Parrot 3.5.0 "Menelaus" released | parrot.org | Log: irclog.perlgeek.de/parrot/today | #parrotsketch meeting Tuesday 1930 UTC | Code freeze in effect except for test failure fixes
Set by moderator on 14 July 2011.
00:09 kid51 is now known as kid51_at_dinner 00:10 fperrad left
dalek R: 7e11dd4 | (Jay Emerson)++ | / (3 files):
Fixing up little grammar things; found more issues surrounding my

HLL in Fixed*Array objects (not allowing literals). There may be arguments each way. I may have chosen the better but more challenging route.		 00:11
jay ~ taking a break to run to Home Depot. 00:12
00:13 rurban_ joined 00:15 rurban left 00:16 rurban_ is now known as rurban 00:17 Kulag joined
whiteknight temporary files are HARD 00:17
I mean, there are a million ways to make temporary files, and none of them do everything we might want
00:18 lichtkind left
whiteknight I'm thinking we use something like tmpnam(), but that is subject to weird race conditions 00:18
tmpnam() has a number of problems, though there are other versions like tmpnam_s on windows which have fewer problems (but not 0) 00:19
00:21 Drossel left
whiteknight windows has a few better options. _tempnam seems good, or GetTempFileName 00:22
but they rely on the user having environment variables set up correctly, etc
I'll send an email to parrot-dev to get some opinions 00:29
00:34 theory left 00:56 whiteknight left 00:57 kid51_at_dinner is now known as kid51
kid51 soh_cah_toa: ping 00:59
01:14 bubaflub_ joined 01:16 bubaflub left, bubaflub_ is now known as bubaflub
dalek rrot: ecfcf90 | dukeleto++ | / (3 files):
[doc] Prettify our test suite readme by converting to POD so that it will render nicely on Github
 01:49
rrot: 19582f6 | dukeleto++ | t/README.pod:
[doc] Add another internal link
rrot: 08481a6 | dukeleto++ | t/README.pod:
[docs] Convert misplaced markdown links to POD links
 01:53
01:54 theory joined
dalek R: f1406db | (Jay Emerson)++ | / (7 files):
Additional testing and minor tweaks.
 01:59
02:04 theory left
dalek rrot: 2c3be63 | dukeleto++ | t/README.pod:
[doc] Add some actually useful information to our test suite readme
 02:06
rrot: a47aa51 | dukeleto++ | t/README.pod:
[doc] Keep the coding standard wolves at bay
 02:08
soh_cah_toa kid51: pong
dukeleto soh_cah_toa: look how please github.com/parrot/parrot/tree/master/t is now 02:09
s/please/pretty/
02:09 woosley joined
soh_cah_toa there we go. nice 02:10
dalek rrot: 24ebeaf | dukeleto++ | t/README.pod:
[doc] Correct a link to TESTS_STATUS.pod
 02:11
Felipe great
dukeleto soh_cah_toa: now you do the one in examples/ :)
Felipe: hopefully that will be useful to you. I wish those docs were there when I started hacking on the test suite :) 02:12
kid51 soh_cah_toa: I posted in your ticket about the two python files.
soh_cah_toa dukeleto: i can give it a try
kid51: i'm reading now
kid51 must sleep 02:13
Felipe dukeleto: it will certainly be useful :) 02:14
dalek rrot: 46ec5ef | dukeleto++ | docs/tests.pod:
[doc] Linkify
02:15 kid51 left
soh_cah_toa loves to see documentation get worked on 02:15
dalek rrot: 31df094 | dukeleto++ | ports/README.pod:
[doc] Add a readme for the ports/ directory

This file did not get picked up by tools/dev/mk_manifest_and_skip.pl		 02:39
soh_cah_toa dukeleto: just an fyi, i opened up a branch for that - soh-cah-toa/tt-2155 02:40
though, it's so trivial it hardly deserves one
dukeleto msg tcurtis i would really like to see another blog post from you soon, last blog post was June 14th. 02:41
aloha OK. I'll deliver the message.
soh_cah_toa and i'm working on the compilers directory so skip that one
dukeleto msg lucian i would really like to see another gsoc blog post soon, last blog post was June 20th.
aloha OK. I'll deliver the message.
Felipe 'night 02:44
03:07 bubaflub left 03:32 bluescreen left 03:35 benabik left 03:39 Kulag left 03:43 Kulag joined 03:53 Drossel joined 03:55 Kulag left 03:57 benabik joined 04:02 Drossel left 04:04 Kulag joined
dalek rrot/soh-cah-toa/tt-2155: 9902af7 | soh_cah_toa++ | / (2 files):
Added a README.pod file for the "compilers" directory.
 04:13
rrot/soh-cah-toa/tt-2155: b41d57d | soh_cah_toa++ | / (3 files):
Converted examples/README to POD format and updated a lot of the information in it.
rrot/soh-cah-toa/tt-2155: 754491d | soh_cah_toa++ | examples/README.pod:
Changed several C<> tags to F<>.
 04:27
rrot/soh-cah-toa/tt-2155: 2fa333c | soh_cah_toa++ | compilers/README.pod:
Changed several C<> tags to F<> in compilers/README.pod.
rrot/soh-cah-toa/tt-2155: b01bd79 | soh_cah_toa++ | compilers/README.pod:
I made a copy/paste error in the previous commit. Now the C<> tags are F<>.
 04:31
04:48 soh_cah_toa left 04:55 woosley1 joined 04:58 woosley left 05:00 woosley joined 05:01 woosley1 left 05:28 Kulag left 05:34 Kulag joined 05:48 Kulag left 05:53 Kulag joined
dalek imera: 8a80598 | dukeleto++ | Chimera.cpp:
Include the new parrot embed api header
 06:27
imera: cd06c72 | dukeleto++ | Chimera.cpp:
Add the code to actually create a Parrot interpreter. Doesn't compile yet because we haven't told cmake to compile against libparrot yet
imera: f4a3461 | dukeleto++ | CMakeLists.txt:
Actually get cmake to link against libparrot, currently with hard-coded paths
cotto a wild Chimera appears 06:32
dukeleto which uses www.firebreath.org . I couldn't resist. 06:42
cmake is a whole different world 06:46
so. many. build systems.
06:51 fperrad joined
dukeleto linguist knows about us: github.com/github/linguist/blob/ma...s.yml#L602 07:04
cotto pbc isn't quite the same thing as pir 07:07
though I'd like to see source highlighting for binary files
Perl is not designed to implement a low-level VM. 07:32
08:07 woosley1 joined, woosley left 08:11 woosley joined, mj41 joined 08:13 woosley1 left 08:15 rurban_ joined 08:17 rurban left 08:18 rurban_ is now known as rurban 08:25 daniel-s joined 08:39 dod joined
dalek rrot/m0-prototype: 4edb3d1 | cotto++ | src/m0/perl5/m0_ (2 files):
add not op, tighten binary op semantics, improve debugging output
 08:57
rrot/m0-prototype: 6a77691 | cotto++ | t/m0/integration/m0_hash.m0:
add passing test that calculates the hash of a string

It's full of hacks but it passes. Making the hacks unnecessary is a good next step.		 08:58
rrot/m0-prototype: d81470d | cotto++ | src/m0/perl5/m0_assembler.pl:
add a note about a bit of screwiness in the assembler
09:07 preflex left 09:10 preflex joined 09:17 Kulag left 09:19 daniel-s left 09:20 jjore left, Kulag joined 09:22 jjore joined 09:48 mikehh left 09:56 woosley left 10:41 Drossel joined, Kulag left 10:51 bubaflub joined 11:19 JimmyZ joined 11:27 bubaflub left 11:40 Drossel left 11:41 Kulag joined
Felipe morning devs 11:49
:)
11:59 Kulag left 12:00 Kulag joined
JimmyZ morning, Felipe 12:11
12:19 whiteknight joined
atrodo aloha: dalek? 12:20
aloha atrodo: dalek is being slow tonight, methinks or gone
12:20 jay left
whiteknight good morning, #parrot 12:20
atrodo Does anyone know who runs dalek?
morning whiteknight
whiteknight is dalek not here?
dalek: owner? 12:21
dalek: help
atrodo No, entirely different issue with ipfy
whiteknight atrodo: I think sorear might, or might know who
Felipe morning whiteknight 12:22
atrodo apparently, isparrotfastyet hasn't saved a revision since Jun 27. And I need to figure out why and see if I can't get back revision data
whiteknight hello Felipe 12:23
atrodo: weird
atrodo not that weird, i'm sure it's my code, I can see the revisions coming in. I just have no idea why they're not being saved 12:24
sorear> ping
12:38 bluescreen joined
atrodo whiteknight> Do you have permissions to hit the "test" button in the service hooks? 12:39
12:45 bluescreen left
whiteknight for which project, parrot? 12:48
atrodo yes, parrot/parrot 12:49
whiteknight yeah. I'll go tap it 12:50
dalek rrot: 24ebeaf | dukeleto++ | t/README.pod:
[doc] Correct a link to TESTS_STATUS.pod
 12:51
rrot: 46ec5ef | dukeleto++ | docs/tests.pod:
[doc] Linkify
rrot: 31df094 | dukeleto++ | ports/README.pod:
[doc] Add a readme for the ports/ directory

This file did not get picked up by tools/dev/mk_manifest_and_skip.pl
atrodo And now it picked it up. Great... 12:52
12:56 bluescreen joined 13:25 JimmyZ left
dalek rrot: 826fb57 | Whiteknight++ | src/platform/win32/dl.c:
Rework win32/Parrot_dlsym to look in libparrot first, and msvcrt second if the library handle given is null. This is a horrible hack, and it smells terrible.
 13:32
whiteknight at the very least, that commit looks like it fixes tests on my machine. I'm running smoke right now to verify
a clean report should make kid51 happy, at least 13:33
I may need to go take a shower to rinse the memory off me
Coke should update the known test failures for PLATFORMS on windows. 13:39
ah. I'll pull and check results of whiteknight's last commit first. 13:42
whiteknight yeah, that commit definitely fixed nciutils.t on my system 13:45
Coke ... ok, now I cannot even build parrot.
whiteknight debug.t still seems to fail
nopaste "coke" at 192.168.1.3 pasted "parrot won't build, win7/strawberry/msvc" (13 lines) at nopaste.snit.ch/62229 13:46
Coke I had no trouble building on friday.
looks like there's only 2 coding commits since then, one from notfound on sprintf, and this one from you. I'll see if I can bisect. 13:48
how do I do the git equivalent of "svn up -r <xxx>" ?
moritz git checkout <commit> 13:49
(if the commit isn't in the local repo yet, 'git fetch' first) 13:50
Coke danke. 13:51
Coke wonders if the tt bots are doing "make world". 13:52
(to catch more build errors)
14:01 ilbot2 joined
moderator Parrot 3.5.0 "Menelaus" released | parrot.org | Log: irclog.perlgeek.de/parrot/today | #parrotsketch meeting Tuesday 1930 UTC | Code freeze in effect except for test failure fixes
14:02 atrodo joined 14:03 knewt joined, jevin joined, benabik joined, fperrad joined, slavorg joined, pmichaud joined, spinclad joined 14:04 PerlJam joined, PacoLinux joined, jlaire joined, szbalint joined, jtpalmer joined
dalek kudo/nom: f2fda00 | moritz++ | src/ (3 files):
add Str.match(Regex, :p, :c)
 14:05
kudo/nom: 19e3ba8 | moritz++ | LHF.markdown:
add regex-related Str methods to LHF.

Also deduplicate an item which seems to have doubled during a merge conflict
14:06 sorear joined, slavorgn left, slavorgn joined, Psyche^ joined, Psyche^ is now known as Patterner 14:07 Tene joined 14:14 dngor_ is now known as dngor
Coke whiteknight: after doing a bisect and pinning blame on that commit, i tried to build your revision just in case... and the build worked. 14:14
running make test now.
whiteknight okay, awesome 14:15
Coke sorry about the false alarm. wonder if I missed a realclean.
whiteknight it's okay. That platform is way too fragile 14:16
and it clearly hasn't been getting enough love
Coke down to 2 failing test files, 3 failing tests. 14:18
and at least one of those is line ending related.
14:22 JimmyZ joined 14:23 jay joined 14:24 jay left
dalek rrot: d778087 | Coke++ | PLATFORMS:
add recent build info for windows.
 14:25
14:26 daniel-s joined
whiteknight Coke: What test failures are you seeing? Can you nopaste? 14:26
14:26 daniel-s left 14:27 daniel-s joined
nopaste "Coke" at 192.168.1.3 pasted "test failures on win7 for whiteknight++" (10 lines) at nopaste.snit.ch/62243 14:28
Coke the streams one generates a far too long error message to nopaste, but seems line ending related. 14:29
whiteknight okay, debug.t fails on my machine too. Nobody seems to understand that test or the failure, so I might rip it out
the streams.t failure is disconcerting
Oh, NotFound was talking about that the other day
nopaste "Coke" at 192.168.1.3 pasted "debut.t failures on win7 for whiteknight++" (20 lines) at nopaste.snit.ch/62244 14:30
whiteknight yeah, I
Coke the streams.t one is probably due to git configuration.
whiteknight 'm on win7 here, and that's what I see too
Coke so, I'm not really worried about any of these failures ATM. the streams one would be a nice-to-fix.
dalek R: 9b05b6e | (Jay Emerson)++ | STATUS:
Re-thinking my grammar; need to study whiteknight's matrixy

now is having consequences. I'm learning.
14:30 Themeruta is now known as NotFound
Coke but given that it's not present for every developer, I think the "git config" is a likely culprit. 14:31
whiteknight yeah. I suspect strongly that the debug.t failure is meaningless, especially considering the poor state of our debugger
Coke aye.
I think any build information in PLATFORMS from 2009 can be excised.
(or earlier.)
whiteknight good call
Coke will do so. 14:32
NotFound In the remaining windows failing tests I've looked, the problems are in the tests, not in parrot.
They are due to different versions of git, perl and windows console doing different things. 14:33
14:35 daniel-s left
Coke PLATFORMS stripped to 2010+ data. 14:36
dalek rrot: b3bdf8b | Coke++ | PLATFORMS:
Remove all old (2009 or earlier) build information.

Please submit recent build information for this file.
benabik ~~ 14:37
JimmyZ wonders how is 6model on parrot 14:44
dalek rrot: 2bac6ed | Whiteknight++ | t/dynoplibs/debug.t:
SKIP the failing test in debug.t. We don't understand the failure and considering the code involved we probably don't care.
whiteknight JimmyZ: What do you mean?
moritz_ JimmyZ: it's not :-) 14:45
JimmyZ ah, my poor english
Coke no, that's a fine way to phrase it.
14:45 moritz_ is now known as moritz
Coke it just hasn't moved yet. I suspect after the release tomorrow we'll be more likely to see movement. (also after rakudo/nom is "done") 14:45
JimmyZ doesn't know how to describe things correctly by using english 14:46
Coke thinks that sentence is pretty good counter example. ;)
is *a* pretty...
whiteknight I've been waiting for rakudo/nom to get started on it. I don't want to be putting demands on the code while it's still in flux
JimmyZ Coke: sorry, I can't follow your 'pretty good counter example' :( 14:47
moritz jnthn__ is on vacation, so now is a perfect occasion wher 6model isn't in flux :-) 14:48
cotto ~~
whiteknight moritz: he's already mentioned to me some changes he plans to make to 6model, and I'm in no particular rush
I need to wrap up my packfiles work first, and I have plenty of other projects to stare at if 6model isn't 100% ready 14:49
...! Just submitted a windows smoke report with 100% passing
JimmyZ did see that from irclog
whiteknight smolder.parrot.org/app/projects/rep...ails/18112 14:50
Coke JimmyZ: you said you didn't know how to describe things in english very well, thereby describing something in english very well.
moritz whiteknight: I wasn't entirely serious :-)
Coke whiteknight: feel free to update my /3 in platforms to a *1 that mentions the potential git config issue. 14:51
whiteknight Coke: Will do
JimmyZ Coke: yeah, Just look up dict, and know what counter example means now
whiteknight Coke: You're on win32-x86? I'm on 64bit here
I'll still update your entry, but I'm adding one of my own 14:52
Coke whiteknight: yes, still 32 bit here. 14:54
whiteknight We don't have a CGoto core, or an Exec core anymore. I suspect we do not need columns for those thigns in PLATFORMS
benabik whiteknight: PF.view() looks good. I'll probably hold off on using it until (a) I actually get bytecode gen working and (b) it's in master. 14:55
Coke whiteknight: true.
whiteknight benabik: I'll merge into master shortly after the release if I can get some test reports. If you have a few spare moments, You could help run some tests on it 14:56
dalek rrot: 54b0299 | Whiteknight++ | PLATFORMS:
Update win32/x86 status for Coke++. Add note about my platform (win7/x64/msvc_16)
 14:57
kudo/nom: 62c49e2 | pmichaud++ | src/core/EnumMap.pm:
Fix hash flattening for empty hashes and enummaps. moritz++
benabik whiteknight: (a) will probably take longer than (b) anyway. :-D But I'll run some test-type things at the branch now.
moritz whiteknight: did anybody test rakudo on top of your branch?
cotto whiteknight, are you seeing the streams.t failure on win7/msvc?
whiteknight moritz: I'm sure not. It's still very early 14:58
cotto: Was failing for Coke. Never failed for me. I'm seeing 100% tests passing on my box
Coke will add a darwin/x86 run later today.
cotto That's the only failure that hasn't been addressed.
ok
curious
whiteknight NotFound and Coke both suggest it has to do with git line-ending configurations
That interpretation makes sense from a quick glance at the tests
cotto that seems likely. The library it tests has hard-coded '\\n'. 14:59
moritz whiteknight: ok, I'll test it when it becomes master :-)
whiteknight moritz: thanks
moritz whiteknight: in general, it's hard for me as a user to figure out which branches are sufficiently mature to need my testing (or rakudo on top of parrot) 15:00
for example some time ago there was a call for testing a branch, and the root email in that thread said it's still early...
whiteknight moritz: yeah, we do need a way to mark branches which are mature enough. I'll send out an email to parrot-dev and parrot-users when I'm ready to start serious testing on my branch
right now I'm only looking for preliminary reports. I haven't even tested myself with g++ yet
moritz and later that branch got merged, broke rakudo, and there was big ruhaha 15:01
whiteknight: if you do, please don't do it as a reply to a previous thread, but as a new thread
whiteknight and I obviously can't merge until wednesday anyway
moritz it's fine, I just wanted to make a general remark
NotFound It's not a suggestion, is a fact. 15:03
On my xp home box, at least, can't talk for other systems.
whiteknight Sorry. "NotFound suggests it is a 'fact' that it's related to git line ending configurations"
NotFound: What is the fix? Is it something Coke can test quickly? 15:08
15:09 jay joined
NotFound whiteknight: no idea... There is some way to tell git to always treat a file as binary? 15:14
whiteknight NotFound: Oh, I don't know 15:16
benabik gitattributes
whiteknight better yet, can we make tests that aren't broken by such obviously-unrelated phenomina? 15:17
we shouldn't have tests that break if the repo is checked out wrong
benabik echo "FILENAME -text" > .gitattributes 15:18
whiteknight We also shouldn't have tests that break on the roll of a dice (t/pmc/threads.t), but we do
benabik whiteknight: Darwin/x86-g++, 4 failures (1 opsc/06-opsfile, 3 dynpmc/subproxy) smolder.parrot.org/app/projects/rep...ails/18115 15:20
whiteknight benabik: wtf 15:21
benabik whiteknight: May be unrelated to your branch, checking a g++ build of where it diverges from master 15:22
whiteknight benabik: oh, that's for the branch? Are you seeing clean build on master?
er, clean tests?
benabik whiteknight: That's your branch, yes. 15:23
whiteknight okay, awesome
benabik I've generally been seeing clean tests on master.
whiteknight not entirely unexpected, Like I said the branch is still young
I wish that the branch name had ended up in the smoke report
benabik That may have been my fault. I think I was on a detached HEAD. 15:24
whiteknight oh, okay 15:26
15:27 alester joined
whiteknight Then I wish there was a way to annotate it. kid51 might pop a gasket if he sees fresh failures 15:27
15:27 jay left
dalek rrot/m0-prototype: f13ab98 | cotto++ | src/m0/perl5/m0_assembler.pl:
start smartening up the assembler wrt constant types

The reason this appeared to work before was because Perl was helpfully converting the string value of an int to its int value when the int was stored as a string in bytecode.		 15:28
15:28 jsut joined
15:39 rurban left 15:42 rurban_ joined
dalek TT #2157 created by pmichaud++: FileHandles aren't flushed and closed on program exit 15:54
TT #2157: trac.parrot.org/parrot/ticket/2157
moritz so is it cosher to set up a destroy() VTABLE that calls Perl 6 code (which in turn might allocated new GCables)? 15:58
pmichaud moritz: I don't think overriding VTABLE destroy is the issue here. 16:00
(if we're talking about flushing filehandles)
16:01 alester left, alester joined
pmichaud the problem is that destroy is never getting called. 16:01
moritz pmichaud: I know, but I'm trying to look at other aspects of the problem
pmichaud okay. 16:02
moritz pmichaud: in the long run we have to provide DESTROY methods
pmichaud since it'll be happening via the 6model destroy vtable, I suspect it'll work out okay.
16:03 Drossel joined
moritz isn't quite so optimistic :-) 16:03
pmichaud afk for a bit 16:04
16:05 fperrad left 16:07 fperrad joined 16:08 Kulag left
whiteknight if FileHandles aren't being flushed and closed on Parrot exit, that's a bug with finalization not working 16:09
Open a ticket
I am not currently optimistic about overriding VTABLE_destroy either, but only because it's never really been done, is untested, unloved, etc
moritz I'll see what I can do :-)
16:15 rurban__ joined 16:17 whiteknight left, whiteknight_ joined, whiteknight_ is now known as whiteknight, rurban_ left, rurban__ is now known as rurban
Coke name your branch "smoke/" something. 16:18
did you need me to check something on win7? 16:19
(smoke) er, to get smoke testing. ISTR p5p is doing something like that, and they have auto-smokers that track that sort of thing. 16:20
16:20 JimmyZ left 16:22 dmalcolm joined
pmichaud 16:09 <whiteknight> if FileHandles aren't being flushed and closed on Parrot exit, that's a bug with finalization not working 16:25
TT #2157
16:32 whiteknight left 16:36 darbelo joined
dalek rrot: 854276e | jkeenan++ | t/README.pod:
Bring into alignment.
 16:37
16:39 kid51 joined
kid51 Can someone please check if I did the right thing in github.com/parrot/parrot/commit/854276edac ? (Am at $job and can't focus on that right now.) Thanks. 16:40
16:40 kid51 left
cotto_work2 ~~ 16:41
Coke odd, my local copy of platforms on darwin doesn't seem to have the PLATFORMS commit that stripped everything out. 16:54
looks like kid51 changed it?
Coke tries to see where this changed on github. 16:55
16:56 lucian joined
Coke I'm super confused. 16:56
yah - if I checkout the commit just before kid51's, it shows the platform the right way. 16:57
if I checkout master at head, I get an old PLATFORM file.
I think his commit is wonky.
anyone? ;) 17:02
benabik takes a poke at it.
Coke looking at the commit graph, he appears to done some wonky merge thing, which reverse merged at least my commit, and probably several others since he branched. 17:04
sorry, s/branch/failed to fast forward/
benabik Yes. `git diff HEAD^2 HEAD` shows a lot of changes. 17:06
Coke so how to undo? 17:07
benabik Could push a rewind and then cherry-pick the original commit he was trying to add. 17:08
Hm. Doesn't cherry-pick cleanly. 17:09
Coke could delete it and let him DTRT next time. :)
so, the first part of what you said. 17:10
?
17:11 mj41 left
dalek rrot: db54a4f | jkeenan++ | t/README.pod:
[codingstd] Shorten line to meet standard.

  (cherry picked from commit f592d3035fc595cf0692fd53a8060a1bab123227)
Conflicts:
  \tt/README.pod		 17:13
benabik That should be better.
Coke: Try pulling again. git may complain a bit about the history going sideways. 17:14
Coke ok. I'll let you respond to my email to parrot-dev, then, either with an "all clear" or notes about how to proceed if someone had pulled that down. 17:15
it said "forced update".
benabik Yes, that's normal.
Coke but nothing else, as I didn't have a commit on top of that. 17:16
benabik But it may try to do a merge.
Coke I always do "git pull --rebase"
benabik Ah.
Coke (just in case I have a commit I forgot about)
benabik Linearizing history is fun. 17:17
(I do the same.)
Coke benabik: can you reply to parrot-dev so folks aren't left hanging?
benabik Yup.
Coke danke. 17:18
benabik++
benabik Am attempting to come up with a couple "what to do"
dalek rrot: 8e90dea | Coke++ | PLATFORMS:
add OS X status.
Coke msg kid51 - please double check that benabik rescued all the changes you intended to commit, he had to change things a bit to clean that up.
aloha OK. I'll deliver the message.
benabik I tried to give a few commands that would help reset to the new history if someone's git tried to merge in the bad commit again. 17:22
Coke benabik++ ! 17:24
benabik git tries to DTRT with merge resolution, but sometimes things get weird.
17:28 whiteknight joined
Coke cheese it, it's the board of directors! 17:30
dalek rrot: 2d4cfad | Whiteknight++ | PLATFORMS:
Remove CGoto and Exec columns from PLATFORMs, since we don't have those anymore. Add update with linux/x64/clang2.8
 17:33
sella: 824cab3 | Whiteknight++ | src/unstable/dumper/ (4 files):
A few rough-drafts for a new object dumper library
 17:36
Coke whiteknight: hey, whiteknight, you deleted my new entry.
github.com/parrot/parrot/commit/2d...d02d#L0R11
17:36 mj41 joined
whiteknight oh, darn. Sorry 17:37
Coke you also added a row under "extra platforms" 17:38
... which is in the commit message, nevermind.
whiteknight my internet connection hiccuped here while I was trying to push
so between when I typed "push" and when it failed with conflict, things happened and I tried again
cotto_work2 Ignore the failing smolder report from win7. I didn't notice that git got confused. 17:48
retrying
17:49 jsut_ joined 17:53 jsut left 18:00 fedov joined 18:18 jsut joined 18:23 jsut_ left
whiteknight heh. Somebody mentioned GCI on the GSOC list. I'm looking forward to that so much this year, I could pass out from excitement 18:26
sorear atrodo: pong
atrodo sorear> I've been told you know who runs dalek 18:27
sorear atrodo: that would be... me
atrodo What, if any, kind of data to you store after the github push notifications? 18:28
whiteknight aloha forget dalek 18:31
aloha whiteknight: I forgot about dalek.
whiteknight aloha dalek is I don't know. Don't ask me, ask sorear. What's with all the questions?
aloha whiteknight: Okay.
whiteknight aloha: dalek?
aloha whiteknight: dalek is I don't know. Don't ask me, ask sorear. What's with all the questions
atrodo whiteknight++ 18:32
whiteknight aloha forget dalek
aloha whiteknight: I forgot about dalek.
Coke dalek is also run by sorear. 18:36
whiteknight dukeleto: ping 18:44
18:45 fedov left 18:48 fedov joined
sorear atrodo: the complete timestamped content of the push packet 18:51
atrodo sorear> Is it possible to get a copy of that data? At least for the past month? 18:53
(wish I had that foresight to save all that)
sorear> I've missed a month of data for isparrotfastyet and I'm trying to get that missing data into the system 18:55
18:55 cotto_work2 is now known as cotto_work 18:57 jay joined
sorear atrodo: do you need the raw data or just the dalek messages? 19:04
19:04 fperrad left
atrodo sorear> The raw data will be easier since that's what isparrotfastyet expects 19:04
19:05 fperrad joined
dalek R: af2ebf8 | (Jay Emerson)++ | src/nqr/Operators.pm:
Fixed bug in * operator
 19:05
sorear atrodo: I only have records going back to Jun 13 19:07
dalek rrot: 6241045 | jkeenan++ | tools/release/update_version.pl:
Add Perl coda.
 19:08
atrodo sorear> That would be good
19:10 fedov left
dalek R: 6b3c490 | (Jay Emerson)++ | t/00-sanity.t:
Marginal testing additions
 19:12
sorear atrodo: host04.appflux.net/drain/packets.tar.bz2 19:19
atrodo sorear++ Thanks! 19:21
atrodo will now implement the exact same system to keep all the data I ever get
19:22 jsut_ joined
dalek R: e091546 | (Jay Emerson)++ | / (3 files):
Negation vectorized
 19:23
19:27 jsut left
dalek sella: 4e67607 | Whiteknight++ | src/query/Provider.winxed:
+comments about qsort
 19:27
19:43 soh_cah_toa joined
jay whiteknight: no question. Making progress, and your matrixy grammar/actions makes far more sense to me now than in my previous looks. !! 20:18
whiteknight jay: Awesome! I'm glad it's serving some purpose
I keep kicking around the idea of trying to revive it somehow, but I haven't quite figured out what I want to do with it yet
Maybe I'll see what you come up with, and draw some inspiration from NQP 20:19
er, NQR
jay I need to work on lvalues. I'm in decent shape with the stuff on the "right side" so to speak.
whiteknight that's good. That's a great first start
does R support multiple return values?
jay No, they would be packed into a list.
I do need to work on some scoping issues with the functions, though. One thing at a time. Unfortunately, life is getting in my way this week. 20:20
whiteknight okay. multi-returns is what really tripped me up with matrixy, because NQP didn't support it at the time 20:21
Still might not, I haven't checked
benabik whiteknight: I asked around a bit when I hit a function in PAST that used it and people seemed to think it didn't. 20:22
whiteknight benabik: NQP-rx uses control exceptions for returns, and exceptions can only carry a single payload object. I think there was some work to try and automatically stuff multiple returns into an array or something, but I imagine that would be a huge pain 20:23
especially true if an array is a valid return value from the function
I think I saw a commit fly by recently that the new NQP was using continuations to implement returns, instead of exceptions. If that's the case, returning multiple values should become possible, or much much easier to implement 20:24
of course, that's in the new NQP 20:25
What I may be very interested in doing is trying to resurrect Matrixy, or a subset of it, on the new NQP as a learning tool 20:26
help to acquaint me better with 6model
Coke ETOOMANYNQPS 20:27
whiteknight for serious
let's all hope that the newest one is the last one
jay Er... I don't think it's likely NQP will teach you much (anything) that you don't know? You're more likely to spot areas for improvement (or for the dustbin).
Coke pmichaud: not that you should stop ;)
whiteknight jay: The newest NQP uses an object metamodel library called "6model", which we are hoping to steal and jam into parrot eventually 20:28
so the more practice I get with that library, the better off I will be when it's time for the big migration
jay Oops... I meant NQR. lol ok, we'll rename sometime soon.
whiteknight jay: "Jay's awesome freaking R compiler" (JAFRC) 20:29
20:38 ilbot2 joined
moderator Parrot 3.5.0 "Menelaus" released | parrot.org | Log: irclog.perlgeek.de/parrot/today | #parrotsketch meeting Tuesday 1930 UTC | Code freeze in effect except for test failure fixes
benabik Netsplit. Wheee... 20:38
20:38 ascent joined 20:39 aloha joined, Maddingue joined
Coke aloha, msg particle blastr.com/2011/07/little-known-sci-fi-fact.php 20:39
aloha Coke: OK. I'll deliver the message.
Coke IRC-- 20:40
benabik karma IRC
aloha IRC has karma of 2.
20:42 PacoLinux joined 20:43 particle joined, autark joined, soh_cah_toa joined, dmalcolm joined, cotto_work joined, tadzik joined, eqhmcow joined, Khisanth joined 20:46 Felipe_ is now known as Felipe 20:56 fperrad left
Coke smolder.parrot.org/app/projects/rep...ails/18143 - 2 test failures on win7 on my box. 21:00
Coke chuckles at the p5p suggestion to rewrite their config system in perl. 21:02
cotto_work Coke: streams.t test? 21:03
Util_ Coke: Those two test failures are due to t/library/perlhist.txt being used as data, but handled like source code.
On Win32, line endings get xlated to \\r\\n, hosing the binary comparisons in streams.t
Coke Util_: yes, we know. ;) 21:04
but I suppose a ticket should be opened to address that.
Util_ Sorry; been away from Net in the mountains.
Coke Util_: note that it depends on how you configured your git.
benabik What files are we having line ending problems in? 21:05
cotto_work t/library/perlhist.txt
benabik Adding a .gitattributes to t/library should fix it. We need it to end in \\n, right? 21:07
Coke is there a git setting that says "no matter what the user's preferences are, treat this file as binary"?
... yay.
benabik We can set it either eol=lf or -text
Coke I can test that. momentito.
benabik echo "perlhist.txt eol=lf" > t/library/.gitattributes 21:09
I'd commit it myself but I can't really test it well.
Coke benabik++ it works. 21:11
dukeleto Util_: i would really like to see some of those pics you took at the parrot/perl6 bof at YAPC::NA 21:12
cotto_work benabik: What do I need to do to make that change take effect?
benabik cotto_work: .gitattributes can be checked in, but you may need to delete .git/index and git reset to make it rescan everything. 21:13
cotto_work no joy 21:15
Coke cotto_work: I'm pushing shortly. momentito.
cotto_work Coke: ok
Util_ dukeleto: I should be able to get them *somewhere* online tonight. I will let you know.
dalek rrot: e1533a7 | Coke++ | t/library/.gitattributes:
force this file to have unix line endings always
 21:16
Coke cotto_work: you might have had to 'git checkout' the file in question again so it got the right line endings. 21:17
(I rm'd it, then git checkout'd it)
all tests pass now on my win7 checkout.
benabik++
dalek rrot: 83cc98a | Coke++ | PLATFORMS:
all tests pass on windows/re-add darwin report.
 21:18
Coke msg whiteknight I readded the darwin report you removed.
aloha OK. I'll deliver the message.
21:19 Util_ is now known as Util
NotFound Will not be easier to delete that file and create a new one with a different name? And preferably without .txt extension. 21:22
cotto_work Coke: awesomesauce. It works now.
NotFound Easier to get smoke reports updated, I mean.
Coke composing a mail to list about win... 21:23
NotFound: easier? no, given that it's already fixed benabik++'s way. 21:25
hurm. running with a high TEST_JOBS on windows doesn't really seem to help.
21:25 bubaflub joined
NotFound Coke: If you neeed to git checkout that file, will that propagate to smolders? 21:26
Coke I don't understand, "to smolders" 21:27
the next time you do a "git pull", it should work.
NotFound The systems that are sending those windows reports with errors.
21:27 preflex left
Coke I would expect it to JFW the next time they 'git pull' and see the new .gitattributes file. 21:28
but IANAGE.
benabik I'm not sure. I think it'll DTRT, but it may need a `git reset --hard`. I knew about attributes, but haven't used them much. 21:30
Coke I included a tickler in the email in case it fails for someone else.
21:31 preflex joined 21:38 lucian left 21:43 Psyche^ joined 21:48 Patterner left, Psyche^ is now known as Patterner
dalek nxed: 91df9be | NotFound++ | winxedst1.winxed:
refactor common parts of using namespace at function and namespace levels into a function
 21:52
cotto_work It's nice to have windows passing. Now for OpenBSD... 21:53
21:59 GeJ_ is now known as GeJ 22:00 wagle_ left 22:01 wagle joined 22:27 kid51 joined 22:33 preflex left 22:34 mj41 left 22:37 preflex joined
kid51 benabik ping 22:47
benabik kid51: pong
kid51 benabik Thanks for helping to straighten out the git commit problem earlier today. 22:48
benabik kid51: No problem. Sometimes it's too easy to get a merge wrong.
kid51 Are you the 'brian@hermes' of smoke reports (e.g., smolder.parrot.org/app/projects/rep...ls/18115)?
benabik And with text, you don't have a compiler yelling you got it wrong.
Yes. That was on whiteknight's branch. 22:49
kid51 I notice that that Smolder report does not include a git commit Version number (on the main Smolder page) or a branch (on the specific report page) 22:50
I believe we recently put some fixes into lib/Parrot/Configure/Smoke.pm in master. Do you think you could get them into that branch?
benabik It may have been confused by the fact I was on a detached head. But I can look at it. 22:51
kid51 In these pre-release hours, it's very helpful to know where a FAIL is coming from.
Hmm, yeah, I suspect a detached head would have neither version nor branch info
Do you think you could submit a straightforward Smolder on master for Darwin/i386? 22:52
benabik Can do. 22:53
22:55 autark left
cotto_work kid51: are you comfortable with the changes I made to the release process? 22:57
kid51: github.com/parrot/parrot/compare/7......e0acdf6 23:00
kid51 You mean: re tools/release/update_version_number.pl ? I have no objection, but haven't tried it out since your commit. 23:02
cotto_work kid51: yup
kid51 I have far greater worries than that ;-)
cotto_work I'll be around if it explodes.
kid51 cotto_work Coke dukeleto anyone: What, if anything, can we do about the OpenBSD build failure reported by Andy D on list today? 23:03
cotto_work seen doughera 23:04
aloha Sorry, I haven't seen doughera.
23:08 bluescreen left
kid51 cotto_work: doughera doesn't do IRC 23:10
cotto_work kid51: funny. I'm sure I've seen him around.
Felipe I haven't an openbsd box in hand to test it. 23:11
cotto_work At this point, we should revert if we know that it'll do some good. 23:15
*iff
dalek TT #2158 created by Felipe++: Appending error messages to IMCC error_message 23:16
TT #2158: trac.parrot.org/parrot/ticket/2158
23:17 alester left
benabik kid51: PASS on Darwin/x86 gcc - smolder.parrot.org/app/projects/rep...ails/18161 23:20
kid51 benabik: Thanks.
23:21 benabik left
dalek rrot: 06131a5 | NotFound++ | t/pmc/socket.t:
provide more info on create socket failing tests
 23:22
23:22 benabik joined 23:23 bluescreen joined 23:24 NotFound_b joined
kid51 NotFound: We are supposed to be in code freeze. 23:27
NotFound_b The test that uses the perlhist file still fails in my XP box, until I delete and checkout the file.
kid51 How does a change in t/pmc/socket.t address that problem?
NotFound_b kid51: this commit was another problem, also related to windows test failures. 23:28
23:28 writtencode1 left
kid51 A change in a test at this point means I have to solicit smoke reports on all systems. 23:28
23:29 elmex left, elmex joined
NotFound_b kid51: if you prefer to have known test failures, I give up. 23:30
23:31 benabik_ joined
kid51 The problem is that while it may fix a failure on one box, it may create failures on others. 23:31
23:31 benabik left, PacoLinux left, tempire left, perlite left, Drossel left, PerlJam left, pmichaud left, jevin left, knewt left, slavorg left, benabik_ is now known as benabik, jevin joined, darbelo left
NotFound_b kid51: it may, but IMO the probability is extremely low. 23:31
kid51 What OSes can you test that out on? 23:32
23:32 spinclad left, sorear left
NotFound_b Windows XP home, ubuntu i386, debian amd64 23:32
23:33 cxreg left, marc left, marc joined
kid51 Can you test that on each of those? Thanks. 23:33
NotFound_b It's already done
kid51 Could you submit a Smolder from the Windows XP box?
23:34 writtencode joined
cotto_work passes on win7 23:34
(msvc)
kid51 cotto_work Can we get a smolder on *that* as well?
NotFound_b kid51: I'll try, but I'm not sure about its cpanness
23:34 dmalcolm left 23:36 PacoLinux joined, tempire joined, perlite joined, Drossel joined, PerlJam joined, pmichaud joined, knewt joined 23:38 spinclad joined, autark joined 23:39 preflex left 23:40 slavorgn left, slavorgn joined 23:41 sorear joined 23:42 dalek left, cxreg joined
cotto_work kid51: running now 23:43
NotFound_b For the stream problem, I think we must wait after the release to decide if will need another solution.
cotto_work NotFound_b: yes
NotFound_b In the meantime, expect failure reports from system when no one does the update manually.
23:43 dalek joined 23:44 preflex joined
kid51 NotFound_b: which update are you referring to? 23:45
benabik kid51: I think he means the gitattribute update. Fresh checkout should work, but it doesn't change already checked out files. 23:46
23:46 slavorg joined
NotFound_b kid51: ensuring that the perlhist file used for the test is updated to take into account its attribute changes 23:46
kid51 Okay, I understand.
NotFound_b smokes sent from my three boxes 23:47
kid51 NotFound_b++
bbial
NotFound_b kid51: thanks, and sorry for the late change, but my XP box is painfuly slow. 23:48
benabik kid51: I resmoked Darwin/x86 as well. 'sall good
NotFound_b Note that some socket tests still fails on it, but at least the test file runs at full and report well, 23:50

Powered by Cro and the Raku® Programming Language.

Please report any issues / comments / feature requests as an issue on App::Raku::Log.

Thank you!