Parrot 1.9.0 "Blue-fronted Amazon" released! | parrot.org | Priorities: merge tt_389_fix and one_make branches | Roadmap: icanhaz.com/parrotroadmap | Latest modified TT's: icanhaz.com/parrotbugs
Set by moderator on 19 January 2010.
00:02 patspam joined
chromatic Release tarball: wgz.org/chromatic/tmp/parrot-2.0.0.tar.gz 00:06
cotto_work nice name 00:09
let's see what fulltest says
chromatic So far, so good.
Same here. 00:12
Hm, now I need my OSL password. 00:14
davidfetter heh
cotto_work don't you just use an ssh key?
davidfetter who's doing packages?
dalek parrot: 9101a84 | (David Fetter)++ | HOWTO:
Added GNU make dependency
chromatic I do use an SSH key, but it's not working for some reason. 00:15
00:15 kurahaupo joined
davidfetter kurahaupo, who was that guy from FSF that talked earlier today? 00:16
smash running fulltest on tarball
davidfetter oh, and hi
chromatic All tests pass for me on tarball.
dalek rrot: r43485 | chromatic++ | trunk (9 files):
Updated documentation and distribution files for 2.0.0 release.
rrot: r43486 | chromatic++ | trunk (5 files):
Updated PBC for 2.0.0 release.
rrot: r43487 | chromatic++ | tags/RELEASE_2_0_0:
Tagged release 2.0.0
davidfetter go, chromatic, go! :) 00:19
chromatic Coke, do you have working SSH access? 00:22
00:29 Whiteknight joined
Whiteknight My build in trunk fails with this: 00:30
Can't find string terminator "'" anywhere before EOF at lib//Parrot/OpLib/core.pm line 35547
looks like it's still building the ops files
I'm trying a fresh checkout 00:31
dalek website: chromatic++ | Parrot 2.0.0 Released! 00:32
website: www.parrot.org/news/2010/Parrot-2.0.0
Whiteknight nope, fresh checkout gives me this problem:
syntax error at lib//Parrot/OpLib/core.pm line 3394, at EOF
Missing right curly or square bracket at lib//Parrot/OpLib/core.pm line 3394, at end of line
cotto_work fulltest on the tarball passes 00:33
dalek rrot: r43488 | plobsing++ | trunk/t/pmc/io.t:
fix TODOed null filehandle test
chromatic If the tarball works and Whiteknight's checkout doesn't....? 00:34
Whiteknight no, false alarm sort of, it doesn't work when I specify a different compiler
(though how that affects my perl .pm files, I have no idea)
chromatic It doesn't; it's probably a #line directive that mentions a .pm. 00:35
Whiteknight it happens during this step: /usr/bin/perl tools/build/ops2c.pl CSwitch --core 00:36
00:36 jan joined
Whiteknight so I don't think that's running any C code 00:36
chromatic Hm, right.
The error probably comes from Text::Balanced then. 00:37
Whiteknight ah, happens on a parallel build 00:38
if I specify a compiler to Configure.pl AND do a parallel build, it breaks
which is a very weird combination 00:39
but I don't think it's a release blocker
dalek TT #433 closed by plobsing++: read on null PMC throws exception - /some crazy exception/ ?
00:39 mtk joined, mtk left
cotto_work Coke, ping 00:39
clock?
purl cotto_work: LAX: Tue 4:39pm PST / CHI: Tue 6:39pm CST / NYC: Tue 7:39pm EST / LON: Wed 12:39am GMT / BER: Wed 1:39am CET / IND: Wed 6:09am IST / TOK: Wed 9:39am JST / SYD: Wed 11:39am EST /
chromatic We're unlikely to fix the build problem until we land one_make. 00:40
davidfetter wonders why only some of those are airport codes
00:46 snl joined
dalek rrot: r43489 | plobsing++ | trunk/src/pmc_freeze.c:
don't need visit_info.id_list when freezing, don't need visit_info.seen when thawing
 00:49
kurahaupo davidfetter: sorry for being slow getting back to you. The short answer is "dunno" (because I was late arriving). Are you here, or would you like me to hunt it down for you?
davidfetter i'm not there :( 00:50
davidfetter is 3 hours away by tz and 13 by jet
cotto_work thinking?
purl Oooh he is soooo fine!!!
cotto_work thinking...
purl Oooh he is soooo fine!!!
00:56 cotto_w0rk joined
chromatic Alright, that tarball will be the official tarball when I can upload it. 00:59
01:04 cconstantine joined 01:06 abqar joined
Coke cotto_work: pong 01:07
chromatic: ssh access to...?
chromatic Our VM at OSL.
Coke yes. 01:08
chromatic Through a password or an SSH key? 01:09
Coke or rather, so i thought. password. checking...
I'm in.
chromatic Do you have time to run the release steps that require access there? 01:10
Coke I have no idea what those steps are, but sure. 01:16
chromatic Step 8 in docs/project/release_manager_guide.pod 01:17
Coke ah. we're talking about different things. 01:18
I have access to parrotvm.osuosl.org
not ftp-osl.osuosl.org
chromatic You can do 10f though. 01:19
Coke yes. (I'm usually the guy that does that anyway) 01:20
chromatic Please do!
Coke url me?
chromatic wgz.org/chromatic/tmp/parrot-2.0.0.tar.gz
Coke, your ftp access password might be the same as your Trac password. 01:29
Coke ... then mightn't yours? =-) 01:33
will try that shortly.
chromatic Nope, I didn't have it set up.
cotto Coke, can work on one_make be parallelized 01:34
?
Coke docs.parrot.org/parrot/2.0.0/html/ 01:35
s/2.0.0/latest/ once the cache updates. 01:36
cotto: yes.
cotto Cool. How can I help?
Coke first, you can write up the todo list! 01:37
Coke ducks.
cotto 1) add items to todo list (assigned to: Coke)
Coke DAMMIT! 01:38
cotto This is fun! What else can I do?
Coke you have time now? 01:39
cotto Hmmm. make seems extravagantly broken
yes
(in branch)
Coke ok. I'm working on killing parrot_include config step.
cotto what's that entail? 01:40
Coke there's a tools/dev/h2pasm.pl that does about 90% of what it needs to. 01:42
er, s/dev/build/
each input file has multiple potential output files.
cotto so you're filling in the other 90%. Got it.
Coke needs to be updated to take the output filename as an arg, and then again so it only outputs the directives for /that/ output file. 01:43
(and then tools/dev/h2perl.pl needs to be written.)
(which is basically a copy. refactor ahoy.)
cotto nom
Coke And I think I missed a few output files when I removed the existing deps. should double check the input files for all the potential out files and make sure they are all in the makefile. 01:45
writing this up... 01:46
dalek tracwiki: v145 | coke++ | WikiStart 01:50
tracwiki: trac.parrot.org/parrot/wiki/WikiSta...ction=diff
Coke Inevitable 01:51
... Mr. Anderson?
... chromatic as Mr. Smith. 01:52
01:54 lucian joined
dalek tracwiki: v146 | coke++ | WikiStart 02:06
tracwiki: trac.parrot.org/parrot/wiki/WikiSta...ction=diff
Coke cotto: added a few more notes... 02:10
02:11 ash_ joined
cotto thanks 02:12
Coke links to tickets... 02:14
there. 02:15
hack away; I can review commits and squawk if anything goes sideways. 02:16
cotto is what you wrote on the wiki? 02:19
recent changes doesn't show anything
nm. there it is
02:22 cconstantine joined
dalek tracwiki: v1 | coke++ | BuildTaskList 02:23
tracwiki: trac.parrot.org/parrot/wiki/BuildTa...ction=diff
tracwiki: v2 | coke++ | BuildTaskList
tracwiki: trac.parrot.org/parrot/wiki/BuildTa...ction=diff
02:46 ash_ left 02:52 kid51 joined 02:53 kurahaupo joined 03:09 bacek joined 03:11 ash_ joined
ash_ so, i know threads in parrot are still experimental, but i was running the thread's tests and i noticed in my activity monitor on my computer it never says parrot has more than 1 thread 03:15
is that because of how parrot is threading? or am i just not seeing the thread or something?
cotto They might just not be long-lived enough. 03:16
kid51 is back from perlsemny and wonders: Have we released yet?
chromatic More or less yes.
I don't know if Coke has put the tarball in place yet.
kid51 Should I hold off on merging branches to trunk? 03:17
chromatic You can merge if you like.
ash_ if i sleep in the threaded function should I be able to see the thread if i sample the processes? (i am on OS X, i duno if that makes a difference)
ah, yeah adding a long sleep, they are showing up as 2 threads now, cool 03:19
cotto chromatic, I'll upload it.
chromatic Thanks. 03:20
cotto done 03:22
(not sure how long it'll take to show up, though)
I ran the parrot-trigger script, so hopefully it'll be visible soon. 03:24
it's up 03:28
chromatic Grr, and I just thought of a better release name.
"Obtanium"
cotto good name for 2.3
I lolled 03:29
chromatic Let's rack that one then.
cotto You mean you want to change the name or save it for 2.3? 03:31
chromatic The latter.
kid51 one_make branch: Still failing make buildtools_tests (t/tools/pmc2cutils/*.t) 03:44
... which, granted, may simply be that Coke hasn't had a chance to update the tests.
cotto kid51, it doesn't even build atm 03:45
kid51 But the problems causing those tests to fail are also causing a sudden 'make' failure.
Coke oh, I never checked my ftp password.
cotto oic 03:46
Coke chromatic: nope.
chromatic Apparently cotto made it work.
Coke <samaritan snare> 03:47
dalek rrot: r43490 | jkeenan++ | trunk (82 files):
Merge noalignptrs branch into trunk. Eliminates configuration step auto::alignptrs, the results of whose probe were used in only one PMC source code file.
rrot: r43491 | jkeenan++ | branches/noalignptrs:
Branch has been merged into trunk and is no longer needed at HEAD.
purl i already had it that way, dalek.
rrot: r43492 | jkeenan++ | tags/noalignptrs-43150:
Branch corresponding to tag has been merged into trunk. Tag is no longer needed.
Coke one_make isn't right atm.
kid51 must sleep 03:53
purl $kid51->sleep(8 * 3600);
cotto Coke, what's the reasoning behind having h2pasm.pl output one file at a time? If multiple .pasm includes depend on a single .h, doesn't it make sense to generate them all at once?
chromatic, if you want I can add your public key to the ftp server. 03:56
chromatic That'd be lovely.
Coke cotto: write me a make rule that doesn't run the command N times anyway on a || run. 04:30
(which you can do, but it involves adding at least one intermediate rule, and then having non-uniform rules for things that use that tool. 04:31
so, I was going to avoid it.
cotto ok 04:32
chromatic Ugh, typo in the release announcement on parrot.org: Februuary 04:35
dalek rrot: r43493 | cotto++ | trunk/compilers/pirc/src (2 files):
[pirc] make pirc stop breaking every time someone runs headerizer
 04:36
04:41 diakopter joined
diakopter /topic out of date 04:42
moderator Parrot 2.0.0 "Inevitable" released! | parrot.org | Priorities: merge tt_389_fix and one_make branches | Roadmap: icanhaz.com/parrotroadmap | Latest modified TT's: icanhaz.com/parrotbugs 04:42
chromatic What now? 04:43
purl The chalice contains human blood cursed in Death's name. The hooded figure kneeling beside you passes it to you and you have to drink. LOSE 3 HIT POINTS.
diakopter heh
chromatic Human blood? I'd GAIN three hit points.
diakopter is the "Inevitable" Imperator-class, Victory-class, or Executor-class? 04:46
04:46 cognominal joined 04:51 kurahaupo joined
diakopter or Venator-class or Acclamator-class? 04:52
eternaleye chromatic: It's cursed in _death's_ name, so it brings you closer to death! 05:00
chromatic Why would the name "Ralph" be a curse? 05:02
eternaleye chromatic: No, no, it's like "We will make war in the Lord's name", i.e. the action is taken for the sake of the named person/concept, rather than actually using the name to accomplish the action 05:05
chromatic Oh. Ralph was confused. 05:06
treed Dagron dan in ephedyn en aran 05:07
cotto That wouldn't be so bad if it were a homeopathic dilution of human blood cursed in Death's name.
diakopter purl, what? 05:11
purl diakopter: i don't know
diakopter purl, what now
purl i think now is the time for the Kwisatz Haderach to arise and conquer America on a sandworm army
treed I agree, purl. 05:12
purl treed: what?
treed purl: what what?
purl treed: bugger all, i dunno
cotto how now? 05:15
purl brown cow
06:07 bacek joined
cotto hio bacek 06:09
06:27 diakopter left
cotto Coke, ping 06:44
Coke, nm 06:48
I'll just barge ahead and let you stop me if needed.
dalek rrot: r43494 | tene++ | branches/pge_no_namespace_methods/config/auto/icu.pm:
Cherry-pick "Minor workaround for icu-config" from trunk.
 07:01
07:03 uniejo joined 07:11 woosley joined
dalek rrot: r43495 | cotto++ | branches/one_make (2 files):
[h2pasm] make h2pasm smart enough to write exactly one output file per invocation and start using it that way
 07:17
rrot: r43496 | cotto++ | branches/one_make/config/gen/makefiles/root.in:
[makefile] add makefile rules for all generated pasm/pm includes
Coke guess I should have written that list down sooner. 07:21
cotto, I think you may have fixed the build. 07:25
cotto++ 07:26
Posit: we can kill the TODO file. 07:31
cotto: if h2pasm and h2perl are the same script, h2macro might be a better name. 07:33
dalek rrot: r43497 | coke++ | branches/one_make/tools/dev/checkdepend.pl:
fix to pir dep search, add todo, minor cleanup
 07:34
rrot: r43498 | cotto++ | branches/one_make/tools/build/h2pasm.pl:
[h2pasm] minor typo fix
07:34 cognominal joined
cotto Coke, it's not quite fixed but it's closer 07:35
07:36 bacek joined
cotto and this gets it closer 07:44
dalek rrot: r43499 | cotto++ | branches/one_make/config/gen/makefiles/root.in:
[makefile] add all generated pasm includes as explicit build dependencies
 07:50
Coke an aside, note that the listed order of files doesn't mean they (must) get built first. 07:57
and cotto++
must zzz.
07:58 cotto joined
cotto It's great to finally have some Parrot hackery going on. 08:04
dalek rrot: r43500 | cotto++ | branches/one_make/config/gen/makefiles/root.in:
[makefile] make generated libraries explicitly depend on generated pasm includes

And now it builds!		 08:06
rrot: r43501 | tene++ | branches/pge_no_namespace_methods (2 files):
Fix two method/ns confusions
rrot: r43502 | tene++ | branches/pge_no_namespace_methods/compilers/pge/PGE/Exp.pir:
[pge] Generate code to fetch methods from the protoobject instead of directly from the namespace for ::-delimited rules
rrot: r43503 | tene++ | branches/pge_no_namespace_methods/t/compilers/pge (3 files):
[pge] Fix some method/ns confusion in tests
rrot: r43504 | cotto++ | branches/one_make (2 files):
[h2inc] rename h2pasm to h2inc, update manifest and makefile template
rrot: r43505 | cotto++ | branches/one_make/config/gen/makefiles/root.in:
[makefile] fix some copy/paste goofs
Coke cotto: line 235 of root.in has a tab in it. 08:12
cotto they're insidious
thanks 08:13
Coke and the PASM_INCLUDES are no longer generated by Config, neh?
and now REALLY must sleep
cotto night
I'll check
I'm checking to see what happens when I give config_includes an empty list of stuff to generate 08:14
clock? 08:15
purl cotto: LAX: Wed 12:15am PST / CHI: Wed 2:15am CST / NYC: Wed 3:15am EST / LON: Wed 8:15am GMT / BER: Wed 9:15am CET / IND: Wed 1:45pm IST / TOK: Wed 5:15pm JST / SYD: Wed 7:15pm EST /
Tene chromatic: Everything compiles in the branch, and after this last commit, PGE's test suite passes successfully. 08:18
Be glad my plans fell through this evening :) 08:19
I sleep now.
dalek rrot: r43506 | tene++ | branches/pge_no_namespace_methods/t/compilers/pge/pge_examples.t:
[pge] One more test
 08:22
rrot: r43507 | tene++ | branches/pge_no_namespace_methods (2 files):
[pge] Fix PGE::Text::bracketed and its test
rrot: r43508 | tene++ | branches/pge_no_namespace_methods/runtime/parrot/library/PGE/Util.pir:
Fix PGE::Util's die and warn. warn is untested, as are apparently the other two un-fixed subs in PGE::Util.
rrot: r43509 | cotto++ | branches/one_make/config/gen/makefiles/root.in:
[makefile] fix some makefile template goofs
rrot: r43510 | cotto++ | branches/one_make/config/gen/parrot_include.pm:
[parrot_include] make gen::parrot_include not generate any files
cotto actually, that's not quite true. Apparently it still insists on generating some stuff.
08:32 szabgab joined 08:33 riffraff joined
chromatic I like this. I make a release, go cook dinner, watch some TV, play some guitar, and other people do stuff! 08:43
cotto Bah. There's a mysterious bus error that occasionally manifests itself and make has to be run multiple times to get a complete build. 08:49
sometimes
purl sometimes I think you're off your rocker.
cotto no, sometimes is <reply>
purl okay, cotto.
cotto time for sleeeeeeep 08:50
08:50 woosley left
Tene chromatic: if you can identify additional failures in that branch, that would help a lot. 08:50
I haven't tried a full test run yet.
cotto msg Coke I think gen::parrot_include can be excised safely now. It's a noop now so any oddities are probably not related to it. 08:54
purl Message for coke stored.
cotto really time for sleep
night
chromatic Tene, I'll try to cherry pick to the TT #389 branch. 08:55
dalek rrot: r43511 | cotto++ | branches/one_make/config/gen (2 files):
[configure] turn gen::parrot_include into a noop and fix some more dependencies
09:03 eternaleye joined 09:12 bacek joined 09:40 AndyA joined 10:06 bacek joined 10:16 eiro joined 10:17 eiro left 10:19 eiro joined 10:46 eternaleye joined 10:58 eternaleye joined 11:55 ttbot joined 12:05 bluescreen joined 12:06 cognominal joined 12:11 cconstantine joined 12:13 ruoso joined 12:21 bacek joined 12:26 payload joined 12:50 payload left 12:54 iblechbot joined
dalek rrot: r43512 | gerd++ | trunk/ports/fedora (1 files):
update the spec file and patch to the current stable release (2.0.0)
 12:57
13:05 davidfetter joined 13:11 cognominal joined
Coke msg cotto: if I gut that step in branch, I get a build failure on: runtime/parrot/include/opengl_defines.pasm 13:35
purl Message for cotto stored.
Coke cotto: ah. that's because that file is created by config, but only conditionally. 13:37
tewk_ I know that pulling versions one at a time over the network from a svn server takes a long time. 13:41
oops
dalek rrot: r43513 | coke++ | branches/one_make/config/gen/makefiles/root.in:
These files are already conditionally added below;
 13:46
Coke msg cotto: fixed that, removed gen::parrot_includes. We'll probably need kid51's help to remove some config tests now. 13:48
purl Message for cotto stored.
Coke msg kid51: in the one_make branch, "perl Configure.pl --test" is now failing some tests because they assume things like lib/Parrot/Pmc2c/PCCMETHOD_BITS.pm are generated at config time. (in branch, this is no longer the case). I haven't dug into the tests, but my gut feeling is we can kill any tests that are now failing as a result. 13:49
purl Message for kid51 stored.
Coke anyone here using ncgen? 13:52
er, ncigen?
dalek rrot: r43514 | coke++ | branches/one_make (12 files):
Remove now-obsolete config step gen::parrot_includes

Note: "perl Configure.pl --test" now fails some tests because it is trying to verify things that no longer happen at Configure time.		 14:02
tracwiki: v3 | coke++ | BuildTaskList 14:07
tracwiki: trac.parrot.org/parrot/wiki/BuildTa...ction=diff
14:15 payload joined 14:20 plobsing joined 14:22 eternaleye joined 14:29 payload joined 14:37 payload joined 14:38 whiteknight joined 14:41 eternaleye joined
whiteknight chromatic++ 14:42
15:04 eternaleye joined 15:10 payload joined 15:17 PacoLinux joined 15:30 bubaflub joined 15:47 eternaleye joined 15:48 Psyche^ joined 15:56 eternaleye joined 15:58 theory joined 16:05 payload joined 16:06 eternaleye joined 16:21 eternaleye joined 16:24 clinton joined 16:29 ash_ joined
cotto good morning 16:55
whiteknight good morning cotto
16:56 theory joined 17:00 eternaleye joined 17:03 payload joined
Coke cotto: hio 17:16
cotto hio coke
Coke I think we need to get perl Configure.pl --test passing before a merge back. I think we are also missing some deps. 17:18
hurm, though that may have been due to the opengl issue.
nope. 17:23
Bad name after experimental' at lib//Parrot/OpLib/core.pm line 13.
cotto sounds like a good excuse to rip out the pmc2c tests
Coke at least some of them, yah. 17:24
make TEST_JOBS=5 -j5 test - this dies with the above error.
17:24 eternaleye joined
cotto I'm off to w0rk so no hacking for me until this evening. 17:25
Coke I think TT#1214 is closable, even on trunk. 17:26
17:35 smash joined
smash hello everyone 17:36
whiteknight hello smash 17:39
dukeleto mornin' 17:44
Coke: i got the same build error with -j :(
17:51 davidfetter joined
Coke dukeleto: on one_make? 17:53
./away
dukeleto Coke: no, on parrot 2.0.0 17:59
Coke well, that sucks. 18:06
purl The rock is now off.
Coke but, there are a ton of errors there waiting to happen. :|
dalek rrot: r43515 | coke++ | branches/one_make/config/gen/makefiles/root.in:
These files are not created by Configure.pl anymore;
 18:08
rrot: r43516 | coke++ | branches/one_make/config/gen/makefiles/root.in:
this file is still handled by config
Coke dukeleto: what's the error?
purl the error is KABOOM!
18:08 eternaleye joined
Coke (is there a ticket, and/or is that error not present in the one_make branch.) 18:09
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#31801), fulltest) at r43514 - Ubuntu 9.10 amd64 (g++ with --optimize) 18:14
dukeleto Coke: i will attempt to reproduce soon
Coke mikehh: I think cotto fixed the bug you reported in one_make branch; can you retry? 18:15
wtf. login to trac. go to buildtasklist. edit. save. changes NOT saved, and login button showing again. click login... changes are saved. back to main page. go to buildtasklist, changes indeed saved. 18:16
mikehh coke: sure - got to go out for a bit - will do it in an hour or so
Coke ;(this is the same issue I ahd trying to edit a page weeks ago during parrotsketch.)
;(but this time I figured out how to save.)
dalek tracwiki: v4 | coke++ | BuildTaskList 18:17
tracwiki: trac.parrot.org/parrot/wiki/BuildTa...ction=diff
tracwiki: v5 | coke++ | BuildTaskList
tracwiki: trac.parrot.org/parrot/wiki/BuildTa...ction=diff
darbelo Coke: ping 18:32
Coke pong 18:34
darbelo Does one_make have any pir-level dep-checker?
Coke yes, but it's untested and probably wrong. 18:37
perl tools/dev/checkdepend.pl
darbelo Oh, I thought that was c only.
Coke in trunk it is.
It is currently reporting errors in branch. 18:38
darbelo I just found one missing dep the hard way. I'll fix that an then try to make the tool happy. 18:41
Coke first see if the tool complains? 18:42
darbelo Odd: ok 135 - compilers/nqp/src/Grammar_gen.pir has no dependencies 18:44
But it load_bytecode()s PGE.pbc 18:45
And the build fails without that dep
Coke ah. yah, the _gen dance makes it very hard to track deps. 18:46
18:46 joeri joined
Coke my plan of attack there is to change the build to facilitate dep checking. =-) 18:47
18:47 payload joined
dalek tracwiki: v6 | coke++ | BuildTaskList 18:50
tracwiki: trac.parrot.org/parrot/wiki/BuildTa...ction=diff
davidfetter hello 18:57
purl hi, davidfetter.
18:57 rblasch joined
dukeleto davidfetter: howdy 18:57
davidfetter anybody know whether gerd pokorra is still in charge of fedora packages?
dukeleto, what a pleasant surprise :) 18:58
dalek rrot: r43517 | darbelo++ | branches/one_make/compilers/nqp/Rules.mak:
Add a missing dep, note that checkdepend.pl doesn't correctly detect this case yet.
 18:59
19:01 eternaleye joined 19:06 plobsing joined
dalek tracwiki: v7 | coke++ | BuildTaskList 19:08
tracwiki: trac.parrot.org/parrot/wiki/BuildTa...ction=diff
19:09 plobsing_ joined 19:15 plobsing_ left 19:16 ash_ joined
cotto_work Coke, in one_make do you want as many dependencies to be generated as possible? 19:16
I was tempted to make all those h2inc rules be generated, but it was late enough that I hardly trusted myself to get the explicit versions right. 19:17
19:18 ash_ joined 19:23 contingencyplan joined
dalek kudo/master: 1d49284 | duff++ | docs/release_guide.pod:
[docs] Add tentative release dates through the end of 2010
 19:28
19:33 japhb joined 19:35 eternaleye joined 19:40 kurahaupo joined
kurahaupo Good morning from LCA2010 (TZ=+13h00)! 19:41
cotto_work good morning back at you 19:43
I hope you're enjoying LCA!
19:44 kurahaupo1 joined
cotto_work Mmmmm. Conference wifi. 19:44
kurahaupo1 grumbles about WiFi/3G transitions dropping connections
19:46 eternaleye joined
kurahaupo1 I have some very strange debugging output from the "array unshift" issue (TT#1399). Anyone else looking at it? 19:46
Coke cotto_work: yes, eventually all the deps that can be generated should be. 19:52
for now, doing them explicitly and verifying them with checkdepend is ok.
(and eventually checkdepend guts will generate instead of verify)
pmichaud note to #parrot folks: Rakudo's ng branch doesn't appear to build/run with Parrot 2.0.0 release. 19:53
Coke has it been working with trunk?
pmichaud we tried cherry picking the callsignature patches from rakudo master into rakudo ng branch, but that doesn't seem sufficient
Coke: I don't know, we've been building ng against the 1.9.0 release
dukeleto hmmm
Coke (right, that was most of the issues I knew about.)
19:54 chromatic joined
Coke and, speak of the devil. 19:54
purl Eeeeek! Where?!?!
Coke points at chromatic.
dukeleto pmichaud: what kind of errors?
dukeleto blinks innocently
pmichaud paste.lisp.org/display/93651 19:55
those errors are after cherry picking the rakudo master branch changes into the ng branch
without the cherry pick, it doesn't build at all
Coke so, not build errors, but test failures? 19:56
pmichaud with the cherry pick, we get test failures that weren't present with 1.9.0
but yes, test failures
come to think of it though, the problem could also be the interface change for contexts 19:57
I'm not sure Tene++ updated those as well 19:58
dukeleto pmichaud: do you need testing on any platforms?
pmichaud dukeleto: just the standard ones
I think we've decided that we're going to hold Rakudo's January release for up to a week to resolve a number of different issues, and to make the ng branch into the master branch.
dukeleto pmichaud: what is *standard* for rakudo?
pmichaud dukeleto: basically, anything that Parrot runs.
Coke linux/osx/win?
dukeleto pmichaud: i can test on darwin+ppc, for instance, which is not tested often 19:59
dalek tracwiki: v121 | fperrad++ | Languages
tracwiki: update Parrot version 2.0.0
tracwiki: trac.parrot.org/parrot/wiki/Languag...ction=diff
dukeleto is just now finding out about NekoVM nekovm.org/ 20:00
Coke meows. 20:01
cotto_work I didn't know the "one VM to rule them all" space was so crowded.
Coke it's actually N VMs. 20:02
dukeleto cotto_work: yep. when did Neko start? 20:03
20:03 ascent joined
cotto_work no idea. I found out about it 2 minutes ago. 20:04
well, 5
dukeleto cotto_work: me too 20:05
cotto_work interesting. Their faq has a question "How is Neko different from Parrot?"
nekovm.org/faq
20:05 smash joined 20:06 kurahaupo joined
pmichaud phone 20:06
PerlJam cotto_work: yeah, but the answer to that question reads like "parrot sucks, but neko doesn't" 20:09
dukeleto there is a lot of misinfo about parrot on that site
20:13 kurahaupo joined
cconstantine congratz on 2.0 :) 20:13
cotto_work and they're using cvs.
It's hard to take a project seriously that's still doing that.
chromatic Like OpenBSD?
cotto_work touche 20:14
ash_ neko's faq hasn't been updated since 2007 so.... is it still active?
20:15 theory joined
dukeleto ash_: good point 20:15
ash_ the sites in doku so if you check the meta tags it has the date the page was updated
PerlJam nekovm.org/?do=recent 20:16
ash_ ah so it has been updated recently, well they seem misinformed about parrot a bit 20:17
PerlJam their mailing list seems active too: lists.motion-twin.com/pipermail/neko/ 20:19
(looks like a good place to write to educate them about parrot :)
20:21 cotto joined 20:23 bluescreen joined
dukeleto joins yet another mailing list.... 20:24
cotto_work It'll be interesting to see how they respond. \\ 20:25
20:25 payload joined, bacek joined 20:34 snl joined
dukeleto looks like neko has some kind of JIT, which could be interesting 20:34
neko has *.vcproj files in their repo. hmmmm 20:35
20:37 snl left 20:39 eternaleye joined 20:40 cotto joined 20:42 kjeldahl joined 20:54 cognominal joined 20:56 rblasch_ joined 21:01 plobsing joined 21:02 cotto joined
dukeleto just sent a tentacle of knowledge to the neko list 21:12
21:12 eternaleye joined
GeJ Good morning everyone! 21:14
21:15 bacek joined 21:17 kurahaupo1 joined 21:18 payload1 joined
dukeleto GeJ: howdy 21:27
GeJ dukeleto: Hiya. 21:29
dukeleto GeJ: top of the localtime() to ya 21:31
GeJ clock? 21:38
purl GeJ: LAX: Wed 1:38pm PST / CHI: Wed 3:38pm CST / NYC: Wed 4:38pm EST / LON: Wed 9:38pm GMT / BER: Wed 10:38pm CET / IND: Thu 3:08am IST / TOK: Thu 6:38am JST / SYD: Thu 8:38am EST /
ash_ its time for beer somewhere 21:39
dukeleto it is always scotch:30 in my locality 21:44
darbelo Quarter to vodka on this hemisphere ;) 21:46
dalek rrot: r43518 | mikehh++ | branches/one_make/config/gen/makefiles/root.in:
fix codetest failure - trailing whitespace
 21:49
ash_ does parrot follow the normal 3 part versioning schema? like major.minor.security fixes? 21:54
darbelo Not really.
major gets increased after minor hits 9. 21:55
ash_ gotcha, so your just counting upwards then?
cotto_work actually major gets increased yearly
(it just happened to increase after .9 last year)
ash_ so, 3 will be sometime in 2011? 21:56
darbelo Oh, didn't know that.
purl didn't know that. are you reading the docs for the new version?
cotto_work we haven't needed to make any x.y.1 releases but the possibility is there.
ash_ just curious, thats cool 21:57
cotto_work 3.0.0 will be jan 18 2011
PerlJam cotto_work: you mean, there hasn't been any point in making x.y.1 releases yet :)
cotto_work I've actually got a #ps question about that, but it can wait until #ps.
chromatic You could ask now. 21:59
darbelo Or you can travel back to yesterday too.
PerlJam A real good reason for a x.y.1 release might be ... say Rakudo won't work against an x.y.0 release without patching parrot. Then a quick x.y.1 release that Rakudo can use would be a Good Thing IMHO
chromatic Exactly. 22:00
cotto_work THe question is about forming a dedicated security list for Parrot for high-impact vulnerabilities. I don't think it'll become relevant until a Parrot HLL starts to be widely used. 22:02
dukeleto cotto_work: i like the idea of a security task force, or something
cotto_work yup 22:03
chromatic Perl 5 has one. It's private by default. That's common and seems workable. 22:04
darbelo First task: write a safe exec/spawnw/etc 22:05
cotto_work first of many
dukeleto darbelo: thanks for volunteering!
darbelo dukeleto: It *is* on my TODO list. 22:07
22:11 kurahaupo joined 22:12 NotFound joined
NotFound hola 22:13
purl privet, NotFound.
22:16 bacek joined
ash_ does anyone in here have rakudo? i was just curious if you could run something and let me know whats up with this script, its regarding ParrotThreads 22:17
PerlJam aye 22:18
22:18 davidfetter joined
nopaste "tene" at 67.137.148.232 pasted "A script like this, ash?" (9 lines) at nopaste.snit.ch/19311 22:20
22:22 mikehh_ joined
ash_ well, almost but basically that 22:22
that exact one gives me a error with fix_lex
nopaste "mikehh" at 81.149.189.7 pasted "test failures in one_make branch at r43517" (96 lines) at nopaste.snit.ch/19312 22:23
Tene ash_: you claimed that threads no longer failed in rakudo. That fails for me with the same assert failure documented in 757
ash_ Tene: nopaste.snit.ch/19313 is what I am running 22:24
cotto_work mikehh_, those are known failures. We need to tri^H^H^Hget kid51 to fix them since he's generally most familiar with that chunk of code.
Tene ash_: can you try running this: trac.parrot.org/parrot/raw-attachme.../tt757.pir
cotto_work I may take a stab at it tonight but my tuits are iffy.
ash_ but, yeah, i am kinda retarded, earlier i didn't realize i had left off the & on bl (line 2) and it was actually just executing the block
mikehh cotto_work: I mentioned the missing file to Coke before and he said he was putting it on his TODO list 22:26
cotto_work perlcritic.t doesn't seem to like return undef 22:27
Tene ash_: Yes, I would love to see tt757 fixed, but I really don't understand that part of parrot well enough to evaluate the patches there. 22:28
cotto_work mikehh, 0 would work just as well 22:29
mikehh all the post-config failures have something like: Error: Can't locate Parrot/Pmc2c/PCCMETHOD_BITS.pm in @INC 22:31
darbelo cotto_work: Isn't a bare return better? 22:32
cotto_work The problem is that PCCMETHOD_BITS.pm used to be generated during configuration but now it's generated during the build.
sure
Configure-time tests that rely on PCCMETHOD_BITS.pm (directly or indirectly) will break. 22:33
mikehh then we need to remove those test, or modify then in post-config - well kid51 probably :-} 22:34
22:38 cconstantine joined
darbelo mikehh: I don't have perlcritic on this box, but r43519 should shut it up about h2inc.pl 22:40
ash_ if your in the gdb and you have a pmc pointer, can you get the gdb to print the data from the pmc?
chromatic p *pmc 22:41
22:42 joeri left
chromatic If you need specific data, you probably have to cast one of its data members to the appropriate type. It'd be nice if we had gdb macros for that, but I don't know if we do. 22:42
ash_ hmm, so, if its a hash, and i want to see the first value in it, i'd probably cast the hash data pointer to a? ... 22:45
chromatic (Hash *)(pmc->data)
That might not be right; might have to dereference pmc->data further. 22:46
plobsing (Parrot_Hash_attributes *)(pmc_data)->hash // I think 22:48
chromatic That looks safer yes.
plobsing s/_/->/
ash_ yeah, thats working, now if i do ->hash->entries it prints a sane number 22:49
davidfetter anybody heard from the rakudo folks about an upgrade since 2.0.0 rolled? the fedora people seem concerned about this 22:53
dalek rrot: r43519 | darbelo++ | branches/one_make/tools/build/h2inc.pl:
Change 'return undef' to a bare return. This should apease perlcritic.
 22:54
purl dalek: that doesn't look right
chromatic pmichaud said they're going to target 2.0.0, and they'll probably delay the release a week to make rakudo-ng the master.
If we have to roll 2.0.1 to make Rakudo work, we will. 22:55
22:58 patspam joined
pmichaud bug in parrot 23:00
outer = outer_ctx['current_sub']
needs to be
outer = getattribute outer_ctx, 'current_sub' 23:01
compilers/pct/src/PCT/HLLCompiler:596
chromatic I thought bacek made a commit to make that work.
pmichaud I thought we were going to have a cycle where both worked also... apparently we don't 23:02
at least, I'm _guessing_ that's the source of my current problems -- checking now
23:04 kurahaupo joined
pmichaud yes, it is. 23:06
CallContext['current_sub'] is returning pmcnull 23:07
ng can work around that for now. 23:09
dalek rrot: r43520 | plobsing++ | branches/pmc_freeze_with_pmcs:
create a branch to convert visit_info into PMCs
 23:10
23:20 payload joined 23:25 Whiteknight joined
dalek rrot: r43521 | darbelo++ | trunk/config/init/hints/openbsd.pm:
Parrot will only use clock_gettime() to obtain timin ingormation if _POSIX_TIMERS is defined. OpenBSD provides clock_gettime() but doesn't #define _POSIX_TIMERS. Rather than add to the macro soup in that file let's fake it on the hints file and avoid producing sub-optimal timing info.
 23:43
23:49 preflex joined
pmichaud for #parrot folks: I have the ng branch running on 2.0.0 now, thanks. 23:57
cotto_work great!
darbelo pmichaud++
chromatic No changes necessary in Parrot 2.0.0? 23:58
pmichaud well, not for the current level of ng :-)
there's a bug in PCT::HLLCompiler in 2.0.0, but ng can work around it.
we may still find some parrot changes that we'll want for the January release, but I don't think it's likely.
(for the January Rakudo release, that is) 23:59
anyway, for the moment Parrot 2.0.0 looks fine for us as it is
chromatic Excellent.
pmichaud we'll know more over the next few days
chromatic I'll work on the TT #389 merge then. I think that's ready soon.

Powered by Cro and the Raku® Programming Language.

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

Thank you!