www.parrot.org/ | Parrot 1.4.0 "Mundo Cani" Released! | For 1.5: Remove Deprecated Features | Planet Parrot planet.parrotcode.org/
Set by moderator on 4 August 2009.
NotFound GeJ: look at the comments at the fisrts lines of include/parrot/vtable.h 00:20
00:36 payload joined
GeJ NotFound: Ah, perfect. Thanks. 00:37
cognominal what means PIR_FRAGMENT_INVALID in a pod? that the the doc is not up to date? 00:58
Whiteknight it means that a test somewhere is parsing the code in that file, and the INVALID snippets don't parse 01:00
it's a system to help verify that the code in our documentation is accurate 01:01
cognominal that may means that the snippets may be valid in some larger context?
Whiteknight yes 01:06
cognominal (Note that in this case you must not use ::Petfood in the inner declaration, 01:09
or it would rebind the type to type of the actual food parameter.)
I can't parse that sentence in S-14-roles 01:10
*to the type?
Coke PIR_FRAGMENT = valid in a larger contxt. 01:24
PIR(_FRAGMENT)?_TODO = bit of pir that doesn't compile yet, but should eventually.
PIR(_FRAGMENT)?_INVALID = bit of pir that doesn't compile and shouldn't. the example is probably wrong.
01:24 theory joined
Coke (perldoc t/examples/pod.t) 01:25
cognominal thx you all 01:32
dalek rrot: r40400 | coke++ | trunk/t/examples/pod.t:
Docu-fix
 01:33
01:48 kid51 joined 01:53 rhr joined, JC1 joined
eternaleye cognominal: role Pet[::FoodType] { method eat( FoodType $noms) {...} } #[ Pet is generic; FoodType will be bound to the type passed at composition. That type will be used int the signature of 'eat'. my $foo = Pet[Chowder].new; my $bowl = Chowder.new; $foo.eat( $bowl ) ] role Pet[::FoodType] { method eat( ::FoodType $noms ) {...} } #[ When you call 'eat' here, it will /redefine/ what FoodType means to mean the type of the 02:11
argument it was passed ]
02:11 Andy joined
eternaleye cognominal: make sense? 02:11
02:13 Socro joined
Socro hi guys 02:21
may i ask 1 stupid question and few not stupid ? 02:22
eternaleye Socro: Sure
Socro :)
1st stupid: how do i get .pbc code from my script? 02:23
i mean:
for example i started to try partcl 02:24
it compiles well, and i got tclsh.pbc as result ( .pbc of interpeter)
and now i can run script like : parrot tclsh.pbc foo.tcl
but how can i get foo.pbc ? :)
02:25 wayland76 joined
eternaleye One sec 02:25
( parrot tclsh.pbc --target=pir > foo.pir ); parrot -o foo.pbc foo.pir 02:27
That's how IIUC
Socro eternaleye, thank you. one sec, i'll try.
eternaleye Whoops, I made an error 02:29
( parrot tclsh.pbc --target=pir foo.tcl > foo.pir ); parrot -o foo.pbc foo.pir
^^^^^^ that was missing 02:30
Socro seems tclsh.pbc doesnt "eat" --target option
so that might be problem of tcl implementation.
eternaleye Socro: did you see my correction?
Socro yes
eternaleye Hm.
It is supposed to work as long as the HLL uses PCT. 02:31
(I copied that code from rakudo's makefile; that's how it makes rakudo)
Socro ok. i got idea for "to what direction i should dig", thank you ;) 02:32
let's move to non-stupid questions
... 02:33
there are few runcores in parrot
(ah, btw, i want to embed parrot machine in my c++ code) 02:34
is it possible to make c++ "host" application to be like "clock generator" for parrot processor
so, i could manually call parrot_execute_next_op() from c++ code, whenever i want it. 02:35
? :)
(parrot_execute_next_op() - just for instance)
02:35 janus joined 02:39 mikehh_ joined 02:56 theory joined
Coke partcl doesn't use PCT, sorry. no way to generate a pbc from a .tcl file atm. 02:56
however, 99% of the other HLLs you can do that from. 02:57
(holy crap someone tried to use partcl. =-)
Socro Coke, thank you anyway. i just need to know that .pbc is possible.
:)
TCL rules.
Coke socro - it will be for partcl eventually. partcl existed before PCT, and we haven't made the jump yet. 02:58
Socro Coke, how many memory eats parrot?
Coke using partcl, at least ,it uses quite a bit. 02:59
Much better than it was even 2 months ago, though.
Socro is there option in embed/extend API to set maximum memory size for application usage?
Coke no, but there's a ticket for that. 03:00
(we should allow for that, yes.) 03:01
Socro is it possible to manually call "parrot, please execute next op" from c++ code?
:)
(please, dont tell RTFM, i am doing it last week) 03:02
Coke Socro: probably, but that's more of a low level question for me. 03:03
if you like tcl, please let me know if you find something missing that would make it more useful for you. (feel free to open a ticket.)
Socro i am able to do hack in c sources of parrot to get "execute next op" for my need, just wanted to know if it is already done there 03:05
Coke if no one is awake, that's a fine question for the dev list. 03:06
wayland76 I'm presuming by "the dev list", you mean parrot-dev@lists.parrot.org Right? 03:08
Socro Coke actually, i want to make some swarm-computing application. and now considering parrot as virtual machine for that. 03:09
so that's like "how many angels can dance on a pin" - "how many parrots can run in 1Gb of RAM" 03:11
davidfetter why not a TB? 03:16
we'll have that much in production machines, pretty soon ;)
Coke wayland76: yes. 03:18
Socro davidfetter , just for rate "ppg" parrots per gb.
Coke nite.
davidfetter you down with ppg?
jdv79 wow, spring. that's a bold move, pmichaud.
davidfetter (yeah, you know me)
03:20 donaldh joined 04:12 TiMBuS joined 04:16 satrac joined 04:28 tetragon joined 04:59 Andy joined
cotto seen kid51 05:07
purl kid51 was last seen on #parrot 1 days, 2 hours, 14 minutes and 27 seconds ago, saying: aspires to be cool some day, too. [Aug 4 02:47:00 2009]
cotto 1 days ago.
05:16 workbench joined 05:27 cottoo joined
cotto stupid crashy xine is making me crabby 05:39
05:53 chromatic joined 05:58 uniejo joined
cotto chromatic, ping 06:00
chromatic pong 06:02
cotto am I likely to step on your toes if I add the profiling runcore to pluggable_runcores? 06:03
chromatic Is it just one runcore?
cotto Yes. 06:04
chromatic in src/runcores/cores.c?
cotto That's the plan.
chromatic Mostly in one main function, say runops_profile_core?
cotto Yeah, except not that function.
since it already exists
chromatic That won't hurt anything. 06:05
I'll have to merge it into my work, but it'll be easy.
cotto (plus whatever code is needed to make it usabke from the cli)
ok
Is the other profiling runcore worth keeping? 06:06
chromatic I don't think it's usefl. 06:07
useful
cotto I'll just use that then. That'll make my life much easier. 06:08
chromatic It shouldn't hurt anything for my changes either. 06:09
nopaste "mikehh" at 90.208.192.110 pasted "codetest FAILures at r40400" (46 lines) at nopaste.snit.ch/17470 06:10
"mikehh" at 90.208.192.110 pasted "PATCH for codetest FAILures at r40400 - make test/ make codetest PASS after patch" (19 lines) at nopaste.snit.ch/17471
chromatic This refactoring is a lot of fiddly bits. 06:18
cotto bits that can't be done in stages? 06:20
chromatic The stages aren't five minutes long.
I guess they could be, but I haven't done them that way.
Mostly it's just adding a lot of code, then moving a lot of code, and finally deleting some code.
Lots of code in parallel. 06:21
cotto sounds familiar
chromatic My branch doesn't even build now, so I'm still working on this step. 06:22
cotto Is clock_gettime portable enough that I won't have to do something special on windows? 06:27
chromatic POSIX 2001, so it should be.
cotto That'd be great. It's perfect for profiling. 06:28
dukeleto how do I modify pasm registers from C? 06:30
mikehh dukeleto: the PATCH nopaste.snit.ch/17471 applies to src/debug.c from your commit ar r40395 06:32
chromatic dukeleto, there are register access macros at least in the .ops files. 06:33
dukeleto chromatic: thanks 06:34
mikehh: touché, i really should run the codingstd_tests. do you have commit access yet? 06:35
moritz he doesn't, but will soon 06:36
mikehh I got to post my CLA today - the printer/scanner in the house is not connected to my computer - I got to fix that 06:37
TonyC MS VC++ doesn't implement clock_gettime 06:38
chromatic Should be a C core library function. 06:39
moritz what about gettimeofday?
TonyC MS don't claim to be POSIX compliant, only ANSI
nope
perl's Time::HiRes uses Win32 API calls to get an accurate time 06:40
urr, precise
06:41 payload joined
japhb remembers writing the first version of the T::HR Win32 code ... ah, horrid, horrid memories ... 06:44
TonyC it's kind of scary
japhb I'm pretty sure it's different now than my original hack, though the replacement may have been just as bad. :-) 06:45
The best timing code is probably from Devel::NYTProf. He seems to have worked hard to get the best out of each OS. 06:46
cotto NYTProf FTW 06:48
I'm glad we have lots of smart people to steal from.
TonyC NYTProf uses Time::HiRes to get the time if gettimeofday isn't available 06:51
moritz (code reuse)++ 06:52
06:52 HG` joined
cotto There's gotta be some C I can steal somewhere in there. 06:53
TonyC kind of cute - Time::HiRes puts the address of it's myU2time and myNVtime functions into perl globals and NYTProf grabs their values and calls them
cotto chromatic, if I commit now and continue to commit changes later, will that cause you much pain? 07:12
chromatic It won't hurt anything. We're in a decent spot of encapsulation now. 07:14
cotto incoming 07:17
purl incoming is pause.perl.org/incoming/
dalek rrot: r40401 | cotto++ | branches/pluggable_runcore (4 files):
[profiling] repurpose the old profiling runcore and remove some now-unused functions

There are many things that need work in this code. I'm aware of some of them.		 07:18
07:20 donaldh joined 07:28 particle joined
dalek rrot: r40402 | cotto++ | branches/pluggable_runcore/src/runcore/cores.c:
[prof] fix a copy/paste bug
 07:31
rrot: r40403 | cotto++ | branches/pluggable_runcore/src/runcore/cores.c:
[prof] and another. cotto--
07:52 mokurai left 08:34 hiroyuki_y joined 09:15 wayland76 joined
moritz it would be nice if somebody could alias planet.parrot.org to planet.parrotcode.org 09:25
dalek rrot: r40404 | NotFound++ | branches/auto_attrs:
Creating auto_attrs branch for TT #895
 09:28
cotto It looks like Parrot_floatval_time will be an easy (if inferior) fallback for platforms that don't have clock_gettime. 09:32
09:38 Whiteknight joined
dalek rrot: r40405 | NotFound++ | branches/auto_attrs (26 files):
put the branch up to date with the changes already tested from TT #895
 09:38
09:49 donaldh joined
cotto Whiteknight, what's a good way to detect (in C) if you're in an inferior runloop? 09:56
NotFound cotto: all runloops are created equal... 09:57
Whiteknight cotto: I'm not sure about the best way 09:58
cotto except the inferior ones
NotFound cotto: inferior in respect to what?
Whiteknight there is a runloop id field somewhere, maybe in the interp struct that keeps a count, but I don't know how it is updated and used everywhere 09:59
cotto I guess another word would be "inner".
Whiteknight "innerferior"
NotFound Any inner runloop related to the first one of the current Interpreter?
cotto yeah 10:00
NotFound Maybe we need to store somewher the id of that first one. 10:01
dalek tracwiki: v2 | kjs++ | PIRCDevelopment 10:05
tracwiki: trac.parrot.org/parrot/wiki/PIRCDe...ction=diff
NotFound Mmmmm... the runloop id is set in runops at src/call/ops.c, and then it calls runops_int. But there are calls to runops_int from other places, so I guess we need some fixes to be able to reliably use the runloop id. 10:06
cotto There's an easy workaround for what I need.
Meh. I was hoping to profile something using PGE tonight, but that'll have to wait until later. 10:12
10:13 bacek joined
dalek rrot: r40406 | cotto++ | branches/pluggable_runcore/src/runcore/cores.c:
[prof] make profiling runcore behave better with nested runloops
 10:19
cotto good night 10:21
(or morning)
dalek rrot: r40407 | cotto++ | branches/pluggable_runcore/src/runcore/cores.c:
[prof] It's really time I stopped writing code and went to bed.
 10:23
bacek o hai 10:31
cotto: go to bed! 10:32
10:36 MoC joined
pmichaud Good morning, #parrot 10:44
10:44 Zak joined 10:45 zak_ joined
bacek hi pmichaud 10:45
But I disagree about "morning" :) 10:46
dalek rrot: r40408 | NotFound++ | trunk (2 files):
Deprecate PASM1 pseudocompiler and PDB_compile, TT #872
rrot: r40409 | NotFound++ | branches/auto_attrs (11 files):
merge trunk r40408, mainly to check that the branch is alive
 11:00
rrot: r40410 | bacek++ | trunk/src/debug.c:
[cage] Fix hardtab.
 11:04
NotFound Uh, sorry.
bacek NotFound: no worries. 11:05
I'm just merging my branch to trunk atm. So I'm running a lot of tests before dcommit them :) 11:06
NotFound I'm familiarizing myself with svn branching and forgot to look at the details of the changes.
bacek NotFound: use git :) 11:07
NotFound bacek: familiarizing with git will be even worse :) 11:08
bacek much simpler to use
But you can do it locally without showing to everyone how bad you are :)
NotFound I'm on holidays, don't want to read manuals ;)
bacek fair enough :) 11:09
Care to fix t/tools/parrot_debugger?
t/tools/parrot_debugger (Wstat: 256 Tests: 35 Failed: 1) 11:10
Failed test: 35
(Or I can do it)
NotFound Uh, there was a test for that? :o
'eval nothing' ? That's funny 11:11
I'll fix/kill it
bacek -Ofun ftw!
dalek rrot: r40411 | bacek++ | branches/keys_cleanup/compilers/imcc/pbc.c:
[cage] Remove unused macro. kid51++ for notice.
 11:14
rrot: r40412 | NotFound++ | trunk/t/tools/parrot_debugger.t:
[t] TODO parrot_debugger eval test, TT #872
 11:17
11:20 donaldh joined
dalek rrot: r40413 | NotFound++ | branches/auto_attrs (11 files):
merge from trunk r40412
 11:21
11:30 uniejo joined
bacek pmichaud: Can I delete isafast branch from svn? 11:30
dalek rrot: r40414 | bacek++ | trunk (13 files):
Merge branch keys_cleanup into trunk.
 11:31
rrot: r40415 | bacek++ | branches/keys_cleanup:
Branch keys_cleanup was merged to trunk
11:35 kid51 joined 11:40 MoC joined
wayland76 purl, seen allison 11:40
purl allison was last seen on #parrot 17 hours, 19 minutes and 16 seconds ago, saying: :)
moritz tests rakudo with the keys_cleanup branch merged 11:44
ttbot bacek: Parrot trunk/ r40414 cygwin-thread-multi-64int make error tt.ro.vutbr.cz/file/cmdout/66253.txt
bacek oh shi... 11:45
dalek rrot: r40416 | NotFound++ | branches/auto_attrs (22 files):
merge from trunk r40415 (keys_cleanup)
bacek Segfault??? 11:46
purl Segfault is xkcd.com/371/
bacek Does anyone have "cygwin-thread-multi-64int" around for "interactive debugging"? 11:48
moritz bacek: after the branch merge rakudo segfaults on precompiling Test.pir 11:49
on amd64
bacek a-ha!
Let me recheck locally (i386 anyway). 11:50
NotFound Looks like there are some important thing that out test suite fails to exercise. 11:51
moritz that's not really new :/ 11:52
bacek key_cleanup branch should be pretty safe... It's just about removing unused stuff... 11:53
moritz well, I didn't test it immediately before the merge 11:56
bacek moritz: I can't reproduce segfault. Can you rebuild parrot/rakudo after "ccache -C"?
moritz bacek: sure, I'll do that
bacek moritz: thanks
moritz so far I never had problems with ccache, but one can never know
bacek I had quite few of them. Especially after merging branches. 11:59
So I've put in my check list: "build without ccache after merge"
Second segfault on darwin-multi... 12:00
yak...
moritz bacek: clearing the ccache didn't help 12:02
bacek moritz: hrm... 12:03
moritz seems to be GC related
when I compile it by passing -G to parrot there's no segfault
nopaste "bacek" at 114.72.150.10 pasted "moritz: can you try this patch?" (14 lines) at nopaste.snit.ch/17473 12:04
bacek oh noes... Not GC again! 12:05
bacek running in circles screaming 12:06
moritz: can you run it under gdb?
moritz bacek: ok, I'll try that. If it doesn't help, I'll use gdb or bisect
bacek Yay! I reproduced it with --gc-debug 12:07
moritz: ignore my patch. 12:08
moritz bacek: would a bisect help you? 12:09
12:10 whiteknight joined
bacek no... Only lobotomy... 12:10
moritz: can you try again on r40417? 12:13
moritz bacek: just a sec, I've got something else running right now
bacek moritz: ok 12:14
whiteknight good morning moritz, bacek
bacek whiteknight: 2 mistakes. It's not morning. And it's not good...
bacek screwed parrot again 12:15
whiteknight okay
bad evening bacek
how did you screw it?
bacek And bad evening to you, whiteknight :)
dalek rrot: r40417 | bacek++ | trunk/src/key.c:
[cage] Don't try to mark non-GCable keys
bacek r40417 12:16
merge keys_revamp branch...
whiteknight ah, I see
good! That branch is very important
bacek whiteknight: not yet. But I almost wrap my head about "keys_refactor" branch. So I can screw parrot even badly! 12:17
whiteknight I don't have my development computer at work and can't really help testing
12:17 JC1 joined
moritz bacek: no immediate segfault on r40417 12:18
kid51 starts a smolder test on 40417 and heads for $job
bacek no worries. We have moritz and ttbot :)
moritz bacek: running spectest now, expect results in ~30min 12:19
the first 10 or so test files look good
bacek moritz: ok, thanks.
I removed such a nice comment in r40376 12:21
I probably have to put it into include/parrot/key.h...
"* Note to self: shoot whoever thought this was a good idea." 12:22
Yay! "Green light" on ttbot! 12:24
dukeleto bacek: hilarious comment 12:31
12:31 MoC joined
dalek rrot: r40418 | NotFound++ | branches/auto_attrs (10 files):
merge from trunk r40417
 12:32
12:32 quek joined
bacek dukeleto: not for me... I'm cursing day when I decided to make hashes/keys less insane. 12:33
dukeleto bacek: i can run tests on mac ppc/intel if you need it 12:34
moritz bacek: you can see yourself as a cursing buffer - you do it now extensively, and safe the curses of a future generations of parrot hackers for something even more worthwile
dukeleto: a 'make smolder_test' on parrot HEAD would surely be welcome 12:35
dalek rrot: r40419 | bacek++ | trunk/src/key.c:
[cage] Put parentheses in if statement.
 12:36
dukeleto moritz: will do
12:36 ruoso joined, bacek_ joined
bacek_ me again... 12:37
12:38 ruoso joined
dukeleto is it safe to pass -j flags directly to Parrot's make or is there an environment variable? 12:39
bacek_ dukeleto: not yet. Andy++ working on proper make dependencies
dukeleto bacek: good to know
12:40 ruoso joined
moritz dukeleto: works for me here 12:41
dukeleto: I always build with 'make -j3', and only if there are build problems I'll retry without it
but that happens only rarely these days
dukeleto moritz: ok, also good to know
moritz but I usually do a 'make realclean' before, so there's no old stuff left 12:42
bacek_: spectest finished, no new problems in rakudo so far
12:42 ruoso joined
bacek_ departing to bed. It will be big meeting tomorrow at $dayjob. 12:43
moritz: ok. thanks for testing.
see you!
moritz sleep well
bacek_ I'll try. 12:44
12:44 bacek_ joined, ruoso joined
dalek rrot: r40420 | NotFound++ | branches/auto_attrs/src/pmc/nci.pmc:
set auto_attrs in nci pmc
 12:46
rrot: r40421 | NotFound++ | branches/auto_attrs/src/pmc/unmanagedstruct.pmc:
set auto_attrs in unmanagedstruct pmc
 12:52
dukeleto moritz: just got a passing "smolder_test" on r40419 for darwin 386 12:56
12:56 quek left
dukeleto how do I tell ASSERT_ARGS about new C functions ? 12:57
NotFound dukeleto: make headerizer 12:58
dukeleto NotFound: thanks! 12:59
i am making good progress on adding the "assign" command to the debugger
NotFound dukeleto: better do a make headerizer before doing any change, to make sure there was no pending headerizing.
dukeleto: good :) 13:00
dalek rrot: r40422 | NotFound++ | branches/auto_attrs/src/pmc/managedstruct.pmc:
set auto_attrs in managedstruct pmc
dukeleto NotFound: good to know
13:01 quek joined
moritz dukeleto: thanks 13:02
dukeleto i just realized that my darwin ppc smolder cron job has been failing due to a certificate warning in svn co 13:06
13:12 payload joined 13:21 quek left
Coke (-j) has been working for quite some time now. it is, generally, safe. 13:30
I am not understand how something that never worked needs a deprecation cycle. Did allison have a compelling argument? 13:35
13:57 TiMBuS joined 14:13 mikehh_ joined
whiteknight that is quite an interesting question 14:15
moritz I think she was under the impression that there were some cases where it was working before
whiteknight perhaps a better idea would have been to grandfather all non-working items that existed before the deprecation policy as being "experimental"
but that's hindsight, and the existance of the PASM1 compiler doesn't really hurt us if it doesnt work and nobody relies on it 14:16
dukeleto am I correct in thinking that REG_INT(interp,3) = 42 should be equivalent to the pasm "I3= 42" ? 14:17
dalek kudo: 3c5210a | moritz++ | build/PARROT_REVISION:
bump PARROT_REVISION after keys_cleanup merge by bacek++
 14:25
14:26 theory joined 14:48 MoC joined
Coke if it's been deprecated and is now throwing an exception when using it, (which is how I read the commit, but didn't check the code), then we've done just as bad as simply removing it, except now we have to remember to clean up our mess later. 14:50
If we're going to leave it alone until 2.0, let's leave it alone entirely.
whiteknight send a message to the ML. you've got my vote 14:51
14:51 Psyche^ joined
Coke hurm. [namespace eval foo {namespace parent}] should return "::", the top level namespace. works in tcl now. if I convert all the helper subs in [namespace] into a switch (removing one parrot sub invocation), it fails. 15:13
(returning the empty string instead.)
odd.
15:20 donaldh joined
mikehh All tests PASS (pre/post-config, smolder, nqp_test, fulltest) at r40422 - Ubuntu 9.04 amd64 15:49
15:55 mokurai joined
mikehh rakudo (3c5210a) builds on parrot r40422 - make test PASS / make spectest (up to 27878) FAIL - same two tests fail that have been failing for the last week 16:09
16:31 jtal joined
jtal does anyone have a moment to tell my why this is wrong? kivasti.com/p.txt 16:35
when I look at --target past the name is not what I expect, instead its PMC 'bla;Grammar' => "myFunc" @ 1 16:36
whiteknight jtal: are you defining your namespaces in those files? 16:37
jtal I'm starting from the boilerplate stuff... so there is .namespace [] in say.pir
I've changed very little
whiteknight the .pg file should have a line near the top "grammar MyGrammar::Grammar is PCT::Grammar;", or whatever your language is named 16:39
jtal ah yes thats there
whiteknight and the .pm file should have something like "class MyGrammar::Grammar::Actions;" 16:40
jtal yes 16:41
whiteknight okay, maybe I don't understand the problem
jtal there is no error either... I forgot to mention that
when I type something that parses correctly it doesnt complain or do what I expect
hm, maybe the problem is in the pir then 16:42
ok, thanks for your help...I think I need to read more :> 16:43
16:46 davidfetter joined
whiteknight okay 16:49
17:29 zak_ joined
dalek rdinal: ba72bd4 | (Danius Michaelides)++ | (3 files):
Correct return value for Array, inc tests.
 17:35
17:35 theory joined
moritz admires the activity level of the cardinal compiler 17:37
treed Thanks. :-)
Mostly a lot of little things, but progress is progress.
moritz aye 17:38
17:38 payload joined 17:48 braceta joined
Coke is there a way to say, "show me the changes in commit id foo" ? 18:08
(I can do a log, and a 'git diff <commit> <previous commit>' but was hoping for a shortcut version. 18:09
18:09 kj joined
dalek tracwiki: v3 | kjs++ | PIRCDevelopment 18:11
tracwiki: trac.parrot.org/parrot/wiki/PIRCDe...ction=diff
18:11 altious joined
Coke packy's "gitsquash" script is verrrrry nice. 18:12
dalek tracwiki: v4 | kjs++ | PIRCDevelopment 18:15
tracwiki: trac.parrot.org/parrot/wiki/PIRCDe...ction=diff
tracwiki: v5 | kjs++ | PIRCDevelopment 18:18
tracwiki: trac.parrot.org/parrot/wiki/PIRCDe...ction=diff
tracwiki: v6 | kjs++ | PIRCDevelopment 18:22
tracwiki: trac.parrot.org/parrot/wiki/PIRCDe...ction=diff
cotto <3 kj 18:23
(or at least his documentation
)
kj heh :-) 18:24
moritz kj++ 18:25
no he's <4 ;-)
s/no/know/
*now
ENOTYPING
18:25 l3t0 joined
cotto btw, moritz, I think your rss feed is goofy 18:25
all the links of the titles point to the rss feed itself rather than the articles 18:26
moritz ouch
cotto: thanks for the information
dalek tracwiki: v7 | kjs++ | PIRCDevelopment
tracwiki: trac.parrot.org/parrot/wiki/PIRCDe...ction=diff
cotto np 18:27
l3t0 Coke: git show can be used to show the changeset of a commit sha1 or branch (or a general 'commitish')
18:28 joeri joined
Coke l3t0: woot. thanks. 18:30
dalek tracwiki: v8 | kjs++ | PIRCDevelopment 18:37
tracwiki: ah while i'm at it, some more information on PIRC.
tracwiki: trac.parrot.org/parrot/wiki/PIRCDe...ction=diff
tracwiki: v9 | kjs++ | PIRCDevelopment 18:44
tracwiki: add link to register allocator paper that i based the reg alloc on
tracwiki: trac.parrot.org/parrot/wiki/PIRCDe...ction=diff
tracwiki: v10 | kjs++ | PIRCDevelopment 18:52
tracwiki: some comments on PIRC vs IMCC
tracwiki: trac.parrot.org/parrot/wiki/PIRCDe...ction=diff
kj ah this always happens... just starting to write without any planning :-( 18:53
18:53 altious joined
kj amazing how fast you can make a mess of things 18:53
18:55 danius joined
dalek tracwiki: v11 | kjs++ | PIRCDevelopment 18:59
tracwiki: trac.parrot.org/parrot/wiki/PIRCDe...ction=diff
kj does anyone know whether I can use the full path of a file in the repository, and have it link automatically to that file on the online repos? 19:10
without making explicit links? Bascially what happens for trac tickets automatically, I'd like to do for files in the repos.
particle [source:/README] 19:12
Coke [source:README] 19:13
kj ok, thanks for that
Coke except that doesn't seem to work. =-)
particle wonder if we need a plugin for that :( 19:14
Coke probably need trunk/
yup. 19:15
[source:trunk/README]
particle coke++
kj mm doesn't work 19:19
i want to link to compilers/pirc/src/pirregalloc.c
so i'd do [source:trunk/parrot/compilers/pirc/src/piregalloc.c]
moritz without the parrot/ in there 19:20
19:20 donaldh joined
kj thanks! 19:21
dalek tracwiki: v12 | kjs++ | PIRCDevelopment
tracwiki: add link to a file that i mentioned before.
tracwiki: trac.parrot.org/parrot/wiki/PIRCDe...ction=diff
19:24 chromatic joined 19:29 Andy joined
Coke Andy, chromatic: morning. 19:30
Andy I've been so busy I haven't even logged in to IRC>
particle morning, where?
purl morning, is probably this a submarine?
l3t0 purl, a yellow submarine ?
purl YELLOW SUBMARINE
Coke we all live in. 19:31
19:34 tekky joined
chromatic Hello. 19:37
kj hello chromatic 19:38
19:38 iblechbot joined
whiteknight hello chromatic 19:42
chromatic cotto, I've been thinking of appending the PID to the profiling output name so as not to clobber existing outputs. 19:50
whiteknight PID? 19:51
purl i think PID is a unique number that identifies a single instance of a program. or pelvic inflammatory disease. or $$ or Passenger Information Display
whiteknight ah yes, we want the pelvic inflamatory disease in the profiling output
particle those are rarely unique 19:53
rather, they're usually a shared pubic key
19:53 altious joined
japhb I caught the Perl 5 SDL_perl guys talking about porting their work to Parrot, and was about to point them to the right places to coordinate with us. Is parrot-dev the right mailing list for people who want to work on Parrot modules, rather than Parrot itself? 20:11
Tene For now, yes. 20:12
It might be nice to have a separate list, a spart of the "aviary" or whatever proposal.
japhb OK, I'll add that to my notes 20:13
OK, pointers sent. Hope they come join us. 20:18
cotto chromatic, that's actually on my todo list 20:19
20:24 payload joined
tekky PID=process ID 20:31
cotto I was going to use a passenger information display, but that'll be considerably easier. 20:34
tekky use a principal ideal domain
moritz parrot idiosyncratic developer ;-) 20:36
tekky Like that one. Is that another name for an intern?
moritz I just came up with that 20:37
20:46 bacek joined
cotto hi bacek 20:54
20:56 donaldh_ joined 20:59 MoC joined
dalek kudo: a948cae | moritz++ | docs/architecture. (2 files):
[docs] add links to architecture.svg, and a HTML document to which it links.

The HTML is take from my blog, needs a bit polishing and some style sheets.		 20:59
21:01 theory joined
bacek good morning 21:04
purl And good moroning to you, bacek.
bacek chromatic: around? 21:06
chromatic pong
bacek chromatic: I've got excellent (fsvo) idea. Deprecate Keys at all. 21:07
chromatic What would replace them?
bacek Create NamespaceKey which will be used for Namespaces.
21:08 JC1 joined
bacek Navigating over "ordinary" aggregates will be "manual" 21:08
chromatic The idea has merit. 21:09
bacek indeed.
No one uses $P0[foo;bar;baz] 21:10
It's almost impossible to implement and used correctly.
And behaviour is very dependant on HLL 21:11
chromatic It's not easy to manage in IMCC either.
bacek e.g. autovivify vs throwing exception.
So, NamespaceKey will has-a RSA 21:12
Simple, clean, comprehensible.
Remove a lot of code in Hash and Array to support Keys.
Simplify IMCC and future of PIRC (which doesn't support complex keys) 21:13
chromatic It's worth considering.
japhb How does one grab Parrot config info from PIR?
chromatic There's a Parrot config library. config.pbc 21:14
21:15 Whiteknight joined
japhb chromatic: Ah! I was mentally stuck on the parrot_config.pbc at root that had no .pir, and it wasn't clear what to do with it. 21:16
Coke bacek; no one uses $P0[foo;bar;baz] because you can't really do that at runtime. 21:19
(compiletime, sure.)
bacek Coke: indeed.
Coke: do you really want to use $P0[foo;bar;baz] for ordinary Hashes in partcl? 21:22
dalek TT #903 created by bacek++: [RFC] Deprecate Keys at all. 21:23
bacek departing to morning duties and $dayjob 21:26
21:32 joeri left 21:46 jan joined 21:53 bacek joined
Whiteknight Far from removing key literals in PIR, I think it would be much better to create an improved, unified aggregate literal syntax that can be used with multiple PMC types 21:54
chromatic Do we use aggregate lookup for anything other than namespaces and class names? 21:55
Whiteknight for I don't think so, but if we had a literal syntax for more PMC types it would be used more
array literals, hash literals, etc. It would be very valuable
Coke Whiteknight: as someone who is now moving those over to .const, is that a /big/ win? 22:01
22:01 MoC joined
donaldh_ anyone familiar with compilers/pirc ? 22:20
there's a macro parser in compilers/pirc/macro that doesn't seem to get used. 22:21
22:23 braceta joined
Coke most of pirc doesn't get used. 22:28
it's a work-in-progress replacement for imcc that's not built by default. 22:29
there's a newly added wiki page (and several tickets) on trac.parrot.org
donaldh_ Yep, I figured that
Coke bah: exists_keyed() not implemented in class 'ResizableStringArray' 22:30
donaldh_ I'm just eradicating use of malloc/free in the generated lexers. The macro/macro.l lexer doesn't seem to get used by pirc.
chromatic It might be a proof of concept. 22:31
donaldh_ I've updated macro.l anyway - just can't test. 22:32
Coke incoming. 22:34
purl incoming is pause.perl.org/incoming/
22:35 wayland76 joined
donaldh_ chromatic: you were interested in a patch that replaces malloc/free calls with mem_sys_allocate et al 22:36
chromatic Yes.
dalek rtcl: r570 | coke++ | trunk/src/macros.pir:
Add TryCatch for more simple exception handling.
 22:37
rtcl: r571 | coke++ | trunk/runtime/builtin/ (8 files):
use .const subs to generate these PMCs
rtcl: r572 | coke++ | trunk/runtime/builtin/namespace.pir:
use more hllmacros
rtcl: r573 | coke++ | trunk/runtime/builtin/namespace.pir:
remove unneeded subcommands
donaldh_ I've updated the lexer files and I've updated the makefiles to correctly generate the .c files in --maintainer mode.
dalek rtcl: r574 | coke++ | trunk/tools/tcl_test.pl:
Allow for a git-svn checkout.
rtcl: r575 | coke++ | trunk/ (2 files):
Issue # 42 - implement [string is ascii]
rtcl: r576 | coke++ | trunk/ (9 files):
Eliminate toDict() multi subs and use a method instead.
rtcl: r577 | coke++ | trunk/ (5 files):
gitignore more generated files
Coke that was squashed down from about 20 commits. 22:38
donaldh_ chromatic: do you want the generated .c files included in the patch, or do you want to generate them in --maintainer mode
Whiteknight Coke: depends what you mean by "win"?
22:38 rg joined
Coke how much of an improvement would that be over the existing system, esp. compared to all the other projects we have unfinished. 22:39
I would say that having a nice syntax for stuff you can already do at compile time with .const and split and JSON is probably a very low priority. 22:40
->
GeJ Good morning everyone. 22:41
22:52 tetragon joined
nopaste "donaldh" at 213.123.171.12 pasted "for chromatic: patch replacing malloc/free with mem_sys_allocate variants" (532 lines) at nopaste.snit.ch/17477 22:55
22:57 ruoso joined
donaldh_ bed 22:57
22:59 ruoso joined 23:01 ruoso joined 23:03 ruoso joined 23:05 ruoso joined 23:06 ruoso joined 23:08 ruoso joined 23:09 slavorg joined 23:10 ruoso joined 23:12 ruoso joined 23:13 ruoso joined 23:15 ruoso joined 23:17 ruoso joined
wayland76 ruoso seems to be having connectivity issues 23:17
Question all: is there an easy way to call the Linux "stat" function from Parrot? 23:18
chromatic There's a wrapper for it in src/platform I think.
The bytecode/program loading call uses it.
23:19 ruoso joined 23:20 donaldh joined 23:21 ruoso joined 23:22 slavorg joined 23:23 ruoso joined 23:24 patspam joined, ruoso joined 23:25 braceta left 23:26 ruoso joined 23:28 ruoso joined 23:30 ruoso joined, gerd joined 23:31 ruoso joined 23:33 ruoso joined 23:35 ruoso joined, slavorg joined 23:37 ruoso joined 23:39 ruoso joined 23:41 ruoso joined 23:42 ruoso joined 23:44 ruoso joined 23:46 ruoso joined 23:47 ruoso joined 23:49 slavorg joined, ruoso joined 23:51 ruoso joined 23:52 ruoso joined
cotto ruoso, it's getting old 23:53
23:54 ruoso joined
chromatic Maybe that'll help notify him. 23:58

Powered by Cro and the Raku® Programming Language.

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

Thank you!