Parrot 3.1.0 Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Goals: merge and extensively test gen_gc2 | Green light from lua, partcl-nqp and rakudo on gen_gc2
Set by moderator on 18 February 2011.
00:13 nwellnhof left
cotto ~~ 00:15
dukeleto cotto: good day, sir 00:21
dukeleto just switched from irssi to weechat and is very pleased 00:22
00:27 bubaflub joined
dukeleto bubaflub: hey stranger 00:32
bubaflub dukeleto: hey!
dukeleto: long time, no chat
dukeleto bubaflub: how is the Wolfram gig?
bubaflub dukeleto: well, i like my coworkers 00:33
dukeleto bubaflub: been hacking on anything lately?
bubaflub dukeleto: cleaned up Math::Primality 00:34
working on the Math::Primality::AKS
i tried to get LibreOffice on Mac OS X to work, but it wouldn't compile 00:35
i've been working on some OCR stuff, though
00:36 kid51 joined
dalek sella: 1f28bf8 | Whiteknight++ | README.md:
+readme
 00:38
whiteknight good evening, #parrot 00:39
bubaflub evening, whiteknight
whiteknight hello bubaflub. How are you doing?
bubaflub whiteknight: good! dropped off for a bit - finished a semester and got married and switched jobs 00:40
whiteknight: but i'm looking to get back into the swing of things
whiteknight congratulations on all counts
bubaflub whiteknight: a-thank-you
kid51 so I guess finishing a semester, getting married and switching jobs does NOT qualify as the swing of things ;-) 00:41
bubaflub kid51: haha. i meant the parrot things. i've been away since the svn -> git switch
kid51 #parrot - much more swinging thatn all that!
bubaflub though i read all the emails
riveting reading material 00:42
kid51 Would you like to join a team working on a roadmap goal?
bubaflub what are the goals? i'm a bit rusty will do whatever i can 00:44
dukeleto bubaflub: what are you interested in?
bubaflub: it is best to harness your interest than throw you at something, but we can surely throw you at something :) 00:45
kid51 bubaflub: here's the short version of the roadmap goals/teams as discussed at Jan 29 PDS: thread.gmane.org/gmane.comp.compile...devel/5195 00:46
dukeleto bubaflub: we really need good benchmarks
bubaflub: maybe you want to hack on some prime-checking in PIR to be used as a benchmark? 00:47
bubaflub kid51 and dukeleto: i'd love to hack on ruby or javascript HLLs too
dukeleto bubaflub: we need stuff that stresses the GC
bubaflub kid51 and dukeleto: as long as i got someone to do some hand holding
dukeleto: yeah, i could do that as well
kid51 bubaflub: Also: See pmichaud's summary of rakudo's needs from Parrot: thread.gmane.org/gmane.comp.compile...evel/5200; these will probably become roadmap goals
dukeleto bubaflub: sweet, clone jaspers and tell me why it doesn't work ;)
bubaflub dukeleto: hahaha, alright
dukeleto bubaflub: i get a weird error from node.js 00:48
bubaflub kid51: reading...
dukeleto: cloning...
kid51: lamentably, i'm not really sure i'm particularly great at any of those tasks 00:59
dukeleto bubaflub: which is why we will find something that you want to hack on
bubaflub: cardinal hacking might be fun for you
bubaflub dukeleto: yeah, that sounds good. a while ago it was a bit stalled out on the object model and i couldn't get it installed 01:00
dukeleto: but hey, if it actually works, maybe i'll see if we can hook up the RubySpec stuff to it
dukeleto bubaflub: 6model is the new perl 6 object model, that will trickle down to parrot 01:01
bubaflub dukeleto: yeah, i was a bit of the brouhaha over that 01:02
dukeleto bubaflub: so we are fixing the foundation to build the floor that Cardinal will use :)
Tene There's still plenty of work to be done on cardinal even before that.
The parser needs to be updated away from PGE
The parser isn't entirely complete, iirc. There are some lingering nasty issues that should be easier to solve in nqp regexes, iirc 01:03
But, yes, the majority of the ruby test suites assume that they're running against an implementation that's mostly complete 01:04
so they assume things like monkey-patching other classes, etc.
bubaflub Tene: ok, so where should i start?
Tene bubaflub: I have no idea what sort of tasks you like working on.
bubaflub Tene: i could probably help fill out the standard library without much hand holding 01:05
Tene You could start by running the test suite, and making sure that all of the failures have open tickets on the issue tracker, and see if there's anything there that could use work.
bubaflub: That would be great.
bubaflub Tene: ok, i'll get started with that
Tene bubaflub: cardinal's policy is that the test suite should always pass; any failing tests should be marked as skip or todo 01:06
with notations in the output for the issue number
bubaflub Tene: ok, i've got HEAD on both parrot and cardinal and will test it
Tene at least most of them are like that; any failures that aren't like that are new since cardinal was worked on before.
Expanding the test suite would be great, expanding the stdlib would be great. 01:07
bubaflub ok
do we have any test coverage tools?
Tene If you want to dig into the compiler, the best option right now is migrating the parser over to nqp-rx from PGE.
bubaflub Tene: i haven't done much with parsers (beyond the basics) and nothing with nqp-rx, but that sounds interesting
Tene bubaflub: The syntax is mostly the same; it would mostly be a case of tearing out the PGE infrastructure, trying to run it against nqp-rx, and then iteratively fixing the failures as they come up. 01:08
It shouldn't take more than a couple of days of beating on to get it running, in the worst case, I'd guess.
bubaflub ok 01:09
Tene bubaflub: in terms of infrastructure, the wiki and issue tracker need to be migrated to the cardinal repo in the parrot org on github, an HLL devs team needs to be created on github, with privs to HLLs like cardinal but not parrot core repos, so others can commit to cardinal, and then we could take down the old cardinal repo on github 01:10
If you want access to the cardinal account on github, ask me or Hackbinary for the credentials
bubaflub Tene: ok. 01:11
Tene You could also document that tasklist in a TODO or whatever in the cardinal repo :)
bubaflub Tene: yeah, sounds good
Tene oh, I think dukeleto has credentials for the cardinal account on github too
Tene gets back to $realjob work now. 01:12
bubaflub Tene: silly question - i suppose i need an installed parrot for cardinal to build
Tene Good luck! Great to have you interested. Please let me know if there's any questions you have
also #cardinal
umm... probably
Yeah, you'll need an installed repo somewhere
I don't recall whether it's been fixed well enough to not rely on a parrot build dir still being available or not; investigating any install issues, getting infrastructure for building distro packages, testing on other platforms like MacOS or Windows might be good... 01:14
bubaflub Tene: i'm on Mac OS X 10.6.6 and got it to build 01:15
Tene That's good to know. I'm pretty sure treed was on OS X full-time when he was working on it a lot, so I'm not surprised. 01:16
I haven't heard of anyone trying it on Windows, but I'd expect it to work.
bubaflub Tene: so i'm running rake test:all, i'll post an update on #cardinal when it's done 01:17
Tene bubaflub: you can also run rake smoke, to submit a smolder report
should be able to, at least. There's been a few problems with that, but I'm pretty sure it's been fixed 01:18
bubaflub Tene: okey dokey
01:42 vmspb left
dalek rdinal: 57bee5f | bubaflub++ | .gitignore:
ignore smolder report/ directory
 01:49
rdinal: 9ef4290 | bubaflub++ | / (2 files):
basic documentation
02:11 whiteknight left
dalek rdinal: e90ade7 | bubaflub++ | Rakefile:
fix some typos
 02:26
rdinal: 2a5c349 | bubaflub++ | Rakefile:
whoops, actually fix these things lateau++
 02:30
02:41 woosley joined 02:49 toei left
dalek rrot/tt1954_eliminate_make_docs: 81165eb | jkeenan++ | / (2 files):
Create html/ directory.
 02:50
02:50 ShaneC joined 02:54 ShaneC1 left 03:00 ShaneC left 03:01 toei joined, contingencyplan left 03:10 bubaflub left
kid51 If I no longer need a repository on github, how can I delete it? 03:27
Example: At one point I forked chromatic's modern_perl_book. The book is published; I no longer need my fork. 03:28
cotto admin -> delete this repository
kid51 Hmm, I don't see 'admin' 03:29
oh, wait
Thanks. 03:30
cotto np 03:31
kid51 cotto: How is work on your roadmap goal progressing?
cotto slowly but surely 03:32
The key seems to be to not care how many other people are working on it and plow ahead.
People will join as they have tuits.
(I mean that I'm encouraging others to join in, but I don't let myself get too attached to whether they contribute to any given aspect of the design or not.) 03:33
kid51 Is there anything to look at yet besides the draft PDD? 03:34
cotto btw, I think it's highly premature to clean up the pod in pdd 32. The document isn't anywhere near final.
no 03:35
kid51 k. That was a byproduct of running tests.
cotto it'll just be the pdd for at least a week or two 03:38
03:41 bubaflub joined
dukeleto kid51: the lorito branch 04:00
kid51: the lorito branch is my crack at implementing the draft lorito pdd with dynops
kid51: nothing close to the final product, but a prototype to flesh the spec out 04:01
github.com/parrot/parrot/compare/m...r...lorito
mostly just a stub, but it is beginning to get a tiny bit of flesh 04:02
cotto good evening, dukeleto
dukeleto cotto: hola 04:03
04:21 woosley left, kid51 left
dalek rrot: 967267f | petdance++ | src/hash.c:
remove unused interp argument
 04:27
rrot: 47cf36b | petdance++ | src/hash.c:
removed unused variable
rrot: c05d2c0 | petdance++ | src/hash.c:
remove extraneous if() wrapping around the for()
rrot: 46d2486 | petdance++ | src/key.c:
removed unused var N
rrot: 178d329 | petdance++ | src/pmc/callcontext.pmc:
remove unused variable
rrot: a5a74ba | petdance++ | / (2 files):
add PARROT_xxx decorators
04:35 plobsing left 04:45 woosley joined 04:52 bubaflub left
dukeleto PARROT_xxx decorators sounds like it could be an interesting movie title... 04:53
cotto Thanks. I'm going to have some really strange dreams now. 04:58
05:00 benabik left
cotto dukeleto, ping 05:36
dukeleto cotto: ping! 05:45
cotto dukeleto, I'm thinking about alignedness and the m0 bytecode format. Do you know what the consequences of having e.g. 5-byte long ops would be and if it'd be better to use padding?
also, pong
Coke seen kid51 05:50
clunker9 kid51 was last seen on #parrot 2 hours, 15 minutes and 25 seconds ago, saying: k. That was a byproduct of running tests.
aloha kid51 was last seen in #parrot 2 hours 15 mins ago saying "k. That was a byproduct of running tests.".
Coke why do we have clunker in here?
clunker, owner? 05:51
aloha, owner?
aloha Coke: No clue. Sorry.
Coke aloha, owner is bacek, but see also: "source" 05:52
aloha Coke: Okay.
sorear imagine filling memory with 5-byte records which are accessed at random 05:53
imagine that there are so many of them that the L1 cache is irrelevant 05:54
a given record will cross a 32-byte boundary if it starts at an address congruent to 31, 30, 29, or 28
therefore, with 32 byte cache lines, the expected number of memory accesses per record access is 1 1/8 05:55
cotto seen clunker9
aloha clunker9 was last seen in #parrot 4 mins 52 seconds ago saying "kid51 was last seen on #parrot 2 hours, 15 minutes and 25 seconds ago, saying: k. That was a byproduct of running tests.".
cotto much better
sorear padding the records (optimally, by 1 byte per 3 records) would give an improvement of 11% in this model
however the model leaves a lot of stuff out... m0 instructions are (hopefully!) accessed sequentially a lot, and increasing the number of sequential records that fit in a cache line is good 05:56
also there are many sizes of cache line, and what improves performance at one level of the hierachy may hurt it at others 05:57
my machine here seems to have 4096 byte TLB lines and 64 byte L1/L2 lines 05:58
[the page size can be ignored since it will always be the same as the TLB size]
this assumes your record is 5 1-byte fields 05:59
nopaste "coke" at 192.168.1.3 pasted "current partcl (old) failures on parrot master." (31 lines) at nopaste.snit.ch/33153
sorear an intelligent assember could probably align labels such as to avoid the vast majority of line-spanning records 06:02
simply starting subs at addresses = 4 mod 64 would prevent spanning for subs of 24 or fewer instructions 06:03
Coke looks like most of those failures are due to HLL issues. I get a String instead of a TclString when invoking lrepeat, e.g. 06:05
06:07 lateau joined 06:09 benabik joined 06:12 rurban_ joined 06:15 rurban left, rurban_ is now known as rurban 06:17 woosley left
Coke is there any way to suss at runtime what the .HLL was set to? 06:18
Coke has bugs open 9 months here. What's an HLL guy to do ? 06:22
Coke throws a new one on the pile and heads off to sleep. 06:32
dalek TT #2022 created by coke++: Something borked with HLL mappings. 06:37
TT #2022: trac.parrot.org/parrot/ticket/2022
06:38 jsut_ joined 06:40 fperrad joined 06:42 bacek left, jsut left 06:43 jsut joined 06:46 jsut_ left 06:49 theory left 06:52 bacek joined
cotto g'day, bacek 06:54
07:21 Hackbinary left 07:23 dwebb joined 07:24 dwebb left
cotto aloha, clock? 07:46
aloha cotto: LAX: Fri, 23:46 PST / CHI: Sat, 01:46 CST / NYC: Sat, 02:46 EST / UTC: Sat, 07:46 UTC / LON: Sat, 07:46 GMT / BER: Sat, 08:46 CET / TOK: Sat, 16:46 JST / SYD: Sat, 18:46 EST
08:16 fperrad_ joined 08:20 fperrad left, fperrad_ is now known as fperrad 08:54 Hackbinary joined 09:12 mtk left 09:19 mtk joined 09:43 contingencyplan joined 10:23 lucian_ joined 10:25 Hackbinary left 10:27 lucian left
bacek ~~ 10:33
cotto ~ 10:34
Tene 10:35
bacek -~ 10:36
10:36 Hackbinary joined 10:43 lucian joined 10:46 lucian__ joined 10:47 lucian_ left 10:49 lucian left 10:51 lucian joined 10:54 lucian__ left 11:05 jsut_ joined 11:10 jsut left 11:29 mikehh left 11:45 whiteknight joined, preflex left 11:46 nopaste left 11:47 nopaste joined 11:53 preflex joined 12:00 lucian left 12:02 lucian joined 12:14 mikehh joined
whiteknight good morning #parrot 12:21
12:46 contingencyplan left 12:52 jsut joined 12:57 jsut_ left 12:59 lucian_ joined 13:00 preflex left 13:01 lucian__ joined 13:02 lucian left 13:05 fperrad_ joined, lucian_ left, preflex joined 13:09 fperrad left, fperrad_ is now known as fperrad 13:19 plobsing joined 13:20 kid51 joined
moritz good morning whiteknight 13:25
whiteknight hello moritz. How are you today? 13:26
moritz whiteknight: fine, thank you. What about yourself? 13:27
whiteknight also doing well. 13:30
13:37 fperrad left 13:50 kid51 left 14:03 woosley joined, woosley left 14:12 rurban_ joined 14:15 rurban left, rurban_ is now known as rurban 14:25 plobsing left 14:29 plobsing joined 14:37 plobsing left 14:55 plobsing joined 14:57 zby_home joined 15:01 plobsing left
Hackbinary hello #parrot 15:10
15:17 bubaflub joined
whiteknight hello Hackbinary 15:18
opbots trust Hackbinary
slavorg Ok
slavorgn Ok
bubaflub mornin' #parrot 15:20
whiteknight good morning, bubaflub 15:22
15:22 preflex left
bubaflub morning whiteknight 15:22
Hackbinary cool, am I a channel operator now?
whiteknight Hackbinary: it's one of the perks of being a contributor
15:40 preflex joined
dalek sella: f6be7b0 | Whiteknight++ | / (6 files):
quick cleanups of the testsuite code
 15:52
sella: d31196b | Whiteknight++ | xunit/ (2 files):
Rearrange the test classes more
sella: 5fbc111 | Whiteknight++ | / (16 files):
start reorganizing the test code. Update tests to match the new style
sella: 340a58b | Whiteknight++ | t (5 files):
property handle TODO tests in the harness.

Refactor the Harness::Loader so we take a filetype proto in the constructor. This removes the need for subclasses here. Multiple other fixes
sella: 0795c0e | Whiteknight++ | / (5 files):
implement my own replacement for Test::Builder, get rid of evil hacks
16:06 kid51 joined
tadzik hello 16:08
whiteknight hello tadzik
tadzik whiteknight: kudos for Rosella, looks freaking awesome 16:09
whiteknight tadzik: Thanks! It's a culmination of many things I've been planning for months
tadzik I wonder how real is language interoperability, so I can use it in Perl 6 without wrappers
whiteknight tadzik: that depends on a lot of factors. At the Parrot level a PMC is a PMC. individual languages and compilers may impose conflicting semantics 16:10
the dream of never needing any wrappers is probably unrealistic, especially if different HLLs are designed without interop in mind 16:11
tadzik hrm 16:15
whiteknight for instance, it would be very easy for the JavaScript developers to not implement VTABLE_find_method at all, which I'm sure P6 is going to rely on to find methods 16:16
it's a trivial addition, but it's ancillary to core JS behavior
likewise, it's very possible that P6 would not be expecting to find methods using getattribute_p_p_s 16:17
16:17 Psyche^ joined, Patterner left, Psyche^ is now known as Patterner 16:19 cognominal left, cognominal joined 16:42 fperrad joined 16:46 lateau left
dalek rrot: 6ac86c9 | jkeenan++ | config/ (3 files):
Apply patch submitted by Coke in TT #1954.
 16:54
rrot: 3c8fed0 | jkeenan++ | config/ (3 files):
Merge branch 'coke_tt1954'
rrot: c743d11 | jkeenan++ | config/gen/makefiles/root.in:
Need to remove 'docs-clean' from 'clean' target.
rdinal: 5faa0f6 | bubaflub++ | src/builtins/cmp.pir:
fix != and !~
 16:58
16:59 eternaleye_ joined
dalek TT #1954 closed by jkeenan++: 'make docs' target: Why are these files not installed? 17:00
TT #1954: trac.parrot.org/parrot/ticket/1954
17:00 mtk left 17:01 eternaleye left 17:06 mtk joined 17:09 fperrad_ joined
dalek rrot: b812b49 | jkeenan++ | NEWS:
Announce removal of 'make docs' target.
 17:12
17:12 fperrad left, fperrad_ is now known as fperrad 17:34 eternaleye_ is now known as eternaleye 17:37 theory joined, theory left 17:54 kid51 left 17:57 jsut_ joined 18:00 bubaflub left 18:02 jsut left, jsut joined 18:07 jsut_ left 18:28 theory joined
dalek rrot: dee6786 | (Gerd Pokorra)++ | MANIFEST.generated:
let "languages/JSON/JSON.pir" be installed
 19:01
rrot: 5b84f21 | (Gerd Pokorra)++ | compilers/data_json/JSON.nqp:
remove unnecessary loading of bytecode
19:03 contingencyplan joined, jsut_ joined 19:04 lucian__ is now known as lucian 19:08 jsut left 19:09 bubaflub joined 19:17 lucian_ joined 19:20 bacek left, lucian left
cotto ~~ 19:22
whiteknight hello cotto 19:23
19:24 kid51 joined
cotto good day, whiteknight 19:24
NotFound $ make -sj 19:27
compilers/imcc/imcc.y -d -o compilers/imcc/imcparser.c
-ocompilers/imcc/imclexer.c compilers/imcc/imcc.l
/bin/sh: ./parrot-nqp: No existe el fichero o el directorio
make: *** [runtime/parrot/languages/JSON/JSON.pir] Error 127
kid51 realizes he should have organized a hackathon to fill up this 3-day weekend
cotto NotFound, that should be an easy fix. 19:28
kid51 is forced to do cage cleaning instead (after all, it's preferable to house cleaning)
NotFound cotto: nothing nqp related is easy to me.
cotto I'm beginning to see why you write winxed. 19:29
cotto just got the same error with a parallel build 19:30
NotFound Yes, without -j it build fine for me. 19:31
cotto which is a little surprising, because I don't have a Spanish localized system
NotFound Uh, forgot the LANG=C that I usually do before reporting. 19:32
19:32 bacek joined
cotto ;) 19:33
dalek rrot: 9859ea7 | cotto++ | compilers/data_json/Rules.mak:
add nqp-rx as a depenency for JSON.nqp
 19:35
NotFound cotto++ Builds fine now 19:37
Coke pokes parrot with a stick? 19:39
cotto NotFound, great. Thanks for reporting. 19:47
19:49 theory left
cotto seen gerd 19:51
aloha gerd was last seen in #parrot 4 days ago saying "it is: ...@perl.org".
19:51 nwellnhof joined
nwellnhof ~ 19:51
cotto msg gerd In dee67865dee67865, shouldn't we be installing a pbc file instead of pir? 19:53
aloha OK. I'll deliver the message.
dalek nxed: r812 | NotFound++ | trunk/pir/winxed_ (2 files):
update installable files
 19:54
kid51 Does anyone know how to run the tests in ext/Parrot-Embed/t?
kid51 is looking at trac.parrot.org/parrot/ticket/488 19:55
dalek rrot: 155e2eb | nwellnhof++ | / (29 files):
Merge branch 'master' into nwellnhof/unicode_dynpmcs
 20:02
rrot: 8b81777 | nwellnhof++ | / (34 files):
Merge branch 'nwellnhof/unicode_dynpmcs'
rrot: 501582b | nwellnhof++ | NEWS:
Update NEWS
 20:03
cotto dukeleto, ping 20:12
20:18 bubaflub left
kid51 afk 20:18
cotto msg gerd That commit should be dee67865c. 20:20
aloha OK. I'll deliver the message.
dalek sella: 7e015cc | Whiteknight++ | / (8 files):
Implement handling of TODO tests which unexpectedly pass. Reimplement differentiating between Assert:: failures and other unhandled exceptions. Fix problems with identifying test files which prematurely exit. A few other fixes
 20:23
bacek ~~ 20:27
Good morning, humans
whiteknight good morning, bace
bacek
bacek aloha, whiteknight
20:38 plobsing joined
cotto aloha, rosella? 20:39
aloha cotto: rosella is like a patterns and practices library
Coke cotto: I have some bugs in parrot that are blockers for tcl. fyi. 20:42
cotto Coke, is there a report that lists them or a tag that you used? 20:52
Coke report 15, IIRC. 20:53
(it's the same report for languages since ever.
cotto "new patches"?
Coke 16.
cotto ok 20:54
Coke #2022 is the biggest blocker at the moment.
21:03 dwebb joined 21:04 dwebb left
whiteknight Coke: I'll take a look at #2022 later tonight or tomorrow for you 21:09
Coke I'd appreciate it, thanks. 21:10
bacek ~~ 21:15
Anyone object if I'll merge gen_gc2 now?
cotto bacek, did it get any testing with partcl-old?
bacek cotto, partcl-nqp only 21:16
afaik
aloha, partcl?
aloha bacek: partcl is broken, so it's not clear your change breaks it more
bacek not helpful
cotto partcl-old is building again and is close to passing its tests. It'd be a shame to break it further now. 21:17
bacek where is it hosted?
cotto aloha, partcl is also github.com/partcl/partcl 21:18
aloha cotto: Okay.
cotto aloha, partcl?
aloha cotto: partcl is broken, so it's not clear your change breaks it more or github.com/partcl/partcl
bacek error: The requested URL returned error: 500 while accessing github.com/partcl/partcl/info/refs 21:19
looks like github is also broken
Coke, ping
Coke bacek: pong.
bacek Coke, can I have commit bit to partcl? 21:20
Coke want me to double check out partcl-old works on yoru branch?
cotto looks good on my end
Coke bacek: NOOOOOOOOOOOOOOOOOOOOOOOOOOk. just a sec.
bacek Coke, and also check partcl-old on gen_gc2 :)
dalek TT #1933 closed by nwellnhof++: stat opcode causes program crash on Win32 with invalid file descriptor 21:21
TT #1933: trac.parrot.org/parrot/ticket/1933
TT #1836 closed by nwellnhof++: Support non-ASCII characters in filenames
TT #1836: trac.parrot.org/parrot/ticket/1836
TT #1898 closed by nwellnhof++: unicode bytecode filenames
TT #1898: trac.parrot.org/parrot/ticket/1898
Coke bacek: added you to both partcl/ projects. 21:22
bacek Coke, thanks. 21:23
Coke, how many tests are broken?
(on master)
21:23 theory joined, whiteknight left
Coke bacek: for old? moment. 21:23
dalek rtcl: 0d849f9 | coke++ | runtime/builtin/info.pir:
remove optional args not used in called func.
 21:26
rtcl: 5c65027 | coke++ | config/misc/select_ (2 files):
avoid mixed-case HLLs.
bacek Coke, it hangs on cmd_for on gen_gc2... 21:27
Coke, on test 5. 21:28
nopaste "coke" at 192.168.1.3 pasted "partcl (old) test failures on master." (37 lines) at nopaste.snit.ch/33270 21:29
Coke bacek: also cmd_while, yes.
bacek Coke, what about on master?
Did you just ^C them? 21:30
Coke that /is/ master ,yes.
no, I did a ps and killed 'em that way.
(otherwise you kill the harness, not the test.)
(look at the nopaste) 21:31
are you running on gen_gc2 now?
bacek yes
dalek rtcl: 4069e65 | bacek++ | src/pmc/tcl (3 files):
Mark vtables with VTABLE explicitely
 21:32
rrot: d6ce563 | jkeenan++ | / (5 files):
Add test file and test sample file to address trac.parrot.org/parrot/ticket/692.
 21:35
bacek Coke, looks like exception handling is going nuts. With "Null PMC access" thrown from exception handler 21:37
21:38 hudnix left
dalek rrot: e3565e2 | jkeenan++ | t/pir/pbc_fallback.t:
Improve POD.
 21:38
nopaste "bacek" at 192.168.1.3 pasted "partcl-old on gen_gc2" (33 lines) at nopaste.snit.ch/33271 21:41
bacek Coke, looks like same failures on master and gen_gc2
Coke prove-- for NOT REPORTING HOW MANY FAILURES THERE ARE 21:42
bacek: yup. so, no worries on breaking me worse than already.
cotto merge away, then 21:43
bacek Coke, meh :)
moritz bacek: just finished another spectest run of rakudo on gen_gc (linux amd64, with --optimize) -- all good
bacek moritz, good.
cotto moritz, awesomesauce
21:48 bubaflub joined
tadzik is partcl from Parrot Tcl or from Part (of) Tcl? I never know 21:52
nopaste "bacek" at 192.168.1.3 pasted "Coke, cmd_for trace. Looks like a bug in grammar." (175 lines) at nopaste.snit.ch/33274 21:53
dalek rrot: 57afed7 | bacek++ | / (57 files):
GenGC has been arrived.
cotto bacek++ 21:54
tadzik has been arrived?
cotto has been arrived
tadzik I'll have to review my english it seems
cotto that are trues 21:55
Coke bacek: until I get the boxing error fixed, I'm assuming it's that. ;) 22:01
(where things are getting boxed to parrot types, which then misbehave badly.)
22:03 jsut joined 22:08 jsut_ left 22:12 rurban_ joined 22:14 rurban left 22:15 rurban_ is now known as rurban
tadzik "Your branch is behind 'origin/master' by 38 commits, and can be fast-forwarded.", appeared after git co master. Is there any command with which I can summon such message? 22:19
bubaflub git status should show that
moritz tadzik: git co master :-)
tadzik -_- 22:20
could, but usually does not. I'm looking for something that will tell me "you are outdated, or not"
bubaflub tadzik: git remote update origin/master
that'll pull the latest updates
without automatically merging it
tadzik hmm
bubaflub you can also pull in other branches 22:21
er updating other branches
tadzik yes, fetches, but no straightforward message. git-checkout somehow knows, I wonder what it uses
bubaflub tadzik: sorry, i usually use a combo of git remote update and git status 22:23
nopaste "kid51" at 192.168.1.3 pasted "post-gen_gc2 merge: Some failures in t/tools/pmc2cutils/04-dump_pmc.t" (54 lines) at nopaste.snit.ch/33277
bubaflub if the remote update downloads anything, then the status will show. if not, there will be nothing 22:24
there should be a better way to do that, though
kid51 test failures may be do to change in number of pmcs
22:24 zby_home left
bacek erm. I didn't introduce new PMCs in branch 22:24
and pmc2c is Perl5...
sorry, have to run.
kids time
bb in 12 hours 22:25
cotto have fun
kid51 The test failure seems to be related to 'make buildtools_tests'. It doesn't occur if I run 'make' and then call 'prove' on that one file only. 22:40
I will have to look at it later. Not critical.
22:46 rdesfo joined 22:49 hudnix joined, kid51 left 22:53 nwellnhof left
dalek rdinal: 8f0069e | bubaflub++ | src/classes/String.pir:
fix String []=
 22:54
bubaflub afk for sushi 22:56
22:56 fperrad left 23:26 sjn left 23:32 bubaflub left

Powered by Cro and the Raku® Programming Language.

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

Thank you!