libs are you seriously considering?
18:21 benabik o.
18:21 benabik o/, even
18:22 cotto_work dukeleto: looking at nanojit, gnu lightning and possibly one other (depending on what I can find)
18:22 cotto_work and llvm
18:41 PacoLinux joined #parrot
18:51 whiteknight NotFound: ping
18:53 NotFound whiteknight: pong
18:54 whiteknight NotFound: Timers aren't working without threads. When Parrot is built --without-threads and in the kill_threads branch, the Timer tests fail now. That wasn't true a few weeks ago
18:54 whiteknight NotFound: I know you were cleaning that code, any idea which change you made would be most likely to have that effect?
18:55 NotFound whiteknight: Are you sure it worked? When I touched it I saw several conditionals on threads available.
18:56 whiteknight NotFound: I saw at least one make test PASS on the kill_threads branch. Also, got an email to parrot-dev today from a person saying it stopped working recently with --without-threads
18:57 whiteknight I assumed it was a problem in the branch, but if it happens in master with --without-threads, the Timer cleanups seem to be the most obvious
18:59 NotFound Looks like the problem is simple: we must just skip the test if no threads.
18:59 NotFound Like 4, 5 and 6 do
18:59 whiteknight so Timers don't work at all if there are no threads?
18:59 whiteknight oh, is that a new test?
18:59 NotFound ok 4 # skip No thread enabled
18:59 NotFound ok 5 # skip No thread enabled
18:59 whiteknight okay, I didn't realize that
18:59 whiteknight weird, I didn't think Timers relied on threading
18:59 whiteknight so kill_threads should rip out timers too, I guess
18:59 NotFound The ones that are not skipped are just creating the Timer object.
19:00 NotFound I think that the dependance on threads is an artifact, doesn't look like is using them.
19:01 whiteknight I don't know why that test wouldn't pass
19:01 NotFound Because of conditional compilation.
19:03 NotFound Mmmm... maybe not
19:03 whiteknight yeah, I wasn't seeing much in there that was conditional
19:04 contingencyplan joined #parrot
19:04 schmooster joined #parrot
19:05 NotFound Funny, 4, 5 and 6 pass if I comment out the skip
19:08 NotFound Urgh, looks like sleep is the problem.
19:08 whiteknight sleep?
19:09 NotFound I used sleep 0.2 in that test, and it seems to be ignored without threads.
19:09 whiteknight oh yeah, shoot
19:09 whiteknight I see that now
19:10 whiteknight all this crap is garbage
19:13 whiteknight okay. At least we know what the problem is now. Do we want to skip the test, or try to fix sleep?
19:13 NotFound No time to testing well before the release, skip should be fine.
19:13 whiteknight okay
19:14 whiteknight merging kill_threads is going to be a very happy day for me
19:21 whiteknight I might call that day "wednesday" in celebration
19:21 NotFound Parrot_usleep(unsigned int microseconds) --- but usleep takes a long
19:23 benabik :-(
19:30 NotFound Got it: Parrot_usleep((UINTVAL) time*1000000); --> This casts time to UINT, and then multiplies. WRONG!
19:31 NotFound Parrot_usleep((UINTVAL) (time*1000000));
19:34 NotFound The fix is easy, but we don't have time to check reports is most platforms.
19:41 dalek parrot: 4fe04fa | NotFound++ | t/pmc/timer.t:
19:41 dalek parrot: skip the Timer start/repeat/stop test when no threads, same as the other Timer .../start/... tests
19:41 dalek parrot: review: https://github.com/parrot/parrot/commit/4fe04fa674
19:51 whiteknight nice find
20:00 dalek parrot/NotFound/fix_sleep: 107e860 | NotFound++ | / (2 files):
20:00 dalek parrot/NotFound/fix_sleep: fix Parrot_cx_scheduler_sleep no threads case and unskip timer tests
20:00 dalek parrot/NotFound/fix_sleep: review: https://github.com/parrot/parrot/commit/107e860881
20:19 bluescreen joined #parrot
20:32 dalek TT #2192 created by NotFound++: Parrot_cx_scheduler_sleep is wrong in the no threads case
20:32 dalek TT #2192: http://trac.parrot.org/parrot/ticket/2192
20:38 cotto_work in space, nobody can hear malloc return NULL: http://users.softlab.ece.ntu​a.gr/~ttsiod/stackusage.html
20:45 tcurtis I hope no one does use that code to determine the stack usage of anything important.
20:45 tcurtis It doesn't appear to take into account push or call.
20:49 dalek rakudo/nom: 7563146 | jnthn++ | src/ (3 files):
20:49 dalek rakudo/nom: Switch to having a $! and $/ per routine, not per block. Also consistently use assignment, not binding, for $! and $/.
20:49 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/7563146c33
20:50 cotto_work "Don't use this to send your spaceship in orbit"
20:50 cotto_work not that it was likely
22:04 sorear tcurtis: he said he was on a SPARC processor.  SPARC doesn't have those instructions in that way.
22:05 sorear cotto_work: sounds like that guy never heard of the Robson theorem.  http://sqlite.org/malloc.html#nofrag
22:05 dalek rakudo/nom: e062a4c | jnthn++ | src/Perl6/Metamodel/ (2 files):
22:05 dalek rakudo/nom: Make role Foo[::T] does Bar[T] { ... } work.
22:05 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e062a4cf2b
22:05 dalek rakudo/nom: 178ddf6 | jnthn++ | src/ (2 files):
22:05 dalek rakudo/nom: Refactor so that we create a BUILDPLAN at a per-class level as well as for the whole hierarchy.
22:05 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/178ddf600c
22:06 dalek TT #2193 created by NotFound++: tailcall to Class.new fails with overriden init
22:06 dalek TT #2193: http://trac.parrot.org/parrot/ticket/2193
22:10 jsut joined #parrot
22:19 tcurtis sorear: Ah, I didn't notice that supports SPARC. Nonetheless, the x86 case still seems to be incorrect in that way.
22:20 bluescreen joined #parrot
22:39 dalek rakudo/nom: d9b19d6 | jnthn++ | src/Perl6/Metamodel/MROBasedTypeChecking.pm:
22:39 dalek rakudo/nom: Add missing .^does.
22:39 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/d9b19d682e
22:39 dalek rakudo/nom: e97afdb | jnthn++ | src/core/ (2 files):
22:39 dalek rakudo/nom: Ensure that when we mix in a role, its attributes are initialized appropriately (for example, any defaults are applied).
22:39 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/e97afdb4b6
22:39 dalek rakudo/nom: 9256ee2 | jnthn++ | t/spectest.data:
22:39 dalek rakudo/nom: We now pass S14-roles/parameterized-mixin.t.
22:39 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/9256ee2bb6
22:39 dalek rakudo/nom: b10ef51 | jnthn++ | NOMMAP.markdown:
22:39 dalek rakudo/nom: Remove completed tasks from nommap.
22:39 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b10ef51901
23:04 ttbot joined #parrot
23:30 bluescreen joined #parrot

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

Parrot | source cross referenced