www.parrot.org/ | Parrot 1.4.0 "Mundo Cani" Released!
Set by moderator on 21 July 2009.
dalek rrot: r40320 | chromatic++ | trunk/compilers/imcc (2 files):
[IMCC] Plugged a memory leak I introduced in r40270 when removing quoted names
 00:03
00:11 wayland76 joined
dalek cnum-dynpmcs: r126 | darbelo++ | trunk/ (4 files):
Remove decTest's Configure.pl, as the decTest Makefile is now generated by the

The bulding and cleaning of decTest are now controlled by the main Makefile.		 00:57
darbelo cotto: ping 00:59
cotto darbelo, pong 01:02
darbelo I'm thinking that we should provide a human-friendly representation of the integers used in set_exceptions(). 01:04
Is there a good place to put that in? install-wise.
cotto Hmm. Lemme dig a bit. 01:05
darbelo The sanest place I can think of is "lib/1.4.0-devel/include/" but I don't know if third parties are supposed to put their crap all over the parrot install tree. 01:08
wayland76 darbelo: That seems to be the Rakudo approach 01:09
I suggest that Parrot come up with a policy on this quick-smart (ie. about what directory names to use, and that sort of thing), before it devolves into a big mess :)
cotto I think the right place to install them would be lib/1.4.0-devel/library/ or include 01:11
I'm not certain though. 01:12
When in doubt, steal from Rakudo or ParTcl.
darbelo cotto: Also, seeing that the user is .inlcuding our constants should we try to be helpful and loadlib ourselves in that same file? 01:14
cotto I can't think of a case against that.
Sure. 01:15
darbelo Cool, and now that we're going that way I can replace the strings for the rounding-mode with INTVAL .macro_consts. 01:17
01:22 Whiteknight joined
Whiteknight is holding his copy of the PIR book in his hands right now 01:23
GeJ Whiteknight: hi, 01:28
Whiteknight: where did you get it?
Whiteknight I think amazon 01:31
hello GeJ
The book is actually a little smaller then I was imagining, although much of it was cut out 01:36
GeJ Ah, found it.
dalek rrot: r40321 | dukeleto++ | trunk (2 files):
Fix TT#884 and add a test for setting a breakpoint in the parrot_debugger
 01:56
cnum-dynpmcs: r127 | darbelo++ | trunk/ (5 files):
Convert the rounding_mode get/set METHODs to use INTVALs instead of strings.

inc/decnum.pasm Add logic to the Makefile to install and uninstall this new file.		 02:14
02:25 eternaleye_ joined
dalek rrot: r40322 | jkeenan++ | trunk/t/codingstd/c_function_docs.t:
A file lacking documentation of all its C functions had its name changed. Updating the list of such files, in effect TODO-ing that file from this test.
 02:27
02:27 kid51 joined 02:33 tetragon joined 02:41 janus joined
pmichaud 01:08 <darbelo> The sanest place I can think of is "lib/1.4.0-devel/include/" but I don't know if third parties are supposed to put their crap all over the parrot install tree. 03:18
01:09 <wayland76> darbelo: That seems to be the Rakudo approach
In the ins2 branch, Rakudo is (as far as I know) following the Parrot guidelines as to where things are to be installed. If this is not the case, I will appreciate any corrections.
afk for a bit 03:19
03:22 s1n joined 03:31 dukeleto joined
dalek TT #886 created by jimmy++: [patch]various consting, memory leak fixes and so on. 03:43
03:55 Zak joined
cotto seen jimmy 04:01
purl jimmy was last seen on #parrot 13 hours, 28 minutes and 26 seconds ago, saying: thank Whitenight for creating that ticket.
04:02 mokurai1 joined 04:04 jimmy joined
jimmy parrot make error 04:07
.\\miniparrot.exe config_lib.pasm > runtime\\parrot\\include\\config.fpmc 04:08
Malformed string
make: *** [runtime\\parrot\\include\\config.fpmc] Error 1
It happend on windows 04:09
cotto Did you make realclean? 04:13
jimmy yes
make realclean && perl Configure.pl && make 04:14
cotto Hmm. trying to duplicate...
is it easy to reproduce on your system? 04:15
jimmy it errors very time on my windows xp 04:16
cotto ok. It works fine for me. 04:17
particle or jonathan, ping
Util jimmy: what is the size of the config.fpmc file after the error? 04:18
jimmy 5,132
cotto Util, are you on win32 04:20
?
nopaste "jimmy" at 119.123.3.190 pasted "config.h" (187 lines) at nopaste.snit.ch/17422 04:21
Util cotto: OS X laptop, Win32 desktop, but my Mingw32 compiler will take a few minutes to set back up. 04:22
jimmy sorry
Util: it's 0 size
F:\\pipp\\parrot-dev>.\\miniparrot.exe config_lib.pasm 04:23
Malformed string
04:30 eternaleye joined 04:39 iblechbot joined
dalek rrot: r40323 | dukeleto++ | trunk/t/tools/parrot_debugger.t:
Add tests for listing breakpoints in parrot_debugger
 04:40
04:42 Zak joined
dalek TT #887 created by jimmy++: [patch]removed unused macro args 04:48
jimmy dalek: where's TT #886? 04:49
cotto jimmy, dalek is a bot 04:50
trac.parrot.org/parrot/attachment/ticket/886 04:51
jimmy I know, but it lost TT #886
bacek_at_work <dalek> TT #886 created by jimmy++: [patch]various consting, memory leak fixes and so on.
<-- eternaleye_ has quit (Ping timeout: 360 seconds)
jimmy ;( 04:52
cotto You can just look at another ticket and change the url 04:53
nm. You thought it didn't make the announcement. 04:54
Util better url: trac.parrot.org/parrot/ticket/886
jimmy I found that parrot used 'Parrot_PMC' or 'PMC *' in some place. but they are the same 04:55
and STRING or Parrot_String 04:56
Util Oh, I see what you mean about dalek missing #886. Dalek sometimes drops tickets and commits on the floor; I have not seen a mention of the particular cause.
jimmy Util: yes 04:57
Util Wait, no, Dalek *did* announce 886. I see it in the scrollback. 04:58
Also in the logs: irclog.perlgeek.de/parrot/2009-07-29#i_1352219 04:59
Util is at 1 minute to midnight local. Fading...
jimmy: what compiler are you using? 05:00
jimmy strawberry perl
purl it has been said that strawberry perl is found at strawberryperl.com/
Util I just tried it with MSVC 13, and it worked fine. Trying to set up Mingw32 (which should come with strawberry perl) now. 05:02
05:05 Zak joined
jimmy where is Parrot_String defined? I found some place used it, but i can't find where it is defined. 05:09
seems that it had been removed from interpreter.h 05:11
05:15 theory joined 05:16 dukeleto joined
Util jimmy: try include/parrot/core_types.h 05:20
jimmy: I have just compiled with Mingw-TDM 4.4.0, which should not differ significantly from the compiler in Strawberry Perl. I cannot reproduce your problem. 05:27
My theory is that you have a file that is not getting removed by `make realclean`, probably due to a Win32-specific path.
Could you try compiling a fresh checkout in a different directory? 05:28
If it works, then compare the directory listing of a fresh checkout to the listing of your original (error-producing) `make realclean` directory, and report the difference to us so we can get realclean to clean up the file.
05:30 petdance joined
Util petdance: Is there a mailing list or other venue to discuss the Coverity errors? 05:32
petdance just the parrot list 05:33
Util ok, thanks
petdance dd you have a question? 05:34
05:40 eternaleye joined
Util petdance: all the RETURN_LOCAL defects are coming from the same use of _return_indexes in NCI functions of PMCs. However, I cannot distinguish what is different (and bad) in those particular NCI functions, vs the NCIs that it finds defect-free. 05:40
petdance yup, I don't know myself.
It's on my todo
Util I did not want to knock myself out over something already being tackled. 05:41
petdance If I did, I'd have fixed 'em. :-)
Util Well, such fixes might have been held up by pending NCI design change, or something :) 05:42
petdance Dunno, I don't even know what an NCI is. :-)
I'm jsut a C guy.
Util If I start any intense work on the issue, I will make a TT to co-ordinate and document any progress or false avenues. 05:43
NCI is Native Call Interface, also called FFI/Foreign Function Interface. In Perl 5, it is handled by XS or Win32::API. 05:44
Util sleeps; will check the scrollback in the morning for any updates from jimmy++ on compiling with a fresh checkout. 05:46
mikehh All Tests PASS (pre/post-config, smolder, nqp_test, fulltest) at r40323 - Ubuntu 9.04 amd64 05:55
06:10 uniejo joined 06:21 eternaleye joined
mikehh rakudo (6999e58) builds on parrot r 40323 - make test PASS, make spectest (up to 27790) FAIL - Ubuntu 9.04 amd64 06:35
2 failures - t/spec/S12-enums/basic.rakudo - Parse errors: Bad plan. You planned 30 tests but ran 28 - fails with Method 'Num' not found for invocant of class '' after ok 28 06:37
06:38 eternaleye joined
mikehh t/spec/S12-introspection/walk.t - Parse errors: Bad plan. You planned 10 tests but ran 4. - fails with Parameter type check failed; expected Any, but got Object for $class in call to block_86 after ok 4 06:39
06:40 TiMBuS joined
mikehh just did a security update - got to reboot 06:44
06:54 mj41 joined 06:59 mikehh joined
dalek rrot: r40324 | dukeleto++ | trunk/t/tools/parrot_debugger.t:
Add tests for tracing and printing registers in parrot_debugger
 07:08
07:11 abesapien joined 07:21 flh joined 07:22 eternaleye joined 07:25 chromatic joined 07:27 knewt joined 07:57 mokurai1 left 08:03 abesapien joined 08:45 masak joined 08:49 desertm4x joined 09:08 szabgab joined
dalek ose: r85 | Austin++ | trunk/src/parser/ (3 files):
Updated PASSTHRU to call DUMP
 09:20
09:23 abesapien joined 09:41 wayland76 joined 09:49 jimmy joined
jimmy Util: It worked before. 09:50
10:04 donaldh joined 10:23 einstein joined 10:38 GeJ joined 11:19 he joined 11:20 donaldh joined 11:39 rblackwe joined 11:42 bacek joined 11:54 whiteknight joined 12:08 payload joined 12:13 jimmy joined
flh I have a dynpmc "foo" which extends another dynpmc "bar", but "foo" doesn't inherit bar's "mark" vtable function (default's mark function is called instead): has anyone ever seen this bug before? 12:26
bacek flh: It's not a "bug" 12:30
wait.
flh a feature maybe? :)
bacek it is
jimmy I don't know why make failes with chinese TEMPPATH 'C:\\DOCUME~1\\ę•£čŠ±ē‰§äŗŗ\\LOCALS~1\\Temp', It didn't failed before with chinse TEMPPTATH 12:32
flh oh, sorry, my mistake this time :)
bacek flh: can you nopaste generated foo.c? 12:33
flh I was calling VTABLE_mark on PMCs which didn't have the custom_mark flag
jimmy It happened in Parrot_str_unescape() fuction 12:34
bacek flh: You shouldn't call VTABLE_mark directly. It's not "public API" really.
Coke yay, trac spam.
bacek Coke: 840? 12:35
flh bacek, I should replace the call to VTABLE_mark with Parrot_gc_mark_PObj_alive, am I right?
bacek flh: erm... May be. It depends on context. Usually you should call Parrot_..._alive from VTABLE_mark _implementation_ 12:36
flh FixedPMCArray uses Parrot_gc_mark_PObj_alive, so since I'm implementing something which looks like an array, it should be ok :) 12:37
bacek in foo.mark it's probably better to call SUPER() to invoke bar.mark 12:38
12:38 ruoso joined
bacek flh: yeah. Rule of thumb: in Foo.mark you should call Parrot_..._alive for each ATTRibute 12:38
(except non-GCable ATTRibutes) 12:39
dalek tracwiki: v8 | coke++ | CageTasklist
tracwiki: trac.parrot.org/parrot/wiki/CageTa...ction=diff
tracwiki: v1 | coke++ | TracSpammers 12:42
tracwiki: trac.parrot.org/parrot/wiki/TracSp...ction=diff
tracwiki: v2 | coke++ | TracSpammers 12:46
tracwiki: trac.parrot.org/parrot/wiki/TracSp...ction=diff
12:46 he left
Coke bacek: why is it wrong to expect that you get your parent's mark?? 12:47
bacek: yes. cleaned up. 12:48
bacek Coke: it is not. My mistake.
Coke finds where 16G of his hard drive is. 12:49
cono not labeled? :D
12:51 tetragon_ joined
Coke family is not emptying their Trash. 12:53
mikehh All tests PASS at r40323 (pre/post-config, smolder, nqp_test, fulltest) - Ubuntu 9.04 amd64 12:54
Coke wonders if posting email addresses of known spammers is a net karma win. 12:56
Coke notes with some irony that the spammer in this case is taking advantage of gmail's lovely anti-spam.
cono I guess google ban him soon 12:57
flh is there a way to force a GC run from PIR? I'd like to write a testfile to check whether my dynpmcs correctly mark their data 12:59
ok, I've asked my question too quickly, was looking for gc in src/ops, and missed the collect opcode 13:01
13:04 bsdz joined, bsdz left
bacek flh: "collect" 13:05
"sweep 1"
iirc
dalek kudo: 3910fce | jnthn++ | src/builtins/guts.pir:
Fix bug in role attribute composition introduced with fix for attribute ordering. Fixes RT#67992.
 13:06
flh collect and sweep 1 are synonymous?
bacek flh: not really. But "sweep 1" will run "collect" unconditionally 13:07
13:08 zak_ joined
bacek seen kj 13:09
purl kj was last seen on #parrot 47 days, 19 hours, 3 minutes and 3 seconds ago, saying: (forgiving, or maybe just keeping silent) [Jun 11 18:00:25 2009]
bacek seen kjs
purl kjs was last seen on #perl 1 years, 35 days, 21 hours, 30 minutes and 7 seconds ago, saying: yo [Jun 23 15:33:30 2008]
dalek TT #888 created by jimmy++: [BUG]parrot make failed with chinese TEMP PTAH 13:21
13:25 HG` joined
mikehh rakudo (3910fce) builds on parrot r40323 - make test PASS/make spectest (up to r27794) FAIL - Ubuntu 9.04 amd64 13:26
2 failures - t/spec/S12-enums/basic.rakudo and t/spec/S12-introspection/walk.t 13:27
t/spec/S12-enums/basic.rakudo - Parse errors: Bad plan. You planned 30 tests but ran 28 - Method 'Num' not found for invocant of class '' 13:30
13:30 hercynium joined
mikehh t/spec/S12-introspection/walk.t - Parse errors: Bad plan. You planned 10 tests but ran 4. - Parameter type check failed; expected Any, but got Object for $class in call to block_86 (+ traceback) 13:32
jimmy perl6? 13:33
purl i heard perl6 was see Perl 6
jimmy #perl6?
purl #perl6 is at irc.freenode.net.
mikehh jimmy: probably - but am testing latest parrot
masak purl: jimmy? 13:34
purl jimmy is translating some docs to chinese. or allegedly a sloth
masak :) 13:35
moritz sloth?
purl sloth is, like, being a sloth. or a kind of animal, now extinct
jimmy parrot support chinese path more worse now. :(
einstein I succeeded in removing the unionval structure from pobj.h see ticket #549 13:40
dalek kudo: 6ff39ce | jnthn++ | src/classes/ClassHOW.pir:
Implement :tree in .^parents; fix Object.^parents which for some crazy reason I once thought should return something other than an empty list.
 14:01
NotFound There is some reason to not delete src/gc/res_lea.c ? 14:21
"As of May 14, 2007, it doesn't work"
14:25 kj joined
Coke NotFound: there's a ticket to remove it, no? 14:27
NotFound Don't know, I'll take a look.
Coke ISTR kid51 removed the options to /build/ them. perhaps the c files were accidentally left behind. 14:28
trac.parrot.org/parrot/ticket/490 14:30
looks like the commit for that just removed the ability to /pick/ the broken GC cores.
I think removing the broken GC cores itself is covered. 14:31
(but not done yet.)
NotFound There are still several files that use or document res_lea. 14:33
Coke if it was one of the options that was removed for TT #490, the intent was to remove the options entirely, not just the ability to specify them. 14:36
bacek trying to remember exact form of "Better to ask for..." and launching "git svn dcommit" 14:37
*incoming*
Coke bacek: "Better to ask for forgiveness than permission"
bacek yay.
Who can give permissions to commit into /languages? 14:38
dalek kudo: b6f792c | jnthn++ | src/classes/ClassHOW.pir:
Implement :tree option for attribute introspection.
bacek Coke: who can grant commit permission to /languages? 14:43
anyway, sleep time. 14:47
bacek must sleep
purl $bacek->sleep(8 * 3600);
Coke any of the trac admins, I think.
I just don't know how.
bacek Coke: ok. I'll ask tomo^W today's morning
14:55 MoC joined 14:58 davidfetter joined
dalek kudo: c4f0a93 | jnthn++ | src/classes/ClassHOW.pir:
[t/spec] Support :tree option for .^roles.
 15:10
15:20 donaldh joined 15:31 solarion joined
dalek kudo: a48b31a | jnthn++ | src/ (3 files):
Implement .rw and .readonly for attribute introspection.
 15:34
Coke Reminder for p6 folks: TPF Q3 call for grants ending friday. 15:44
msg Andy buzzable? - news.perlfoundation.org/2009/07/200...oposa.html 15:45
purl Message for andy stored.
Andy tweeted 15:46
I wish Alberto made his announcements a little more interesting.
TimToady may you live in interesting announcements :P 15:47
Andy I wish I could. 15:48
Rule 1 of Perlbuzz articles: Someone has to have written it from the point of view of "What will make someone wantto read this?"
PerlJam Andy: MONEY! MONEY! MONEY! TPF has money to give and *YOU* could be the recipient! 15:50
Andy That would be a start.
As I am fond of saying these days, "Your awesomeness is not self-evident."
davidfetter www.youtube.com/watch?v=3_iQZiVD_zA 15:52
15:52 theory joined
PerlJam Andy: how about "Want to make up to $3000 working from home? Give TPF a call!" :) 15:52
Andy PerlJam: Don't tell me, tell Alberto. 15:53
PerlJam He's not going to lean that way in his announcements I don't think. Geeks are mostly just about the facts.
Andy PerlJam: And yet, that's not the case. 15:54
davidfetter what a silly conceit! there's no such thing as an unframed fact
Andy Because nobody is going to read an announcement of any kind without there being a compelling reason to do so.
davidfetter PerlJam, check out The Political Mind by George Lakoff for a fuller explanation 15:56
Andy For that matter, nobody's going to look at your project without a compelling reason to do so. 15:58
Nobody goes surfing through sourceforge reading READMEs at random.
NotFound "I'm the widow of the former president of The Perl Foundation and I have a proposal..."
davidfetter that "just the facts" framing has damaged a lot of geeks' lives :P
NotFound, excellent! :)
Andy NotFound: That scared me for a bit. 15:59
Bill Odom has just been replaced as pres of TPF, because he doesn't have time to do his duties
and I'm certain that a lot of it has to do with his wife's health.
So when you said that, I got scared, until I got the Nigerian spam joke.
NotFound Andy: sorry, I'm not following TPF news. 16:00
Andy I know, I'm just explaining.
Coke what position did Richard have? 16:08
Andy not sure
purl somebody said not sure was appliable, one more setjmp in a C function that already heavily depends on one, is not much more risk.
Andy purl, not sure is <reply>
purl ...but not sure is appliable, one more setjmp in a C function that already heavily depends on one, is not much more risk....
Andy no, purl, not sure is <reply>
purl okay, Andy.
Andy I think Chairman or something 16:09
Coke Andy: You'd be proud, I think I just converted someone from bitching about a CPAN module to submitting a grant to do something about it.
Andy fantastic 16:10
particle richard was president, and stepped down for the grant.
jim brandt took over as president.
bill odom was chairman
Coke ah: news.perlfoundation.org/mt/mt-searc...p;limit=20 16:11
(wow, the blog was helpful!)
blog++
... that sounded sarcastic. it's not, I'm happy!
dalek kudo: ce21ffe | pmichaud++ | docs/spectest-progress.csv:
spectest-progress.csv update: 423 files, 12048 passing, 8 failing

   S12-enums/basic.rakudo aborted 2 test(s)
   S12-introspection/walk.t aborted 6 test(s)
16:15 soxet joined
Coke seen merlyn? 16:24
purl merlyn was last seen on #moose 44 days, 14 hours, 36 minutes and 57 seconds ago, saying: one more thing to install though. :) [Jun 15 01:41:42 2009]
16:25 Psyche^ joined
NotFound What's "Parrot_Exec_OS_Comman" ? Just a typo? 16:32
config/gen/platform/ansi/exec.c 16:33
Coke finds out that perlmonk passwords have been hacked by following a monk node mentioned with no context in merlyn's fb status. 16:34
s/status/the comment notes he posts instead of using the status, but you knew what I meant/ 16:35
16:43 mokurai joined 16:50 hercynium joined 16:57 sekimura joined 17:00 whiteknight joined
dalek rrot: r40325 | NotFound++ | trunk (2 files):
[cage] add pod doc to example program cotorra
 17:07
17:18 desertm4x joined
Coke is inspired to buy the full 1Password app and start going through his online haunts. :| 17:28
einstein I made a patch to ticket #549, to remove the unionval struct from the pobj.h. 17:31
But I do not know whether i did everything correctly.
the build and tests all succeed with the patch 17:32
I do not know who is responsible for this ticket at the moment or who can give me a confirmation whether it is correct? 17:33
NotFound einstein: there is a recent message on that ticket 17:34
17:35 darbelo joined 17:36 clinton joined
NotFound Uh, is just on the mailing list, not in the ticket. 17:38
einstein oh ok thanks
Coke email2trac-- 17:43
einstein I have read that message, I'am going to take a closer look at the gc system 17:45
17:50 hercynium_ joined 18:07 chromatic joined
particle coke: see passwordsafe.sourceforge.net 18:33
damn, temps just reached triple digits here, before noon. 18:36
cotto Woohoo! 18:38
NotFound Oh, no, t/pmc/packfilefixupentry.t random failures again.
moritz be glad that you measure in Farenheit; 100Ā°C wouldn't be bearable :) 18:39
18:39 szabgab joined
Coke particle: 1password integrates very well with my iphone & my mac. 18:40
particle moritz: 38C isn't very bearable, either :)
NotFound Beer able?
cotto Any temperature is beerable. 18:41
18:42 Zak joined, japhb joined 18:46 davidfetter joined 18:59 whiteknight joined 19:10 allison joined 19:21 he joined, mokurai left 19:39 payload joined
darbelo cotto: ping 19:50
pmichaud it's cool here today -- 84degF 19:52
very atypical for DFW in late July
TimToady phone 20:00
pmichaud darn, beat me to it by 2 sec
cotto darbelo, pong 20:11
darbelo cotto: I was updating the decTest parser to the changes I commited to the rounding mode. 20:15
cotto ok 20:16
darbelo While there I noticed some of the case-insensitive matches don't need to be case-insensitive, which if I understood pmichaud correctly could speed up our parser a bit. 20:17
cotto That possibility exists.
darbelo Do you have any timings from last time so we can compare the performance? 20:18
20:20 ruoso joined
cotto Are your changes committed? 20:20
also, I get an error when I run Configure.pl 20:21
value for 'win32_libparrot_copy' in aux/decTest/cfg/Makefile.in is undef at /usr/local/parrot-svn/lib/1.4.0-devel/tools/dev/../lib/Parrot/Configure/Compiler.pm line 547, <$in> line 42.
darbelo not yet. I'm chasing a bug in the changes I made.
cotto ok.
darbelo I don't think I'm getting that error here. 20:22
Wait, no. I seem to be getting that error. I just didn't notice it before. 20:23
But it doesn't affect anything I can see. 20:24
Does 'make' work for you?
cotto yup. I just had to update the installed Parrot. 20:26
probably for the SUPER changes
20:28 szabgab joined
cotto about 11.96 seconds (avg of 6 runs) to parse multiply.decTest with --target=parse 20:32
darbelo while we're at it. Can you try add.decTest? That one was failing for me before. 20:33
20:33 flh joined
cotto super slow 20:36
darbelo Biggest test file we have. On my box it is EPIC slow, and the dies with "out of memory".
cotto about 2:12:70 20:38
to parse
and generates a 2.5M of output
darbelo That's a lot of output. I'm guessing you ran it with --target=parse, right? 20:40
20:41 dukeleto joined 20:42 dukeleto_ joined
cotto about 12.47 (avg of 6 runs) to parse multiply.decTest after I stripped out :i from the grammar 20:43
yes
It appears to be slower.
nm. I goofed. 20:45
20:47 flh joined
japhb What is the state (or not) of the Parrot event loop and IO interaction? Are we able to add file handles to watch, and callbacks for them, from PIR? 20:50
cotto darbelo, it looks like there are some nice speedups to be had. 20:53
darbelo I turned some changes I made here and there into separate commits, the :i removal went into r130 if you want to test it. 20:55
the generated code is having failures, but I'll fix that in a separate commit.
cotto s/nice/major/ 20:56
jonathan japhb: If you meant async IO, last I knew it's NYI.
japhb: I think I remember Whiteknight having plans on it. 20:57
japhb jonathan: I was thinking about allison's comment that they want to have one new "standard library" a month implemented, and I was thinking of doing dbus, since that opens up a world of power on *nix systems. Unfortunately, it's an async message passing system, that needs a running mainloop somewhere, and the ability to manage socket callbacks, timeouts, etc. 20:58
jonathan japhb: dbus would be very cool to have (guy in my local pm group was playing with that and Perl recently and gave some talk on it...so unusually it's *nix and I know what it is ;-)) 21:00
japhb: But I agree, from what I know of current state of Parrot in those areas, may be tricky. 21:01
21:01 Whiteknight joined
jonathan I know there is some scheduler stuff in place. 21:01
Ah, here's Whiteknight. He'll have more clue than I. :-)
Whiteknight ???
I'm fresh out of clues
japhb jonathan: DBus is fully ported to Windows, IIRC, as it's required for several open source GUI applets to even function. But on a modern *nix, it's basically required to get much beyond booting. :-) 21:02
Whiteknight: check irclog since about 15 minutes ago.
cotto darbelo, about 6.3 (avg of 6 runs) to parse multiply.decTest with r130 and a few extra changes
I'd call that better. 21:03
Is this a good time to commit?
dalek cnum-dynpmcs: r129 | darbelo++ | trunk/aux/decTest/src/inc/procs.pir:
Add the 'apply' pseudo-operation the the decTest support lib.
Whiteknight irclogs?
purl irclogs is irclog.perlgeek.de/parrot/today or see also: infrared clogs
Whiteknight japhb: I've been planning to implement an AIO system for some time now 21:04
cotto dalek, you should make a test to ensure that all the test files you expect to parse can be parsed.
Whiteknight I dont now much about dbus, but I can look it up and take that into consideration
cotto add.decTest is down to about 47 seconds. 21:07
to parse
GeJ Good morning everyone. 21:08
japhb Whiteknight: Thank you. Do you have any time frame (or release goal) for the AIO stuff? 21:10
cotto Hi GeJ
darbelo, you there?
darbelo cotto: Just got back, was afk. commit what?
cotto some changes to speed up the parser a little further 21:11
nopaste "cotto" at 74.61.2.46 pasted "further parser optimizations for decTet" (54 lines) at nopaste.snit.ch/17431 21:12
darbelo I've already commited the speedups I made. dalek hasn't picked them up yet, but they're on goolge-code.
Oh, I missed those. Commit away! 21:14
cotto done
21:15 payload joined
cotto Parsing is still slow, but it's much more reasonable now. 21:15
dalek cnum-dynpmcs: r131 | christop...@mksig.org++ | trunk/aux/decTest/src/pct/grammar.pg:
a few further speedups to the decTest parser
 21:19
21:22 hercynium joined
cotto karma christop...@mksig.org 21:32
purl christop...@mksig.org has karma of 1
cotto christop...@mksig.org--
rakudo: say "hi"
polyglotbot OUTPUT[Parrot VM: Can't stat languages/perl6/perl6.pbc, code 2.ā¤main: Packfile loading failedā¤]
dalek rrot: r40326 | cotto++ | trunk/CREDITS:
[CREDITS] add Google-mangled email for cotto
 21:33
21:39 bacek joined
darbelo cotto++ 21:42
There, now the karma balance has been restored. 21:43
21:48 mokurai joined
cotto feels peaceful 21:53
davidfetter mmm...peacefulness
darbelo Hmm. I was about to say that dalek missed a commit, but r130 isn't on the google-code RSS either. 22:02
cotto how odd 22:03
davidfetter ...too quiet 22:04
darbelo google is stealing our karma! OH NOES! 22:06
jonathan FEAR NOT! Microsoft and Yahoo! are joining forces to fight this almighty evil! 22:07
cotto btw, do a search for "recursion" and note the correction
jonathan :-)
particle whee! the temp just reached 40F 22:12
22:13 MoC joined
darbelo Holy shit! 0x40F degrees? 22:13
darbelo once built a hexadecimal thermometer. 22:15
cotto particle, what's that in American? 22:18
;)
bacek good morning
purl Here I am, brain the size of a planet, and all they say is 'Good Morning'
bacek bad girl 22:19
particle 104F 22:20
i meant 40C of course.
jonathan particle: I was gonna say. If you are saying it's too hot to sleep at 40F... ;-)
particle the last time seattle hit 100F was 115 years ago 22:23
jonathan Wow. 22:24
cotto It's kinda nice. I don't need a toaster to make toast.
jonathan Quite an event then.
cotto: Which might be a good job, if everyone else is cranking the AC...
;-) 22:25
cotto Heh. My condo complex had a power outage a couple nights ago. I'm glad to have new cables in place now. 22:26
22:32 Limbic_Region joined 22:38 rg joined 22:39 kid51 joined 22:49 abesapien joined
dalek kudo: 23a640d | jnthn++ | (6 files):
Implement .^can returning something that as well as being useful in boolean context can also be invoked to run the first method that would be handed back or used as an iterator to get all methods that we could call. Also re-work a few things in P6Invocation to look more like it probably should have in the first place - I'd seen this coming anyway. One bug .^can on proto-objects + iteration does not yet work.
 22:55
kudo: 13ba2f3 | jnthn++ | perl6.pir:
Oops, forgot a file in the last commit.
22:55 patspam joined 22:57 tetragon joined 23:02 allison joined 23:16 dukeleto joined
dalek TT #889 created by dukeleto++: parrot_debugger generates a Bus error when deleting a breakpoint 23:19
dukeleto it is over 100 F in PDX this week. the locals don't know what to do with themselves 23:20
treed Heh. 23:23
mid 80s here in Los Angeles
(Still pretty warm feeling, especially in the sun.)
jonathan rakudobug? 23:24
purl rakudobug is mailto:rakudobug@perl.org
jonathan japhb: There ^
japhb thx. :-)
dalek cnum-dynpmcs: r132 | darbelo++ | trunk/aux/decTest/src/pct/ (2 files):
Update the decTest code generation to use the new [set|get]_rounding_mode
 23:30
purl methods are slower than constants.
23:31 MikHel joined
MikHel Hi! 23:31
How do I debug PAST creation? I got something that get parsed right, but the past step does not call the method for the object I have and I cannot figure out why not. 23:32
rg you can try --target=past 23:33
MikHel rg: That's what I do and I get "No result object" 23:34
rg that sounds like you're not generating any past at all
MikHel Even though the --target=parse shows that it was parsed correctly. One of the objects is not calling the method.
Well, one sub-past is happening. 23:35
rg did you create actions to generate past?
MikHel yes.
Ah. I think I have found something ... Though not sure why it would be a problem. 23:36
darbelo MikHel: Try checking if some of your rules are missing their {*}s 23:37
rg iirc you can use builtins in parser actions, so you could put a say or print in there (if you have such a builtin) to see if the action is actually called
MikHel I put a say and I now see that one action is called but chokes on the "make". 23:38
Not sure why though. But I admit to not understand much yet to this :D 23:39
OK I figured it out :) 23:43
It was my misunderstanding of what $/ means I guess. 23:44
23:46 AndyA joined
rg well you figured it out, that's what counts ;) 23:46
MikHel Yup :) 23:47
How proficient are you guys with this parrot? :)
darbelo It varies both with the person and the part of parrot :) 23:48
rg good answer :)
darbelo I had to say *something*. Even if it didn't tell him anything :) 23:49
dalek cnum-dynpmcs: r133 | darbelo++ | trunk/t/ (4 files):
Kill the decTest2pir.pl-generated tests, the script is dead and the generated
 23:50
23:50 dukeleto joined
MikHel :) Well I have read the squaak tutorial and I do not like the way he handles the scope of variables. It seems to me, if I understand correctly, that if you were to run two parser in different threads, the results would end up incorrect. 23:51
Anybody has an opinion in this?
darbelo MikHel: I'm not very familiar with the Squaak titorial, but I think every parrot thread gets it's own interpreter. 23:52
rg i haven't used threads in parrot at all, so i have no idea.
MikHel darbelo: Well, the squaak interpreter creates an apparently "Global" variable that is a stack of contexts and this is used to resolve where variables are scoped. Reading the "our" explanation, it seems that this variable is associated with the package... 23:54
Whiteknight japhb: I was hoping to get the first bits of AIO in by 1.5.0, but don't know what my schedule is looking like now
MikHel I think it is better to resolve the scope from the root down later.
Whiteknight Coke: ping
japhb Whiteknight: $work sidetrack you?
Whiteknight $work and $family
japhb sorry to hear that! 23:55
jonathan MikHel: I think that you're probably right about the risk of Squaak's approach leading to such issues if two parses were running in parallel, yes. 23:59
Whiteknight it could be worse! I could have no work and no family!
treed is in the first boat.
jonathan MikHel: I guess a better approach would be associating them with an instance of the parser or similar.

Powered by Cro and the RakuĀ® Programming Language.

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

Thank you!