Parrot 1.9.0 "Blue-fronted Amazon" released! | parrot.org | Roadmap: icanhaz.com/parrotroadmap | Latest modified TT's: icanhaz.com/parrotbugs
Set by moderator on 16 December 2009.
00:02 bacek joined 00:16 cotto joined 00:27 ash_ joined
ash_ does parrot's versioning system do the major.minor.secuity_fixes way of doing things or does it just add one? 00:31
i guess i am asking because the next version is 2.0, is that breaking any backwards compatability? 00:32
mikehh trac.parrot.org/parrot/wiki/ParrotVersionNumbers 00:33
although I think we decided at a parrotsketch meeting that we will do three month deprecation cycles for the forseeable future 00:35
What would happen is that if there was a critical problem with 2.0 say we might release 2.0.1 to fix this 00:38
dalek TT #1399 created by plobsing++: Array unshift/access broken 00:40
ash_ mikehh: okay, thats cool, just curious, i mean parrot evolves very quickly so it would make sense not really to follow the 3 number schema since it would mean you'd be changing things so rapidly 00:50
00:52 ash_ left
Coke to be clear, we allow for the third level of versioning, but it's never come up. 00:53
(since 1.0; before that we were just making stuff up anyway.) 00:54
Coke is getting a failure in PCT after juggling the make process. 01:01
svn hang? 01:24
dalek rrot: r43294 | coke++ | branches/one_make (9 files):
convert pct.in to an include'd makefile, including it in the main makefile.
 01:37
purl I don't know how to convert pct.in to an include'd makefile, including it in the main makefile..
Coke stares at the complicated PGE build and wonders if there is a way to simplify it. 01:39
02:25 payload joined
dalek rrot: r43295 | coke++ | branches/one_make (6 files):
Convert pge's recursive make into an included one.
 02:25
03:47 JimmyZ joined 04:38 mikehh joined
dalek rrot: r43296 | coke++ | branches/one_make (8 files):
convert tge's recursive make into an include; add some deps in the library.
 04:52
purl I don't know how to convert tge's recursive make to an include; add some deps in the library..
Coke anyone about? 05:20
purl goes postal or
Coke fixes it 05:25
Coke has nearly eliminated "compilers.dummy" in the one_make branch. 05:27
dalek rrot: r43297 | coke++ | branches/one_make (9 files):
convert json's recursive make into a top level include
 05:40
purl I don't know how to convert json's recursive make to a top level include.
rrot: r43298 | coke++ | branches/one_make/compilers/json:
fix svn:ignore for this dir.
rrot: r43299 | coke++ | branches/one_make (5 files):
pull imcc build rules out into an includable makefile and clean them up.
 06:13
rrot: r43300 | coke++ | branches/one_make (7 files):
Fix svn properties on new files
Coke does order of definition of rules matter in make? if I refer to a variable I haven't set yet, it's just "" at that point, neh? 06:14
Coke ponders splitting up the new Makefile.mak's into setup/rules 06:19
06:52 bacek joined 06:54 mikehh joined 07:58 szabgab joined 08:30 iblechbot joined 08:45 bacek joined 09:28 fperrad joined 09:41 JimmyZ joined 09:46 TiMBuS joined 09:56 bacek joined
bacek o hai 10:00
dalek rrot: r43301 | bacek++ | branches/boehm_gc/compilers/imcc/main.c:
Fix one more free vs mem_sys_free in IMCC...
rrot: r43302 | bacek++ | branches/boehm_gc/compilers/imcc (2 files):
And more realloc vs mem_sys_realloc fixes in IMCC.
rrot: r43303 | bacek++ | branches/boehm_gc/compilers/imcc/optimizer.c:
Fix free vs mem_sys_free in IMCC optimizer.
bacek Tene, around? 10:12
dalek rrot: r43304 | bacek++ | branches/boehm_gc/src/pmc/packfilefixupentry.pmc:
Fix strdup vs mem_sys_strdup in PackfileFixupEntry PMC.
 10:33
rrot: r43305 | bacek++ | branches/boehm_gc/src/pmc/os.pmc:
Fix mem_sys_free vs free usage in OS.cwd. getcwd doesn't allocate memory using mem_sys_alloc, so don't try to use mem_sys_free to free it.
rrot: r43306 | bacek++ | branches/boehm_gc/compilers/imcc/pbc.c:
One more free vs mem_sys_free usage in IMCC.
rrot: r43307 | bacek++ | branches/boehm_gc/src/dynpmc/rational.pmc:
Fix free vs Parrot_str_free usage in Rational PMC. mpq_get_str doesn't allocate string using Parrot's mem_sys_allocate, so don't try to free is with mem_sys_free.
nopaste "bacek" at 122.110.71.39 pasted "Ho-ho-ho. New Year gift. Test results on boehm_gc branch" (11 lines) at nopaste.snit.ch/19181 10:46
10:52 payload joined 11:05 bacek_ joined
dalek rrot: r43308 | bacek++ | branches/boehm_gc/src/gc/alloc_memory.c:
Use GC_MALLOC_UNCOLLECTABLE in mem_sys_alloc. We are manage this memory manually anyway.
 11:07
rrot: r43309 | bacek++ | branches/boehm_gc (2 files):
Add G option to t/harness to support GC switch
rrot: r43310 | fperrad++ | trunk/tools/install/smoke_languages.pl:
[install] add language Forth
 11:40
11:53 bluescreen joined 11:56 mikehh joined 11:58 ruoso joined 12:07 bacek joined 12:39 masak joined 13:00 kid51 joined
dalek rrot: r43311 | bacek++ | branches/boehm_gc/src (2 files):
Initialize Boehm GC early. We have to restructure CLI args handling to do it properly
 13:17
13:32 payload joined
hejki i'm trying to implement the stat(2) S_ISLNK-macro in parrot. i've added it in config/gen/platform/generic/stat.h heikki.mehtanen.fi/parrot_patch/patch1.txt and in include/parrot/stat.h heikki.mehtanen.fi/parrot_patch/patch2.txt . i get no compile-time errors and everything seems ok. calling 'stat filename, 4' doesn't return the S_ISREG-style behaviour, but it doesn't work as S_ISLNK is intended either. what am i missing here? 13:45
13:50 cognominal joined
masak Coke: ping 14:07
I've now bisected. 14:09
Rakudo stops building after r43175, which is a merge of context_unify3_simple
the next commit, r43176, which is a fix of the merge commit, either uncovers or introduces the error I'm seeing when building Rakudo with the latest Parrot. 14:10
tewk_ masak, r43176 just fixes pirc which you shouldn't be using, so r43175 must be the real culprit 14:19
masak tewk_: thanks. yes.
not surprising that it's a Parrot merge which breaks Rakudo.
hejki :) 14:26
14:26 mikehh joined 14:27 bluescreen joined
hejki a quick question: how-to grab ARGV in PIR? :) 14:28
14:32 patspam joined 14:35 whiteknight joined
whiteknight good morning #parrot 14:36
masak mornin' 14:37
hejki hiya
dalek rrot: r43312 | tewk++ | branches/pmc_freeze_cleanup/src/pmc_freeze.c:
Refactored goto into do while loop
 14:42
masak hejki: see docs/running.pod, line 298. 14:43
hejki that just defines that extra arguments are passed in ARGC 14:45
ARGV*
masak in what way doesn't that answer your question?
hejki .param string argv inside my :main subroutine captures the stuff i'm passing, but i guess it's an "array" or some sort?
it doesn't have example code for taking in params :) 14:46
hejki realizes that by the time passed he would've already typed in a PIR loop that does all four checks instead of making something that i'd use in conjunction with bash for
debugging just parrot (and not recompiling rakudo everytime) seems lotsa easier way of debugging the S_ISLNK macro 14:47
:P
"surprise! \\o/"
14:51 mikehh joined
hejki hmm.. www.parrotcode.org/docs/pmc/pmc/file.html this already has is_link-method O_o 15:11
wonder if i could utilize that instead of trying desperately figuring out parrot and adding the S_ISLNK :P 15:12
masak there's only one way to find out. 15:13
hejki ye
i'm already doing so :p
Coke hejki: .param pmc argv 15:23
hejki Coke: ye.. got it already by testing. thanks anyways :) 15:24
15:24 fperrad_ joined
dalek kudo/master: db84bc0 | masak++ | src/builtins/any-str.pir:
[src/builtins/any-str.pir] implemented :l
 15:33
Coke trac sllllow. 15:43
masak: what is your build error again? 15:44
(and are you passing any args to Configure.pl? or make?)
masak passing no args to neither. I have rakudo in parrot/languages/perl6. 15:45
build error here: gist.github.com/264688
Coke ... why do you have rakudo /in/ parrot? 15:48
(and not parrot /in/ rakudo) ?
CallSignature doesn't exist anymore. =-) 15:49
(that's my guess.(
yup. so you /can't/ update until rakudo is updated. =-)
masak Coke: the rakudo-in-parrot setup precedes the parrot-in-rakudo one. I just never made the switch.
it allows me to catch this kind of errors early. :) 15:50
Coke I'm fairly certain pmichaud is already aware of this.
masak goodie.
Coke but I also suspect master wouldn't get fixed.
masak I'd help in upgrading Rakudo, but I'm afraid I don't know enough about CallSignature and what's replacing it. 15:51
Coke does ng work against a recent parrot?
masak I don't know.
it has its own PARROT_REVISION, but I don't know what... I'll check.
Coke Might be useful to try that; if it does, then no worries. if not, then it might be worth opening a perl6 ticket.
I'm sorry I made you chase the bisect; I should have asked to see the error first. =-) 15:52
masak no problem.
Coke I had thought you and the other reporter were having similar errors.
my bad. :|
masak ng is at r43076 as well.
hejki we were
i'm just having the libfakeroot errors cause by archlinux build system
masak I haven't tried, but my guess is that it hasn't been adapted to the latest Parrot.
hejki (it builds stuff using fakeroot, and most likely that's why it's appended into my LD_PRELOAD before compilation takes place) 15:53
but those errors are just ignored (so they're not the real cause, but masaks stuff is)
15:55 jan joined 15:56 Psyche^ joined 16:02 mikehh joined 16:07 mikehh_ joined 16:11 ash_ joined 16:16 poet joined
poet I can't get parrot to compile on OS X 10.6.2. The error message is "i686-apple-darwin10-g++-4.2.1: c: No such file or directory". Can anyone help? 16:19
ash_ how are you compiling parrot?
I have OS X 10.6.2 and its working alright for me, i'll see if I can help 16:20
poet ash_: great thanks, much appreciated. I am just doing "perl Configure.pl" to generate the makefile and then make to compile
ash_ can you paste the output of the make op in like gist? 16:21
poet yea one second
this is with parrot 1.9.0 by the way, I can do a different version depending on what you have
ash_ i do svn trunk checkouts, currently revision 43076 16:22
poet ok ill try with that
ash_ which is a version of 1.9.0 so you know
poet ok
ok it's building now, ill post the output in a sec
ash_ k
poet ash_: here it is pastee.org/4rfzs 16:23
ash_ hmm, your using a newer version than 16:26
poet ah, just checked out the svn repo 16:27
ash_ i'll try that revision then, one sec
poet same error happens with 1.9.0 though, may be unrelated to version
Coke why do you have a \\c on the end of that command? 16:28
ash_ have you built it before?
poet im not sure 16:29
I have not built it successfully before
ash_ k, try doing a make realclean 16:30
poet ok, make realclean fails with perl -MExtUtils::Command -e rm_rf html 16:31
make -C src/dynpmc clean
Makefile:29: *** missing separator. Stop.
make: *** [dynpmc-clean] Error 2
ash_ hmmm thats not supposed to happen
poet i found the \\c in the makefile, should I delete it? 16:32
ash_ which version of perl do you have?
Coke the makefile is generated. we want to figure out how it got in there.
poet ahh this might be it
it's not using the system perl, it's using a perl from a porting system which is 5.8.9
Coke in config/gen/makefiles/root.in, I think this is the rule: 16:33
$(LIBPARROT_SHARED) : $(O_FILES)
we theoretically support 5.8.9, but if you have a more recent perl, yes, use that when running Configure.pl
(so, I'm guessing one of those variables is ending up with the wrong value.)
(not those variables, but the ones in the actual linker line.) 16:34
ash_ i have macport's perl 5.8.9 installed and it builds fine with that, but i normally use 5.10.1, you can try using a different perl to configure and see what happens
poet yeah it's building now with a makefile generated with 5.10.0 16:35
ack, same error
still have the \\c at the end
if I remove the \\c, the build still fails and I get the same error I got when I tried to make realclean, missing separator on line 29 16:36
16:38 plobsing_ joined
Coke that is a generated file. can you post it? 16:39
(the one with the missing separator)
poet yeah trying to figure out which one that is
Coke the \\c, I think you'll have to look in the generated top level makefile to see where the \\c is 16:40
poet yeah the \\c is in the top level makefile
make -C src/dynpmc
Makefile:29: *** missing separator. Stop.
make: *** [dynpmc.dummy] Error 2
does that mean the makefile in src/dynpmc?
Coke poet: yes.
poet pastee.org/zkmk4 16:41
Coke look, it's the \\c again.
poet yeah lol
Coke so somewhere, you're getting a "\\c\\n" inserted into your linkargs.
poet quite straneg 16:42
straneg
Coke LINKARGS := $(LDFLAGS) $(LD_LOAD_FLAGS) $(LIBPARROT) @icu_shared@ @libs@
(that's from dynpmc.in) 16:43
so, narrow down whether it's icu_shared or libs.
(you can just put ## ICU: @icu_shared@ as a comment in root.in or dynpmc.in, rerun configure.pl, and see if that's it. 16:44
(and if not, try @libs@)
the @foo@ syntax is used by configure to substitute things at config time into the makefile.
(and I'd do : ## ICU: @icu_shared@ ## 16:45
(so you can see where it ends.)
poet ok I will do that now, trying to find root.in
Coke config/gen/makefiles/*.in 16:46
root.in -> top level Makefile
poet gottacha
so add the string " ##ICU: @icu_shared@" anywhere? 16:47
ah gottcha 16:49
"## ICU: -lpthread -lm -L/opt/local/lib -licuuc -licudata -lpthread -lm \\c" in the generated make file
so how does the value of that variable get determined ? 16:51
ash_ poet: do you have icu installed? 16:54
17:00 mikehh joined 17:03 Zak joined 17:05 iblechbot joined
poet ash_: yeah I do 17:10
ash_ thats weird, i don't see where the \\c is coming from 17:11
do you have ack?
Coke auto/icu.pm
poet i dont have ack
Coke I would run that step with verbose output. 17:14
perl Configure.pl --verbose-step=auto:icu
er, auto::cu 17:15
er, auto::icu
note that configure.pl also has a --without-icu option to disable ICU entirely. 17:16
ash_ poet: what version of icu do you have? this is strange that I don't get this bug, mine's 4.3.1 17:17
17:27 ruoso joined 17:52 davidfetter joined 18:01 theory joined
mikehh nobody seems interested in #ps at the moment 18:51
19:18 theory_ joined
Tene bacek_at_work: pong 19:50
Coke Tene! 19:54
19:54 payload joined
Tene Coke: yes? 19:54
Coke you got me. I got nothing.
19:54 Zak joined
Tene Oh. :( 19:54
Not even a "pls maek partcl go!"? 19:55
Coke standing request.
also, feedback on the one_make branch is spiffy.
Tene oh, there's a branch?
Coke yah. i was going to do it piecemeal, but ripping stuff out is more fun. 19:56
Tene I'm make-clueless enough that I wouldn't be able to give any more feedback than whether it builds or not for me. Is that useful at this point, or no?
Coke especially useful if you're not using gmake. 19:57
I suspect there are a few missing deps at the moment as I try to detangle the recursive make.
Tene gmake ≟ "GNU Make 3.81"
Coke anyone using nmake, that's probably the killer.
Tene I don't have nmake. :( 19:58
Coke 'sfine.
so, no, don't bother. =-)
Tene Okay. 20:01
AFK now. I need to go show up WAY TOO LATE at work.
Coke mikehh: moving #ps over here. 20:02
patches welcome if you can figure out where I broke it, or just a paste of the build failure is fine, too. 20:03
20:05 joeri joined 20:51 bacek joined 21:21 s1n joined
ash_ has anyone made any progress on the OpenGL issue in OS X? just curious 21:55
22:12 poet joined 22:13 bacek joined 22:49 bacek joined 22:54 payload joined 23:10 ash__ joined 23:34 payload joined 23:39 plobsing joined 23:49 TiMBuS joined 23:56 poet joined

Powered by Cro and the Raku® Programming Language.

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

Thank you!