Parrot 2.4.0 "Sulfur Crest" Released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | GSOC Students: trac.parrot.org/parrot/wiki/GSoCersStartHere
Set by moderator on 2 June 2010.
00:16 whiteknight joined
dalek rrot: r47408 | mikehh++ | trunk/src/pmc/orderedhash.pmc:
fix g++ build - remove cast to (PObj *) as change from Parrot_gc_mark_PObj_alive to Parrot_gc_mark_PMC_alive
 00:17
bacek pmichaud, did you ever install parrot not in /usr/local??? Looks like it was always wrong logic in finding config.fpmc... 00:20
whiteknight kid51: ping 00:24
kid51 pong 00:27
pmichaud bacek: We _always_ install parrot to not /usr/local 00:29
that's what happens with --gen-parrot -- it installs to parrot_install/
bacek pmichaud, interesting... But path was always hardcoded...
jnthn ... 00:30
bacek pmichaud, r47409 with some ugly workaround. We need "installation" expert for properly fix it. 00:32
pmichaud bacek: it doesn't appear to have been hardcoded in 2.4.0
in 2.4.0, I correctly have: conf_file = "/home/pmichaud/parrot/RELEASE_2_3_0/install/lib/include/config.fpmc" 00:33
bacek pmichaud, hmm...
pmichaud (sorry, 2.3.0)
pmichaud checks 2.4.0
whiteknight kid51: I can help out with the docs, if any still need a'doin'
dalek rrot: r47409 | bacek++ | trunk/config/gen/config_pm/config_pir.in:
Unhardcode installed config.fpmc path slightly.
 00:34
bacek meh
plobsing--
kid51 Thanks. I just posted to list. I'll create a branch once I've finished revising my YAPC talk.
So you can follow along and kibitz.
bacek ops_massacre branch...
kid51 But my head is spinning from a cold today, so I'm not feeling alive enough to write documentation. 00:35
pmichaud yes, I get the correct result in 2.4.0 also 00:36
so it's something since 2.4.0 (yes, ops_massacre is likely)
bacek pmichaud, fixed 00:37
r47410
pmichaud bacek++
yes, that appears to have fixed config.pir 00:38
bacek nope, bacek-- actually. It was my fault.
I just misread diff 00:39
plobsing++ # to keep it even :)
mikehh yes but bacek++ for all the work on coretest 00:42
pmichaud karma pmichaud
purl pmichaud has karma of 3935
pmichaud just curious to see where things were today :)
uh oh 00:43
/home/pmichaud/parrot/trunk/install/bin/pbc_to_exe nqp.pbc
Can't read '/home/pmichaud/parrot/trunk/install/runtime/parrot/include/config.fpmc' : No such file or directory
no joy yet :-(
jnthn Whoa
karma jonathan 00:44
purl jonathan has karma of 1764
jnthn karma jnthn
purl jnthn has karma of 431
jnthn Miiiiles behind!
bacek pmichaud, make reconfig?
pmichaud bacek: I did a fresh rebuild, but yes, I'll try again.
make realclean isn't sufficient?
note that the error I get this time is indeed different from before 00:45
it's now correctly looking in mylocal parrotinstall, but it's looking in runtime/parrot/include (which is wrong) 00:46
bacek pmichaud, sigh... It was me again...
r47411 00:47
bacek definitely need to take a break
kid51 So I see that the 'make coretest' failures are now confined to t/pmc/io.t 00:49
t/pmc/io.t (Wstat: 5632 Tests: 40 Failed: 22)
Failed tests: 1-2, 5-14, 21, 24-25, 27, 29, 31, 33, 36-37
40
Progress!
purl it has been said that progress is progress
jnthn purl. If it's not rude, it's almost certainly useless. 00:50
pmichaud I find it amusing when people somehow claim that purl started "working" again a few days ago. :)
dalek rrot: r47410 | bacek++ | trunk/config/gen/config_pm/config_pir.in:
Fix path to installed config.fpmc
rrot: r47411 | bacek++ | trunk/config/gen/config_pm/config_pir.in:
Sigh... And again. bacek-- to break build.
nopaste "kid51" at 192.168.1.3 pasted "make coretest: Remaining failures are in t/pmc/io.t" (54 lines) at nopaste.snit.ch/20979 00:51
kid51 That was at r47407
mikehh make corevm/make coretest - t/pmc/io.t - Failed tests: 1-2, 5-14, 21, 24-25, 27, 29, 31, 33, 36-37, 40
all other tests PASS (pre/post-config, smoke (#34229), fulltest) at r47408 - Ubuntu 10.04 amd64 (gcc)
t/op/annotate-old.t - TODO passed: 1 in testf
pmichaud kid51: I know people don't say it enough: Your contributions to Parrot are awesome. 00:52
bacek mikehh, just move t/pmc/io into t/dynops :)
kid51 pmichaud: "awesome"? a term rarely applied to kid51, in any respect
pmichaud pick any other positive-sounding term you like, then :) 00:54
mikehh bacek: what about splitting the test - the passing subtests remain - the failing subtests to dynops 00:58
bacek mikehh, better to move related tests to t/pmc/filehandle.t. 01:00
I already stated this process.
mikehh bacek: 'k - that sounds like a viable option 01:01
ok it's after 2am for me - definately need to take a break - cul8r 01:05
dalek kudo: f1042d7 | jonathan++ | src/Perl6/Module/Loader.pm:
Ensure we read in modules as utf8.
 01:06
kudo: 6ba4b22 | jonathan++ | src/Perl6/Grammar.pm:
carp about 'package Foo;'
rrot: r47412 | pmichaud++ | trunk (5 files):
[pct]: Add capability for PAST::Val to generate pasm constants.
 01:22
rrot: r47413 | pmichaud++ | trunk/compilers/pct/src/PAST/Compiler.pir:
[pct]: Refactor internals of PAST::Val pasm constant handling a bit.
rrot: r47414 | pmichaud++ | trunk/compilers/pct/src/PAST/Compiler.pir:
[pct]: Change PAST::Control generation a bit, fix except_[types_severity] includes.
rrot: r47415 | pmichaud++ | trunk/t/compilers/pct/past.t:
[pct]: Fix test; it's "except_types.pasm" and not "exception_types.pasm".
rrot: r47416 | plobsing++ | trunk (2 files):
C function docs for compilers/imcc/pcc.c
01:44 ash__ joined 01:46 ash___ joined
dalek p-rx: 6841519 | pmichaud++ | (2 files):
Adjust regexes to use @?BLOCKPAST dynamic var instead of @?BLOCK global.
 01:53
rrot: r47417 | pmichaud++ | trunk/compilers/pct/src/PAST/Compiler.pir:
[pct]: Change PAST nodes to produce pasm constants instead of integers.
 01:55
02:45 jsut_ joined 02:48 janus joined
dalek kudo: c0efe09 | (Solomon Foster)++ | src/core/Any-list.pm:
Allow min, max, and minmax to handle single-argument block Orderings.
 02:57
03:34 Andy joined 03:39 gbacon joined
dalek rrot: r47418 | plobsing++ | trunk (2 files):
C function definitions for compilers/imcc/reg_alloc.c
 03:51
rrot: r47419 | bacek++ | branches/gc_massacre/src/gc/gc_ms2.c:
Use string allocation functions for buffers.
 04:07
rrot: r47420 | bacek++ | branches/gc_massacre (31 files):
Merge branch 'master' into gc_massacre
04:55 LoganLK joined
dalek rrot: r47421 | plobsing++ | trunk (19 files):
change instances of Interp.stdhandle to make use of stdio.pasm macro constants
 04:58
purl dalek: that doesn't look right
05:27 theory joined
dalek p-rx: c341588 | pmichaud++ | src/NQP/Actions.pm:
Refactor loop exceptions slightly.
 05:41
rrot: r47422 | pmichaud++ | trunk/ext/nqp-rx/src/stage0 (4 files):
[nqp]: Update bootstrap files with pasm constant versions.
 06:05
06:24 JimmyZ joined
pmichaud aaaaaarrrrrrrrgggggggghhhhhhhh! 07:35
okay, I've made a change to src/ops/core.ops ... how do I get that change into core_ops.c ? 07:37
(...after spending 90 minutes trying to figure out why my change to the exit opcode wasn't working)
07:38 fperrad joined
sorear there's a make target for that 07:38
pmichaud "make opsrenumber", perhaps?
sorear the op parser runs in parrot now, so it needs an explicit bootstrap 07:39
no
dynops-bootstrap or so
pmichaud I didn't change a dynop.
I didn't add an op.
I just changed an op's implementation.
sorear you still need to bootstrap the ops
ops2c.pl was rewritten in nqp
pmichaud yes, I know.
I understand.
I just don't know the target.
plobsing make bootstrap-ops 07:40
pmichaud is this documented somewhere, ooc?
plobsing damned if I know. I don't read the docs anyways, they're mostly lies. 07:41
pmichaud it ought to be mentioned in the *.ops files, at least
"If you change this file, be sure to 'make bootstrap-ops' to ..." 07:42
or we should get a make warning if any of the *.ops files are newer than core_ops.c
plobsing ./ops2c --core would also do much of what you're looking to do 07:43
the help on that tool is terse but accurate
dalek rrot: r47423 | jimmy++ | trunk/config/gen/makefiles/root.in:
[makefile] update help for bootstrap-ops
 08:01
rrot: r47424 | pmichaud++ | trunk/ext/nqp-rx/src/stage0 (3 files):
[nqp]: Update bootstrap with full pasm constant versions.
rrot: r47425 | pmichaud++ | trunk (4 files):
[core]: The argument to the 'exit' opcode shouldn't be the exception type.

that for the 'exit' opcode.
rrot: r47426 | pmichaud++ | trunk/compilers/pct/src/PAST/Compiler.pir:
[pct]: Add CONTROL_EXIT to PAST::Control handling.
cotto pmichaud, plobsing is correct. I'll add documentation to the .ops file and to core_ops.c. 08:09
pmichaud cotto++ plobsing++ 08:10
(and yes, "make bootstrap-ops" did solve my problem) 08:11
dalek p-rx: 7adae67 | pmichaud++ | src/NQP/Actions.pm:
Fix except_types error.
p-rx: 1d96f81 | pmichaud++ | src/stage0/ (4 files):
Update bootstrap.
p-rx: 610d80c | pmichaud++ | (4 files):
Bump PARROT_REVISION, rebuild bootstrap with full pasm constants.
p-rx: f4ffc61 | pmichaud++ | (4 files):
Bump PARROT_REVISION, update bootstrap to get latest CONTROL_EXIT fixes.
rrot: r47427 | pmichaud++ | trunk/ext/nqp-rx/src/stage0 (3 files):
[nqp]: Final bootstrap to get CONTROL_EXIT fixes in HLL::Compiler.
 08:18
cotto pmichaud, is there a nice way to get line number information in nqp-rx? 08:25
done 08:33
hopefully that makes life a little less frustrating
dalek rrot: r47428 | cotto++ | trunk (19 files):
[ops] update docs and generated code to make bootstap-ops build target more discoverable
 08:34
cotto JimmyZ, you need to be careful when editing the makefile templates. r47423 should have had a tab instead of spaces. 08:35
It's an unpleasant quirk of makefiles that they fail mysteriously when you use spaces instead of tabs in that context. 08:36
It's fixed in r'28 08:37
'night
cotto is glad that he got to write some kind of code, even if it's simple
08:54 jhelwig joined 09:31 TiMBuS joined 10:17 radu joined 10:50 whiteknight joined
mikehh make corevm/make coretest - t/pmc/io.t - Failed tests: 1-2, 5-14, 21, 24-25, 27, 29, 31, 33, 36-37, 40 10:57
t/compilers/imcc/syn/clash.t - Failed test: 13 in testr
all other tests PASS (pre/post-config, smoke (#34232), fulltest) at r47428 - Ubuntu 10.04 amd64 (g++)
t/op/annotate-old.t - TODO passed: 1 in testf
nopaste "mikehh" at 192.168.1.3 pasted "strange failure with g++ - t/compilers/imcc/syn/clash.t - Failed test: 13 in testr" (35 lines) at nopaste.snit.ch/20983 11:00
mikehh it looks to me as though the test should pass, (it does when built with gcc) 11:01
11:37 davidfetter joined
whiteknight good morning, #parrot 11:54
12:02 lucian joined
davidfetter mornin', whiteknight 12:05
dalek kudo: 24a823f | jonathan++ | t/spectest.data:
Turn on integration/advent2009-day18.t; patrickas++.
 12:11
12:17 jsut joined 12:53 jsut_ joined 12:57 patrickas joined 12:58 patrickas joined
dalek kudo: 53da4c1 | jonathan++ | src/core/Cool-str.pm:
Fix chomp to handle Windows newlines too.
 13:04
13:07 lucian_ joined 13:15 lucian joined 13:21 JimmyZ joined 13:22 kid51 joined 13:57 jan joined 14:38 tetragon joined 14:52 wagle joined
dalek kudo: a3de524 | (Ævar Arnfjörð Bjarmason)++ | src/Perl6/Grammar.pm:
Fix 'package Foo;' error so that it suggest a Perl 6 alternative

didn't suggest what syntax to use instead. Explain that you should use package blocks, or the module declaration syntax. Signed-off-by: Moritz Lenz <moritz@faui2k3.org>		 15:35
16:12 patspam joined
dalek rrot: r47429 | khairul++ | branches/gsoc_instrument (267 files):
Sync with trunk
 16:23
16:27 plobsing joined 16:32 theory joined 17:28 davidfetter joined
nopaste "pmichaud" at 192.168.1.3 pasted "weird string numification in trunk" (19 lines) at nopaste.snit.ch/20998 18:09
pmichaud I get the same results in 1.4.0 18:10
(trying 1.0.0) 18:11
pmichaud files trac ticket 18:12
dalek TT #1674 created by pmichaud++: Strange string (non)numification 18:25
TT #1674: trac.parrot.org/parrot/ticket/1674
18:28 lucian joined
dalek kudo: 1bac469 | pmichaud++ | src/Perl6/ (2 files):
Move fake signature handling into <colonpair>, eliminate term:sigterm .
 19:02
kudo: 3d19599 | pmichaud++ | build/PARROT_REVISION:
Bump PARROT_REVISION to get latest exit() fixes and improvements to pasm
19:16 hudnix joined 19:19 tcurtis joined
dalek kudo: 7f2b754 | moritz++ | Test.pm:
[Test.pm] diagnostic output if is_approx fails
 19:51
GeJ Good morning everyone. 20:00
20:03 Psyche^ joined
whiteknight good morning, GeJ 20:04
20:10 hudnix joined 20:15 whiteknight joined
tcurtis pmichaud: ping 20:26
Or anyone knowledgeable about NQP: ping. 20:27
cotto tcurtis, ask and you may get an answer. 20:34
20:38 LoganLK joined
tcurtis Is there any way in a class declaration in NQP to declare a method named "!make"? Other than Foo.add_method("make", sub {...}). If I do "method !make {...}", NQP thinks that there's no signature and panics. 20:39
s/add_method("make"/add_method("!make"/ 20:40
sorear I suppose you meant method !make () { ... }
tcurtis sorear: Yes. That was also a typo in that message. Not in the source, though.
jnthn afaik, NQP doesn't let you do that, though it is valid Perl 6, so I guess it would be OK to make NQP support it, perhaps. 20:41
tcurtis Second question: should the sub I pass to add_method take an additional argument for the invocant or should it just use self as if it were a normally declared method? 20:44
pmichaud: unping 20:45
jnthn tcurtis: If it's a sub, should have an argument for the method.
tcurtis Thanks. 20:51
jnthn er, for the *invocant*
...wish I could remember where I left my brane. 20:52
sorear parrot, like most OO platforms, handles method calls as just subs with an invocant first 20:53
NQP has special sugar for an implicit, sigil-less self parameter in functions declared using "method"
whiteknight sorear: not entirely true. The invocant is typically kept separately
sorear in real Perl 6, you can even get that sugar on closures - anon method :: (...) { ... }
whiteknight: oh dear, I thoght we killed the last trace of that 20:54
jnthn whiteknight: Really?
I thought that died a death in the last PCC refactor.
dalek kudo: 4bc8125 | pmichaud++ | src/Perl6/Actions.pm:
Refactor circumfix:sym<{ }> as hash constructor, fixes RT #74050
 20:55
kudo: 06723b4 | pmichaud++ | Test.pm:
Merge branch 'master' of github.com:rakudo/rakudo
whiteknight hmmm, calls can be made either way. I can't remember exactly how it is passed internally 20:56
jnthn sorear: You can even just write "method (...) { ... }" for an anonymous method. 20:58
tcurtis sorear: I think method (...) {...} works in NQP.
jnthn sorear: Though anon lets you give the anonymous method a name.
(yes, I know...)
:-)
dalek p-rx: 820f547 | pmichaud++ | (2 files):
Empty angles "< >" return a list (see RT #73772).
 21:16
21:21 lucian_ joined 21:26 patspam joined 21:32 Coke joined
Coke returns. 21:32
purl somebody said returns was supposed to use 'iterate over arguments' ?
21:38 hercynium joined 21:39 eternaleye joined
dalek kudo: 841262f | pmichaud++ | src/Perl6/Actions.pm:
Empty <> quotes with only whitespace should return empty parcel. (RT #73772)
 21:47
kudo: 359dbdf | masak++ | (5 files):
[Buf] improvements to Str.encode

to the bytes in a Parrot string, pending the Parrot people's long-term solution. Thanks to it, we now do string encoding of UTF-8 strings right.		 21:59
rrot: r47430 | tcurtis++ | branches/gsoc_past_optimization (7 files):
Matching PAST::Pattern::Node subclasses returns a PAST::Pattern::Match object.
 22:43
22:58 patspam joined
dalek rrot: r47431 | tcurtis++ | branches/gsoc_past_optimization (3 files):
Matching PAST::Pattern::Constant and PAST::Pattern::Closure also return PAST::Pattern::Match objects.
 22:59
rrot: r47432 | bacek++ | trunk (2 files):
Fix numification of short strings. Closes #1674
 23:06
TT #1674 closed by bacek++: Strange string (non)numification
TT #1674: trac.parrot.org/parrot/ticket/1674
rrot: r47433 | tcurtis++ | branches/gsoc_past_optimization/t/library/pastpattern.t:
Added a test for correct .from when matching a subtree of a PAST::Node.
 23:22
rtcl: 10737a0 | Coke++ | README:
Clarify "partcl is not dead".
 23:38
rrot: r47434 | tcurtis++ | branches/gsoc_past_optimization/examples/library/pasttransformerdynamic.nqp:
Add #!, copyright, Id, and coda to PAST::Transformer::Dynamic example.
 23:55

Powered by Cro and the Raku® Programming Language.

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

Thank you!