Camelia, the Perl 6 bug

IRC log for #parrot, 2012-04-17

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 whiteknight good evening, #parrot
00:04 kid51 release tomorrow?
00:05 kid51 make test passes the usual way on Darwin/PPC and the usual 4 ways on linux/i386
00:05 kid51 at HEAD
00:05 kid51 running smolder_fulltest now
00:16 whiteknight I'm testing everything I've got now
00:25 plobsing joined #parrot
00:28 kid51 http://smolder.parrot.org/app/​projects/report_details/27604
00:28 kid51 pass
00:30 whiteknight okay, I'm building on windows, I'm running my smoker
00:40 whiteknight build on windows worked. Testing now
00:50 whiteknight all tests pass linux x86-64 with clang, normal and --optimize
00:52 whiteknight oh, looks like some failures on windows with strawberry
00:55 whiteknight t/pmc/sockets.t, t/pmc/socket_ipv6.t, t/pmc/os.t
00:55 whiteknight looks like mostly the ipv6-related tests. Probably need to be TODOd on that platform
00:57 whiteknight And the os.t failure looks like something wrong with the platform encoding. That test is probably wrong too, if it assumes a specific platform encoding
01:05 whiteknight msg cotto We've got 3 failures on win32/strawberry. socket.t and socket_ipv6.t fail, I don't think we have ipv6 sockets there. t/dynpmc/os.t fails, looks like a bad test. Not blockers
01:05 aloha OK. I'll deliver the message.
01:28 dalek rakudo/nom: 424936c | coke++ | docs/announce/2012.04:
01:28 dalek rakudo/nom: Add preliminary release announcement
01:28 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/424936c740
01:28 whiteknight hmmm.... config fails with icc
01:29 whiteknight I haven't tried in a few months, but it used to succeed
01:30 cotto hio
01:31 whiteknight icc can't seem to find a "bits/predefs.h" include file, which is required for several steps
01:31 cotto whiteknight, thanks
01:34 whiteknight oh, i obviously needed to install libc6-dev-i386 in order to keep building with a compiler I had been using for months
01:34 whiteknight sometimes, Ubuntu bugs the shit out of me
01:40 kid51 whiteknight: Is it okay if I direct people at 10gen to your blog post about working with mongoDB?
01:40 whiteknight kid51: what do you mean?
01:40 * kid51 will be over there for an ny.pm meetup wednesday night
01:41 kid51 mongoDB is a product of 10gen, correct?
01:41 whiteknight kid51: yeah
01:41 whiteknight I don't really have much to say about mongodb, yet
01:41 kid51 So, they may be interested in what you are doing with mongo
01:41 whiteknight okay, that's cool too
01:43 kid51 http://tinyurl.com/7sn8v6j
02:35 brambles joined #parrot
02:38 alvis` joined #parrot
05:22 Coke joined #parrot
05:59 fperrad joined #parrot
06:22 nine msg whiteknight Hosting space is not a problem. I got some 600 GB of unused space on my webserver and virtually unlimited traffic on a 100 MBit/s connection. The problem is that it's probably not very legal to offer an XP VM for download.
06:22 aloha OK. I'll deliver the message.
06:23 plobsing joined #parrot
07:08 brrt joined #parrot
07:20 votz joined #parrot
08:24 lucian joined #parrot
08:38 mdupont joined #parrot
08:41 bacek joined #parrot
09:23 plobsing joined #parrot
09:30 alvis joined #parrot
09:55 mdupont joined #parrot
10:15 bacek joined #parrot
11:50 kid51 joined #parrot
11:55 benabik joined #parrot
12:04 whiteknight joined #parrot
12:05 whiteknight good morning, #parrot
12:05 benabik o/ whiteknight, #parrot
12:05 whiteknight hello benabik
12:52 PacoAir joined #parrot
13:04 bluescreen joined #parrot
13:07 dalek rakudo/nom: b403135 | coke++ | docs/release_guide.pod:
13:07 dalek rakudo/nom: allocate a release name.
13:07 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b4031355bf
13:11 dalek rakudo/nom: 95a6d8c | moritz++ | docs/ (2 files):
13:11 dalek rakudo/nom: update changelog and release announcement a bit
13:11 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/95a6d8c310
13:20 dalek rakudo/nom: b401415 | moritz++ | docs/announce/2012.04:
13:20 dalek rakudo/nom: [announce] note two incompatible changes in the release announcement
13:20 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/b401415d11
13:51 rich joined #parrot
13:52 dalek rakudo/nom: 36bbb40 | (Felix Herrmann)++ | src/core/ (2 files):
13:52 dalek rakudo/nom: add a few typed IO exceptions
13:52 dalek rakudo/nom:
13:52 dalek rakudo/nom: Signed-off-by: Moritz Lenz <moritz@faui2k3.org>
13:52 dalek rakudo/nom: review: https://github.com/rakudo/rakudo/commit/36bbb4035c
14:39 alester joined #parrot
14:45 NotFound ~~
14:51 whiteknight hello NotFound
14:53 jashwanth joined #parrot
14:55 jashwanth hello whiteknight
14:56 whiteknight hello jashwanth
15:22 Justin joined #parrot
15:22 Justin hello
15:24 lucian joined #parrot
15:32 whiteknight hello Justin
15:46 hercynium joined #parrot
15:49 hercynium_ joined #parrot
16:00 cxreg joined #parrot
16:52 PacoAir joined #parrot
16:57 cogno joined #parrot
16:58 cogno joined #parrot
17:08 dmalcolm joined #parrot
17:08 nbrown joined #parrot
17:13 Justin joined #parrot
17:31 dukeleto ~~
17:35 cotto ~~
17:41 whiteknight good afternoon, dukeleto and cotto
17:41 whiteknight cotto: how we looking for the release?
17:42 cotto I was too wiped out last night to do anything.  I'm planning on starting it after #ps today.
17:44 whiteknight what time is #ps?
17:44 whiteknight is it in 45 minutes or 1hr+45min?
17:44 cotto in 1:45
17:49 Coke I'm doing the rakudo release later this week. Any big changes I should know about?
17:49 cotto good #ps question
17:50 whiteknight Coke: No, I don't think anything big
17:50 cogno joined #parrot
17:51 whiteknight are they using a recent parrot revision with rakudo for testing?
17:51 benabik I tend to build blead on both, and it WFM
17:52 whiteknight okay, good
17:52 whiteknight I actually haven't built Rakudo in a little while, but I don't think we've seen any breaking changes recently
18:03 rich left #parrot
18:13 cogno joined #parrot
18:22 benabik whiteknight: Is ParrotStore going to have a common interface to multiple DBs or just provide a minimal Parrot "native" interface to each?
18:23 cotto I'm seeing a checkdepend.t failure and a dynpmc/select.t failure.  istr that the select.t failure is noise, but the checkdepend failure might be legitimate
18:44 Coke doing a build of threads branch on osx 10.7.3
18:46 lucian joined #parrot
18:46 Coke https://gist.github.com/2408144 - test failures on threads.
18:46 Coke not sure if they are also failing on master. checking.
18:51 whiteknight benabik: I've been kicking that idea around a lot. I think I will provide some kind of high-level interface for storing STRING and frozen PMCs
18:51 whiteknight benabik: but since Memcached is just key/value and mongodb is document based with nested objects, I don't know what kind of common interface I can provide that will make sense for too many things
18:52 Coke same failures on master.
18:52 Coke so, cotto: TESTS ARE FAILING ON master.
18:52 whiteknight hmm.. I didn't see those failures last night when I tested
18:52 benabik I did not see those failures this morning.
18:53 cotto Coke: you could say that
18:54 Coke gcc 4.2.1 on osx 10.7.3.
18:54 whiteknight Coke: PCC or x86?
18:54 whiteknight er, PPC
18:54 Coke I dont think apple even makes ppc anymore. x86.
18:54 benabik 10.7 doesn't run on PPC
18:54 whiteknight they don't make them anymore, but some people have not upgraded from them
18:56 Coke also, 64bit.
18:58 Coke how to run a single test again?
18:58 dukeleto Coke: prove -v t/foo.t
18:59 Coke dukeleto: if I run it thatway, I get 31 failures in t/pmc/bignum.t instead of 1
18:59 benabik Just re-ran `make test` and did not see those errors.  Also on 10.7.3 64b.    :-/
18:59 Coke # dyld: Library not loaded: /usr/local/lib/libparrot.dylib
18:59 Coke ... I am not installing parrot here.
18:59 benabik OS X is... difficult wrt shared libraries.
19:00 Coke I get these errors with another random .t file with prove.
19:00 Coke benabik: "this all used to work.". ;)
19:01 benabik I think you need DYLD_LIBRARY_PATH=$PARROT/b​lib/lib:$DYLD_LIBRARY_PATH in the environment.
19:02 benabik Testing parrot on OS X is full of small gotchas.  Have to not have Parrot installed, have to set DYLD_LIBARAY_PATH correctly.  Don't remember if there's another one.
19:02 Coke benabik: that helps.
19:02 benabik The Makefile sets the environment and Configure warns if you have parrot installed.
19:02 Coke benabik: I never had to do that on 10.6
19:03 benabik Coke: In either 10.6 or 10.7, dyld got much pickier about paths.
19:03 Coke here's the error on that failing test:
19:03 Coke # error:imcc:add_1_const:Integer overflow '9223372036854775808'
19:03 benabik I thought it was .6, but I guess it was .7
19:04 Coke benabik: i say .6, but I upgraded that box from 10.5 at some point.
19:05 Coke ah, the test file warns: # GMP version 5.0.4 is buggy with huge digit multiply - please upgrade
19:06 benabik I get that warning too, but the overflow check works.
19:06 Coke there is no version higher than that in macports, though.
19:06 whiteknight Coke: is there a version lower than that?
19:07 benabik Is IMCC trying to optimize it?  IMCC should never see 9223372036854775808 as a constant.
19:09 contingencyplan joined #parrot
19:09 Coke https://gist.github.com/2408144 updated with verbose outputs from those 2 failures
19:10 benabik That add_1_const error is from IMCC_int_from_reg (compilers/imcc/pbc.c:1885)
19:10 Coke rant: test output should not include the strings "ok/not ok".
19:10 cotto would it be appropriate to add an exception for platform.c to checkdepend.t, since it's generated?
19:11 whiteknight cotto: I think so, but I hate that test and would love to add more exceptions to it
19:11 cotto why do you hate it?
19:11 Coke what error are you getting?
19:11 cotto it's complaining that there's no rule for src/platform.c in Makefile
19:12 Coke whiteknight: that test is the only thing keeping "make -j" working.
19:12 cotto yeah
19:12 Coke cotto: error or warning?
19:12 benabik If src/platform.c is generated, shouldn't there be a rule in the Makefile?
19:13 benabik_ joined #parrot
19:13 Coke aye, prolly.
19:13 cotto it doesn't cause either an error or a warning during the build
19:14 Coke huh. I don't have a src/platform.c
19:14 Coke cotto: I mean during the test.
19:14 Coke are you failing a test? or just getting some diag output?
19:15 Coke also, your makefile should probably have a rule that says:
19:15 cotto Coke: actual test failure
19:15 Coke src/platform$(O) : src/platform.c
19:20 cotto Coke: I don't see that line
19:20 whiteknight cotto: in Makefile or in config/gen/makefiles/root.in?
19:21 cotto in either
19:25 cotto now it's passing
19:25 cotto after I didn't do anything
19:28 cotto sigh
19:28 cotto looks like that failure went away
19:28 cotto as does t/dynpmc/select.t
19:30 cotto #ps now
19:40 benabik joined #parrot
19:43 mdupont joined #parrot
19:53 NotFound Am eof test failure? Maybe some line ending conversion?
19:53 NotFound Ops
21:02 cotto alester: what's the version of gcc that you've been playing recently?
21:03 cotto 4.8?
21:23 losinggeneration joined #parrot
21:51 benabik cotto: In #perl6, alester said he was using 4.8 snapshots.
21:51 alester You want I should look?
21:52 alester It's a weekly snapshit.
21:52 alester hahaha
21:52 alester snapshot.
21:52 alester gcc-4.8-20120401.tar.bz2
21:52 cotto benabik: ok.  that's what I thought
21:52 alester so I'm two weeks back.  Why do you ask?
21:52 cotto want to know which gcc version to credit
21:54 alester for what?
21:55 cotto what you've been using for your recent batches of consting changes
21:56 tadzik lol, snapshit :)
21:58 alester cotto: Is this going in some release notes or something?
21:59 cotto alester: yes
22:41 cotto bacek_at_work: ping
22:42 cotto bacek_at_work: you requested that kill_current_object be merged.  How important is it to get that in before the release?
22:45 cotto It looks like it almost entirely deletes things
22:45 cotto moritz: ping
22:55 alester One of the bummers of using the snapshots is that they don't always build.  The 20120415 GCC blows up with a syntax error in a .h file.  Oh well, gotta wait 'til next week.
23:00 cotto afk for a bit (meeting).  moritz, same question for you as for bacek_at_work
23:35 Coke_ joined #parrot
23:45 benabik joined #parrot
23:46 whiteknight joined #parrot
23:47 whiteknight good evening, #parrot
23:50 Coke joined #parrot
23:51 masak joined #parrot
23:53 Util joined #parrot
23:53 PerlJam joined #parrot
23:54 pmichaud joined #parrot

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

Parrot | source cross referenced