bluefeet Hey, nothingmuch over on perl.org#catalyst told me that someone in here would be able to point me at some code that generates colors for nicknames. 00:00
Anyone know what/where that code would be?
obra bluefeet: I think you want to see ian langworth's cpan directory 00:01
comet^ 'generates colors for nicknames' is an impossibly vague description. 00:05
xinming1983 Juerd: ??? what does serious mean? :-S
bluefeet obra: thanks! 00:06
00:10 bluefeet left
svnbot6 r8896 | rafl++ | r27913@ata: rafl | 2006-01-31 08:42:28 +0100 00:24
r8896 | rafl++ | util/build_pugs.pl:
r8896 | rafl++ | * only use directories as include or library dirs if they really exist.
r8896 | rafl++ | * clean up include and library dirs using canonpath.
r8897 | rafl++ | r27914@ata: rafl | 2006-01-31 08:42:41 +0100
r8897 | rafl++ | * Adjusted debian/ for 6.2.11.
Daveman :o 00:32
rafl Daveman: What's up? Did I break anything? :-) 00:37
Daveman rafl, nah ;) 00:55
rep ^_^ 00:56
01:16 rantanplan_ joined 01:19 justatheory joined 01:33 Cryptic_K joined 01:34 feng123 joined, Debolaz joined 01:52 stevan_ joined 02:40 sky joined 02:49 stevan joined 03:09 jp73137 joined
jp73137 hi, I'm searching for a list of surnames and Mr. Google is not helping me too much.... some clues??.... 03:21
03:26 _SamB_ joined
svnbot6 r8898 | audreyt++ | * undef.t and named_rule_capture.t: Probe for missing-rules-support 03:26
r8898 | audreyt++ | condition correctly. Reported by lwall++.
03:29 _SamB_ is now known as SamB
tewk Is there a perl5 version of rules floating around out there ? 04:01
04:04 jasho joined
jasho pugs binary for linux? 04:07
buu cpan: Perl6::Rules 04:22
er
04:22 stennie joined
buu search.cpan.org/~DCONWAY/Perl6-Rule...3/Rules.pm 04:22
jasho pugs binaries for linux? 04:33
rhesa jasho: which linux distro? 04:36
jasho suse or fedora. 04:39
04:43 putter joined, Amnesiac joined
putter tewk: not really. did you have a particular objective in mind? 04:44
04:44 PROO joined 04:47 PROO is now known as PRO, PRO is now known as proller 04:49 proller joined 04:51 dduncan joined
putter good night & 04:55
04:56 dduncan left 05:14 dduncan joined
dduncan hello hello 05:14
I was offline for the last 6 hours due to heavy winds bringing down my ISP's connections 05:15
but a smoke of r8895 was completed during that time, and I just uploaded it
this shows 13 failures of 11120, or 99.88% success 05:16
on darwin/haskell normal
those 13 failures were in t/pugsrun/09-dash-uppercase-c.t (6) plus t/rules/from_perl6_rules/capture.t (7) 05:20
I've pulled to 8898, which may fix some of those
05:25 jasho left
audreyt not afaics. I'll look at your smokes 05:32
what's the parrot version?
<- just woke up, going to find some lunch, bbiab
dduncan I am now pushing a fix 05:33
audreyt woot
dduncan essentially, there was a test that was supposed to bow out if rules support was missing (I haven't linked PGE)
but it only printed out 65 skips, and didn't exit too 05:34
audreyt ahh.
dduncan so the output was 65 skips plus 11 failures
svnbot6 r8899 | Darren_Duncan++ | r2285@darren-duncans-power-mac-g4: darrenduncan | 2006-01-31 21:32:59 -0800
r8899 | Darren_Duncan++ | t/rules/from_perl6_rules/capture.t : made the script exit when it determines rules support is missing, rather than running those tests anyway, in addition to skipping everything
dduncan now I'll check the other file
audreyt dduncan++
05:38 Jooon_ joined
dduncan audreyt, I see you're a late-to-bed-late-to-rise person, as am I 05:38
typically I go to bed at 1-3 am and get up at 11am-noon, in time for lunch 05:39
currently its 9:39pm here
as for t/pugsrun/09-dash-uppercase-c.t, I reran it directly and there were no errors 05:40
so either the errors only appear in smoke mode, or that was a transient problem which went away, or a recent checkin fixed it
I will now start another smoke, for the record 05:43
audreyt k
or you can use "prove" 05:44
dduncan what is that?
audreyt env HARNESS_PERL=./pugs prove t/pugsrun/09-dash-uppercase-c.t
will run one test using harness
very useful
see "perldoc prove" for details 05:45
dduncan and it will update the smoke.html in-place?
audreyt er, no, but at lesat you can simulate the harnessed confition
condition, even
dduncan okay, that's useful there for ruling out the problem
I also suggest the power brownouts I was having at the time may be related 05:46
they were during that last smoke
I don't have a UPS, just a surge protector
audreyt bbl 05:48
dduncan running env HARNESS_PERL=./pugs prove t/pugsrun/09-dash-uppercase-c.t shows all tests successful 05:51
05:52 weinig joined 05:53 gaal joined 05:55 justatheory joined, p5evalbot joined 05:56 stevan joined 05:57 justatheory joined, Bit-Man joined 06:00 justatheory joined 06:01 comet^ joined, theorbtwo joined, justatheory joined 06:04 justatheory joined 06:06 justatheory joined
dduncan now starting smoke of r8899 06:10
06:11 justatheory joined 06:14 justatheory joined 06:16 justatheory joined 06:19 sri_ joined 06:20 drbean joined 06:28 justatheory joined 06:34 justatheory joined 06:39 justatheory joined 06:40 justatheory joined 06:48 justatheory joined 06:49 beppu_ joined 06:55 justatheory joined 06:56 justatheory joined, iblechbot joined 06:57 justatheory joined 06:58 justatheory joined 07:03 justatheory joined 07:04 justatheory joined 07:06 justatheory joined 07:07 justatheory joined 07:09 nothingmuch joined, justatheory joined 07:10 beppu__ joined 07:11 BeoTW joined, justatheory joined 07:12 justatheory joined 07:16 justatheory joined 07:21 justatheory joined 07:23 justatheory joined 07:25 justatheory joined 07:27 justatheory joined 07:28 dduncan joined
dduncan hello 07:28
I shut down and disconnected my computer temporarily since there was a lightning strike so close to our house that we could feel the house shake like a small earthquake 07:29
07:29 justatheory joined
dduncan it came rather without warning, and there weren't more since; I decided to risk coming back on 07:29
said lightning didn't affect the power or anything, though 07:30
07:30 stevan joined
hachi heh 07:31
07:31 justatheory joined
hachi good to hear that... check to make sure trees aren't on fire in your yard 07:31
dduncan actually, I didn't go out side to check
but there is a lot of wind and rain, has been all day
normally, lightning comes in groups, and we can hear ones further away before they get closer 07:32
07:32 justatheory joined
dduncan maybe the wind drowned them out 07:32
anyway, my last smoke got through the ext/ and to builtins or so before I killed it for said shutdown; I have just restarted the smoke now 07:33
07:34 justatheory joined 07:38 justatheory joined 07:39 justatheory joined 07:40 justatheory joined 07:42 justatheory joined 07:43 justatheory joined 07:45 justatheory joined 07:46 justatheory joined 07:47 Kattana joined, r0nny joined 07:48 justatheory joined 07:49 justatheory joined 07:51 justatheory joined 07:53 _bernhard joined 07:54 justatheory joined 07:55 justatheory joined 07:57 justatheory joined
hachi damn, he's bouncing over here too 07:57
07:57 DesreveR joined 07:58 justatheory joined
mlh_ justatheory: you're bouncing quite a bit 07:59
08:00 justatheory joined 08:01 justatheory joined
hachi I banned him on #perl@irc.perl.org 08:01
sent him an email saying that I just banned his nick
08:02 justatheory joined 08:04 justatheory joined 08:05 justatheory joined
dduncan hachi, why would you ban justatheory? 08:08
hasn't that person been an untroublesome participant?
hachi his client is bouncing... you just ban the nickname, he can change his nick and rejoin when his client fixes itself 08:09
it's a motion to make the channel less cluttered, not to exclude them
08:12 justatheory joined 08:13 justatheory joined
lypanov agrees with hachi here 08:13
my
my logs are empty no
s/$/w/
lypanov gets coffee before typing any more
lypanov ignores justatheory joins parts
08:14 justatheory joined
dduncan so hachi, should we ban justatheory here too, for the reason you gave? 08:16
also, can anyone do it?
hachi sorry, no... I don't mean to advise you
08:18 justatheory joined 08:21 mtve joined 08:31 beppu joined
beppu ? 1 + 1 08:33
lypanov hachi, dduncan: seemingly stopped anyway. banned and i missed it or? 08:38
hachi quit on the last shot 08:39
never came back
dduncan I didn't do anything
lypanov neat 08:41
08:43 proller joined 08:50 DesreveR is now known as r0nny
dduncan my smoke of 8899 is about 2/3 done 08:56
09:09 kane_ joined 09:11 marmic joined 09:30 gugod joined 09:33 christo joined 09:38 nothingmuch joined
audreyt rehi 09:42
09:43 orafu joined
comet^ yells at the Perl6 design team. 09:43
dduncan smoke's almost done
comet^ (I've been asked to.)
www.nntp.perl.org/group/perl.perl6....uage/24584 <---> dev.perl.org/perl6/doc/design/syn/S03.html 09:47
dduncan done, and uploaded 09:51
audreyt comet^: submit a S03.pod patch? 09:52
dduncan I'm pleased to say that my smoke of r8899 on haskell / darwin passed all tests and has zero failures
audreyt comet^: or rather, /demand/ one :)
dduncan: excellent
dduncan its now almost 2am here
will there be more release candidates, or are you going straight to release?
and do you have any idea when that will be? 09:53
audreyt I'm investigating parrot embedding segfaults
but it's not really showstopping
I'm happy to call the tree -pre1 at this point
I'll ask around for more testers
meanwhile writing release notes and chglog
and then push the final release out somewhere in the next 6 hours 09:54
dduncan sounds good ... should be done when I get up tomorrow, then
audreyt nod
leo_ audreyt: is by any chance interpreter->lo_var_ptr not set properly?
dduncan I have a question about ChangeLogs
audreyt leo_: I have no idea what it is
so that's possible
dduncan I notice that some ext/ have their own ChangeLog files, and others record their own changes in the Pugs master changelog 09:55
currently I am doing the latter
leo_ are you entering parrot runloop multiple times from different C stack depths?
audreyt dduncan: I'd rather each ext/ report the major changes
dduncan do you have any preference for what all/none of the ext/ do, or ...
audreyt dduncan: in the main chglog, and put more details in their own chglog if desired
dduncan: essentially what perldelta does
(as a high level, one-paragraph-per-module summary) 09:56
leo_: very likely
leo_: in fact, almost guaranteed
leo_ then you have very likely a GC bug with embedding currently
dduncan so if a module's delta can fit in a paragraph, it goes in the master pugs changelog, and if its longer, it goes in a module's own?
audreyt and the workaround is to disable GC?
leo_ you could test by turning off DOD & GC
audreyt dduncan: yeah, and leave a sentence or two in the main chglog 09:57
leo_: yow... ok, will do
leo_ Parrot_DOD_block(interp) ; Parrot_GC_block(interp);
if yes, a change in the init sequence will very likely help
audreyt k
dduncan I think the details I put in the master changelog for my ext/ are short enough, so they should stay there 09:58
audreyt dduncan: they look fine, indeed
dduncan: thanks a lot for the various releng help... it's wonderful to have you around here :)
09:58 hugues23 joined
dduncan fyi, as-of Pugs r6.2.11 (and Rosetta r0.720.0), there will be synchronized initial perl 5 and 6 versions of the Rosetta (rewritten) framework on cpan 10:00
starting there, with the versions in sync, I will be including all the change details with the perl 5 version, and only a brief summary with pugs
since the changes to each will largely be the same, the details with the perl 5 version would apply to the perl 6 one too 10:01
audreyt sounds reasonable
dduncan likewise with Locale-KeyedText, which is not part of Rosetta, but just used by it
see you tomorrow 10:02
good night
audreyt leo_: amazingly, the segfault it is
#1 0x292d20c1 in parrot_split_path_ext (interpreter=0x87be700, in=0x29705090, wo_ext=0xbfbfc548, ext=0xbfbfc544) at src/library.c:394
pos_sl = CHARSET_RINDEX(interpreter, in, slash1, len);
leo_ looks like 'in' is NULL then 10:03
dduncan p.s. I see another smoker has a rather crippled box ... taking 1605 minutes to smoke 10:04
over 20 hours
leo_ bah - not NULL, just borked
audreyt leo_: nod, it's loading a null bytecode
leo_: hmm?
leo_ in=0x29705090
that doesn't look like a valid STRING address
audreyt it's a normal char* 10:05
load_bytecode now takes STRING?
since when?
leo_ audreyt: since 2 months or so
audreyt *sigh*
ok, that's the problem
10:05 G2 joined
dduncan on the other hand, if a box started the smoke, then hibernated, and then woke and finished, the timer may be dumb enough to just subtract the start from end time to get the result, which could be what happened here 10:06
audreyt changes CString to ParrotString
yay, another segfault 10:10
0x2928bea0 in real_exception (interpreter=0x87be700, ret_addr=0x0, exitcode=14, format=0x294209cc "Couldn't find file '%Ss'") at src/exceptions.c:691
691 the_exception->severity = EXCEPT_error;
hm, does it mean I have to call parrot_init_library_paths nowadays? 10:12
leo_ a sec
audreyt shouldn't Parrot_init go there anyway? 10:13
leo_ Parrot_set_config_hash();
interp = Parrot_new(NULL);
audreyt ok
10:14 Limbic_Region joined
audreyt leo_: can you look at this sequence and see if there's anything missing? 10:14
parrot_set_config_hash 10:15
parrot_new nullPtr
parrot_set_run_core interp PARROT_CGOTO_CORE
parrot_imcc_init interp
parrot_compreg interp pugsSTRING pugsCALLBACK
parrot_load_bytecode interp (toSTRING "PGE.pbc") 10:16
parrot_load_bytecode interp (toSTRING "PGE/Hs.pir")
...start calling pge...
leo_ looks ok so far 10:17
audreyt ponders
leo_ are you always running the whole sequence?
audreyt no, just at init
and also there's this segment after compreg that I forgot what for 10:18
pf <- parrot_packfile_new interp 0
pf_dir <- get_pf_directory pf
seg <- withCString "pugs" $ \p -> do
parrot_packfile_segment_new_seg interp pf_dir 4 p 1
set_pf_cur_cs pf seg
parrot_loadbc interp pf
(that's it)
leo_ if you have GC related segfaults, you should use Parrot_init_stacktop instead of Parrot_init 10:19
audreyt dist/build/libHSPugs-6.2.11.a(Parrot.o)(.text+0x1f45): In function `s5Fy_slow':
: undefined reference to `Parrot_set_config_hash'
leo_ the former needs the address of a local var though 10:20
audreyt hm, I need to explicitly link in _config.o, right?
leo_ yep 10:21
audreyt is parrot_config.o installed? 10:22
leo_ for installing you get another config 10:23
audreyt what's its name?
leo_ src/install_config.o
which has the correct --prefix
10:24 wilx`` joined
leo_ make src/install_config.o # part of make install 10:24
audreyt but it's not installed
just part of src/
right?
leo_ no has to be linked into the $exe
audreyt somehow it doesn't sound right 10:25
but ok 10:26
leo_ the idea is that any parrot is able to find it's libs, therefore prefix is compiled in via the _config.o 10:27
well, _config.o has all the Config.pm inside 10:28
audreyt does a parrot-config remember its original src tree? 10:30
the location that is 10:31
leo_ if there is a config var for that yes
audreyt aha, build_dir.
leo_ yep 10:32
audreyt ok... let's hope this works
leo_ see also t/pmc/config.t how to extract items from the builtin config via PASM/PIR 10:33
audreyt ok 10:35
new segfault
0x2929787f in run_sub (interpreter=0x87c1700, sub_pmc=0x8a37a50) at src/packfile.c:267
267 CONTEXT(interpreter->ctx)->constants =
leo_ that's from the packfile init sequence you pasted - missing const_table I presume 10:37
Limbic_Region audreyt - I assume you still read Dan's journal but if you don't get a regularly feed there have been some interesting posts this month you may interesting in. You seem to be interested in weird crazy blow your mind stuff 10:39
leo_ PF_create_default_segs should do it
10:41 nothingmuch joined
audreyt new_cs = PF_create_default_segs(interp, name, 0); 10:46
old_cs = Parrot_switch_to_cs(interp, new_cs, 0);
like this?
leo_ PF_create_default_segs(interp, "dummy", 1); // add to directory 10:47
I'm not sure if you need the switch_to_cs, but doesn't harm I think
another solution of that PF issue could be: 10:50
pf = Parrot_readbc(interpreter, "temp.pbc");
Parrot_loadbc(interpreter, pf);
where that load PBC is the PGE loading code
i.e. your pugs_init.pir
audreyt k 11:06
bbiab 11:08
Limbic_Region gets to the mudslinging post by Dan and says ignore that one while reading the others :-( 11:09
lypanov was kind of bored by all the posts... 11:17
nothingmuch leo_: vtable question 11:21
does parrot analyze to see if there are any subsequently non virtual methods for a given object?
thus doing away with the vtable indirection if there's no possibility of method ... going to any other method? 11:22
11:25 scook0 joined 11:34 pdcawley joined
audreyt leo_: ping 11:50
nothingmuch audreyt: www.woobling.org:3000/ 11:51
audreyt cool 11:54
leo_: I'm not sure where to put PF_create_default_segs 12:11
audreyt gives parrot embedding another try 12:20
nothingmuch audreyt: it also likes unicode 12:25
nothingmuch likes it when things like unicode with little effort
audreyt yay, it worked 12:26
12:26 broquaint joined
audreyt "if you see this, parrot embedding has lived again" 12:26
nothingmuch (the chat thing)
audreyt++
audreyt ok, no showstoppers anymore 12:27
audreyt rolls preflight #1
do we have a Win32 user here?
cygwin and/or normal win32
scook0 audreyt: not sure if you've noticed, but someone posted a patch to p6c -- have you sent a ci invitation? 12:30
svnbot6 r8900 | audreyt++ | * embedded parrot lives again, thanks to leo++'s hints.
audreyt scook0: no, would you do it? :) 12:31
(though in that particular case it's already fixed)
scook0 audreyt: do I need meta-ci? (I don't have it)
audreyt you now do 12:32
but I'll send out the invitation anyway :) 12:34
scook0 audreyt: I was just about to, but I notice it's already done (I had to reset my OpenFoundry pw...) 12:37
audreyt :D
pre1: pugscode.org/Perl6-Pugs-6.2.11-pre1.tar.gz | pugs.blogs.com | pugscode.org | pugs.kwiki.org | paste: paste.lisp.org/new/perl6 or sial.org/pbot/perl6 12:39
audreyt preflight #1 calls for test :) 12:40
azuroth I'm quite behind in times, but just wondering... what's the `./method` syntax this week?
audreyt azuroth: self.method
azuroth cool
would that be self.:method if method were private..? 12:41
audreyt self!method
12:49 pdcawley joined 12:59 kolibrie joined 13:03 chris2 joined
leo_ audreyt: pong 13:08
audreyt leo_: I think I've fixed it
it's a quite absurd sequence 13:09
leo_ great audreyt++
audreyt had to call loadbc twice
parrot_imcc_init interp
pf <- parrot_packfile_new interp 0
parrot_loadbc interp pf
seg <- withCString "pugs" $ \p -> do
parrot_pf_create_default_segs interp p 1
set_pf_cur_cs pf seg
parrot_loadbc interp pf
and _then_ compreg
and all worked
still not sure why two loadbcs are required 13:10
leo_ looks good
audreyt but I cargo culted it :)
Limbic_Region sticking my nose in a conversation it doesn't below with no idea what is being discussed but is the loadbc being called twice an unnecessary performance hit? 13:32
s/below/belong/
13:34 kanru joined
leo_ Limbic_Region: it's an one-liner and only in the init sequence 13:34
Limbic_Region thanks leo 13:36
audreyt another segv 13:39
0x293f6d0c in Parrot_ResizablePMCArray_shift_string (interpreter=0x87c2300, pmc=0x8835210) at resizablepmcarray.pmc:176
13:39 elmex joined
audreyt hm, seems to be PGE specific 13:39
leo_ audreyt: with GC enabled? 13:40
audreyt leo_: yeah
leo_: what was the init sequence fix you mentioned that may work around GC bugs?
leo_ use Parrot_init_stacktop() instead of Parrot_init() 13:41
13:41 putter joined
putter audreyt: fyi, INSTALL and Makefile.PL both currently say "emberdded parrot is NOT recommended with this release"... something to strip if appropriate. 13:42
audreyt putter: please strip, then
it's now working happily 13:43
putter will do
audreyt (well, modulo a couple test failures, but I'm fixing them)
danke
leo_ audreyt: the void *arg has to be the address of a local var on the stack 13:44
audreyt leo_: ok... Haskell has no stack btw
but I'll see what I can do with stable pointers
leo_ yeah - I knew that this isn't easy :-) 13:45
putter in 13:46
svnbot6 r8901 | putter++ | INSTALL, Makefile.PL: removed disrecommendation of embedded parrot.
audreyt cool 13:48
I'll reboot to win32 and do a smoke there
putter is this the right thing for pre1 make?: Current version is 0 13:56
audreyt huh? 13:58
14:00 nothingmuch joined
svnbot6 r8902 | audreyt++ | * add a parrot_init() call. 14:00
r8902 | audreyt++ | * switch to -f (fast core) for portability.
putter /usr/bin/perl -Iinc util/version_h.pl src/Pugs/pugs_version.h 14:01
Current version is 0