Parrot 2.6.0 | parrot.org Log: irclog.perlgeek.de/parrot/today | Nopaste: nopaste.snit.ch:8001 | fix 'make html' (talk to Coke), update tutorial (talk to tcurtis)
Set by moderator on 21 July 2010.
dalek nxed: r584 | NotFound++ | trunk/examples/fly.winxed:
smooth movement in manual mode in example fly
 00:02
00:10 NineCross joined
NineCross please advice me a good manual for basic perl6 syntax, something like "Learn perl 6 in 24 hours". i tried to read perlgeek.de/blog-en/perl-5-to-6/ . But those examples to complicate for me 00:13
tcurtis NineCross: have you tried the Perl 6 book included as docs/UsingPerl6-draft.pdf in Rakudo Star? 00:14
NineCross: I'm about to go, but I recommend asking in #perl6 on freenode.
NineCross thanks, I'll try 00:15
00:25 ruoso joined
cotto_work seen whiteknight 00:31
purl whiteknight was last seen on #parrot 11 hours, 15 minutes and 0 seconds ago, saying: 1) Install Linux ...
cotto_work sounds like good advice 00:32
00:40 snarkyboojum joined 01:00 tommyd joined
tommyd evening all! I'm currently in the process of re-viving the macports parrot package 01:01
one particular issue I stumbled upon so far is that the soname is wrongly hardcoded to the intermediate DESTDIR path instead of the correct (later available) PREFIX path 01:02
this is easily fixable through the install_name tool afterwards, but since parrot_config echoes the wrong configuration still, dependent software like rakudo picks up the wrong default as well and isn't runnable without the install_name fix either 01:03
01:04 mikehh_ joined
tommyd so in line 90 of config/init/hints/darwin.pm the libparrot_soname is apparently set up with $lib_dir - has anybody a deeper understanding of the code there how this place can be safely tweaked to build the correct path? 01:06
01:19 rurban_ joined
cotto ~~ 01:32
mikehh opbots, names 01:42
02:11 tcurtis joined 02:28 wtgee joined
atrodo evening, #parrot 02:42
cotto hi atrodo 02:44
atrodo smell that cotto? that the smell of hacking! 02:45
cotto ooc, how much background in cs do you have, especially the low-level stuff that Lorito deals with?
atrodo i've been playing mmo's too much. all I could read was "out of character" 02:46
i'd say i got experience. I enjoy the asm challenge
most of it in x86, and had my start hacking dos nasm 02:47
sorear does "background in CS" imply formal training? 02:49
02:49 wtgee left
cotto formal or informal 02:49
atrodo luckily, i got both 02:50
well, if by formal you mean a degree and some computer architecture and compiler courses 02:52
cotto I'm realizing that there are a few things I need to brush up on. 02:56
atrodo like?
cotto general CPU architecture 03:03
atrodo Ah. Yep, cpu architecture and compilers have always fascinated me. that's why I'm more interested in parrot then perl6 03:05
dalek TT #1722 created by LaVolta++: osx build failed caused by errant '\\c' 03:06
TT #1722: trac.parrot.org/parrot/ticket/1722
atrodo p&w? 03:17
purl it has been said that p&w is tinyurl.com/23dfwut
03:19 janus joined
cotto that's a proposed object model, though allison has some reservations about it 03:24
atrodo i'm curious what her reservations are. seems like a good base 03:27
cotto me too
she didn't say what they were at the time
atrodo okay. I'll see if I can try and catch her sometime and ask 03:28
cotto me too, though it's tricky 03:29
atrodo yea. I'm trying to read through it again. I get the basic idea, but the exactness of it hasn't been fully revealed
and reading this late at night probably doesn't help 03:30
Okay, now it makes sense. But it's time to call it a night 03:41
goodnight, #parrot
cotto 'night 03:45
have nice dreams about metaobject protocols 03:46
03:57 petdance joined
darbelo msg mikehh r48223 tries to pacify splint by annotating function arguments in generated code. That might be a good place to start cutting down on the noise. 04:52
purl Message for mikehh stored.
dalek rrot: r48223 | darbelo++ | trunk/lib/Parrot/Pmc2c/PMCEmitter.pm:
Add ARGIN_NULLOK() markers to parameters in generated functions. It makes splint a bit less cranky.
 04:53
05:01 plobsing joined
dalek rrot: r48224 | darbelo++ | trunk/lib/Parrot/Pmc2c/PMCEmitter.pm:
Add ARGIN_NULLOK() markers to parameters in generated functions. It makes splint a bit less cranky.
 05:09
05:15 LoganLK joined 05:17 jsut joined
dalek p-rx: a3386d1 | pmichaud++ | (2 files):
Fix general quantifier in regexes with charclasses (RT #76792).
 05:38
p-rx: a9afc70 | pmichaud++ | src/stage0/ (4 files):
Update bootstrap.
rrot: r48225 | pmichaud++ | trunk/ext/nqp-rx/src/stage0 (4 files):
[nqp-rx]: Update with fixes for RT #76792.
 05:42
05:52 uniejo joined
cotto FPA is bigger than I'd expect. 05:57
sorear FPA? 05:58
purl i heard FPA was empty.
sorear oh fixedpmcarray
cotto no, fpa is FixedPMCArray 05:59
purl okay, cotto.
cotto FPa?
purl i heard FPa was FixedPMCArray
cotto hey purl 06:00
lorito? 06:13
purl lorito is "little parrot" in spanish or xkcd.org/707/ or github.com/atrodo/lorito or trac.parrot.org/parrot/wiki/Lorito
cotto someone else was working on another Lorito prototype
lorito is also github.com/ekiru/yalp-asm 06:15
purl okay, cotto.
dalek kudo: 50e0e7e | pmichaud++ | build/PARROT_REVISION:
Bump PARROT_REVISION to get fix for RT #76792.
 06:16
cotto sees very quickly how FPA or RPA is a good feasibility test for any aspiring Lorito implementation. 06:19
dalek rrot: r48226 | NotFound++ | trunk/src/pmc/oplib.pmc:
fix off-by-one error in OpLib
 06:32
parrot: 7b9d02d | (Jonathan "Duke" Leto)++ | html/community.html:
[html] Add a link to the new IRC logs
 06:44
parrot: 193bf38 | (Jonathan "Duke" Leto)++ | plp (2 files):
Refactor PL/Perl6 wrapper code
06:50 jsut_ joined 06:56 Casan joined 07:01 fperrad joined 07:08 baest joined
nopaste "cotto" at 192.168.1.3 pasted "concept for fpa init and init_int in Lorito" (41 lines) at nopaste.snit.ch/22459 07:16
cotto almost certainly half-baked and unworkable 07:17
flames appreciated
sorear I find your ambition lacking. 07:19
NotFound cotto: size * 4?
sorear Why are you copying the flags system? BURN BURN BURN
Current PMCs have a 4 or 5 word overhead 07:20
GHC boxed objects have 1 word, Mono boxed objects 2
cotto NotFound, 4-byte ints 07:21
oh. nm. I got confused and thought I was working on fia.
Mmmm. GHC
dalek rrot: r48227 | NotFound++ | trunk/config/auto/icu.pm:
update workaround for spurious \\c in icu config, TT #1722 LaVolta++
 07:22
NotFound cotto: We don't hardcode int size
sorear has a half-baked replacement storage layer for P5 in /home with a 0 word object header
cotto If there's a decent vtable structure, those flags won't be needed
sorear unfortunately I maxed out my ambition quota and the code can't seem to be finished 07:23
cotto I guess figuring out the object system would make this easier to figure out
sorear You do need a little bit more than the vtable. Like say, a mark bit.
cotto wishes allison wouldn't say she has misgivings about p&w without being more specific.
sorear Fortunately the bottom two bits of the vtable pointer are always 0 and can be reused during nonstandard execution phases. 07:24
cotto sure. Those are free.
cotto goes to bed 07:26
dalek parrot: f7b08d1 | (Jonathan "Duke" Leto)++ | plparrot.c:
Use PMC_ISA, and make the lack of sausage an ERROR
 07:31
parrot: f06bf9b | (Jonathan "Duke" Leto)++ | html/plperl6.html:
Update the PL/Perl6 page and give and example with explanation
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#35171), fulltest) at r48227 - Ubuntu 10.04 amd64 (gcc) 08:03
09:04 LaVolta joined 09:12 jan joined 09:16 AndyA joined 09:19 rurban_ joined 09:27 another_bacek joined
another_bacek aloha, humans 09:28
09:39 snarkyboojum joined 09:56 LoganLK joined 10:09 tommyd joined
moritz we have two baceks? \\o/ 10:20
sorear I think one of them is on a telephone or something 10:24
10:36 snarkyboojum joined 10:50 another_bacek joined
another_bacek this one in SIngapore on corporate laptop :) 10:50
moritz: ping 10:58
11:02 kurahaupo joined 11:12 kid51 joined 11:17 mikehh_ joined 11:30 robin-gvx joined
moritz atrodo: pong 11:32
sorry, meant to pong another_bacek 11:33
moritz + tab fail
11:44 kurahaupo joined 12:30 bluescreen joined 12:35 mj41 joined 12:40 ruoso joined
atrodo moritz> :( I thought i was somebody! 13:02
moritz atrodo: yes, but you didn't ping me :-) 13:03
atrodo moritz> minor detail 13:04
13:11 lucian joined
Coke cotto: nopaste.snit.ch/22459 - you can save an assignment by putting the new FPA in p0 to start with. 13:24
13:27 tommyd joined
Coke wow, the folks in #perl are discussing lorito. 13:35
s/are/were/, perhaprs (in backscroll)
13:36 mj41 joined
szbalint Coke: when? this network? 13:38
Coke yes. hours ago.
it was surprisingly positive. ;) 13:39
atrodo Any logs?
Coke none of which I am aware. 13:40
it was the idea of lorito, not necessarily your implementation, sorry.
atrodo I figured, but i wanted to see what they said
Looks like my boss may have a log of it 13:41
13:52 whiteknight joined
atrodo Coke> Yep, quick and positive. It's nice to see 14:01
szbalint I have it in my scrollback buffer :) 14:02
14:17 ter joined 14:19 plobsing joined
kthakore hi 14:20
something keeps killing my rakudo compile ...
atrodo kthakore> finally decided to awake? 14:21
14:21 ter left
kthakore atrodo: awake? 14:22
atrodo no, not really. Still rather sleepy. I could use a nap 14:23
kthakore aahahah ok
brb
$work
14:24 bubaflub joined
kthakore btw jnthn did we get a perl6 windows installer? 14:24
bubaflub kthakore: i saw something about a perl6 windows installer (.MSI) via twitter, i'll find a link for ya 14:27
jnthn kthakore: Yes
kthakore: Same download page as the tarball
Coke rt cli? 14:28
purl rt cli is probably www.perlfoundation.org/parrot/index...mmand_line
Coke purl, forget rt cli 14:29
purl Coke: I forgot rt cli
Coke (sadly that url is defunct)
dalek rrot: r48228 | NotFound++ | trunk/src/pmc/capture.pmc:
optimize and clarify Capture setter vtable functions
 14:31
atrodo p&w? 14:33
purl it has been said that p&w is tinyurl.com/23dfwut
Coke unshorten that? 14:39
lazy purl
14:43 bubaflub joined
jdv79 how come blizkost can't use 5.8? 14:48
i don't know if $work will ever upgrade:(
Coke jdv79: if you're worried about work upgrading, how on earth are you going to get six in there? 14:49
jdv79 well, it'll be harder:) 14:50
6 is differnt - that'd be all new stuff. 5 is a legacy/risk issue. 14:51
i was just wondering why not 5.8 14:52
moritz because it was somehow harder to do 14:53
jdv79 i could of guessed that:) thanks. 14:54
kthakore jnthn: hmm ... can't seem to find it 14:55
PerlJam 5.8 is a legacy/risk issue :)
jnthn kthakore: github.com/rakudo/star/downloads 14:56
kthakore: Look for the link to the .msi file
NotFound Being alive is a legacy/risk issue
kthakore jnthn++ 14:57
Coke jdv79: ... just install another perl5. 15:01
no reason not to have 5.8 and 5.14, if you must have 5.8
kthakore oooh! 15:02
new logo in perl6 book
jnthn: um ...
perl6 died
jnthn: "load_bytecode" couldn't find file 'P6Regex.pbc'
Coke kthakore: you may get faster help in #perl6. =-) 15:03
kthakore oh right
DOH!
Coke++ for being so deliciously awesome!
kthakore hugs Coke
Coke You sir, have a problem. =-) 15:17
15:18 mbp joined
dalek rrot: r48229 | NotFound++ | trunk/t/pmc/oplib.t:
some more OpLib tests
 15:21
15:27 bubaflub joined
cotto_work ~~~ 15:54
15:56 theory joined
mikehh opbots, names 15:58
jdv79 Coke: except all the XS modules. i'll try it.
16:07 davidfetter joined 16:26 jan joined 16:28 bubaflub joined 16:32 bubaflub_ joined
darbelo bacek: ping 16:52
16:59 patspam joined 17:11 eternaleye joined 17:13 hercynium joined 17:18 rurban_ joined
dukeleto does PIR have an equivalent to Rakudo's .WHAT ? I want the string representation of the PMC type 17:19
sorear typeof__sp 17:23
17:23 cotto_work2 joined
cotto_work2 dukeleto: typeof looks like what you want 17:23
dukeleto cotto_work2, sorear: oh yeah, forgot about that. thanks! 17:24
jnthn (.WHAT returning isn't a string...) 17:38
dalek rrot: r48230 | darbelo++ | branches/gsoc_nfg (31 files):
Sync with trunk
 17:49
18:01 chromatic joined
chromatic Did someone suggest putting GC flags in PMC headers in Lorito? I'd like to avoid that. 18:02
Coke nopaste.snit.ch/22472 18:03
cotto_work Coke++ 18:05
atrodo Coke++
sorear see: 10h 40m ago in this channel logs
cotto_work chromatic, that was just a late-night prototype to get the discussion started. 18:06
s.prototype.concept.
sorear And I already brought up the flags.
cotto_work It'd probably be a good idea to get the object model nailed down first. 18:07
18:07 tcurtis joined
chromatic Okay good. I want to avoid the current memory thrash problem our GC has. 18:07
cotto_work chromatic: do you happen to know what allison's objections to p&w were?
chromatic Nothing in specific.
atrodo i would think that ideally, GC details should be transparent to the pmc author 18:08
chromatic I know one of her concerns is a model flexible enough to allow Python style ("Hi, I don't know how object systems work!") and Perl 6 style ("Alan Kay was right!") and JavaScript ("I'm obsessed with my Self!") object systems.
atrodo, agreed.
Ultimately our GC should *not* write to memory pages strewn throughout the system just to note those objects which are alive. 18:09
kthakore Is there a *handle.pmc for sockets? 18:11
18:12 jsut joined
NotFound kthakore socket.pmc 18:17
kthakore NotFound: ok thanks 18:20
:p
dalek kudo: cf9997a | moritz++ | src/Perl6/Actions.pm:
refacor adverb application and checking into a separate method - it will be
 18:44
kudo: c9c98fa | moritz++ | t/spectest.data:
[t/spec] run new tests for attributive params
18:50 lucian joined
Coke Hey, still have failing smolder tests. 18:51
Be nice if someone fixed those
(TODO, SKIP for those platforms.)
mikehh Coke: which platforms? 18:53
purl which platforms are 8.3 an issue on, anyway?
Coke aix, win32. 18:55
mikehh don't have access to test them 18:56
Coke smolder.plusthree.com/app/public_pr..._reports/8
mikehh: don't need to run the tests. just need to check the output and insure that the failing tests are todo'd for the specific platforms.
then our smolder server will be green except for new failures.
dalek rrot: r48231 | NotFound++ | trunk (3 files):
kill Parrot_PCCINVOKE
 18:57
Coke NotFound++ # WITH FIRE
chromatic Yeah!
mikehh MSWin32/amd64 seems to be failing but not MSWin32/i386
Coke so I figured the pprof2cg.pl script is slow. Guess what I'm going to do! 18:58
mikehh: ayup.
NotFound Deleting code is the better way to improve coverage.
Coke chromatic, cotto_work : any objections if I make pprof2cg.pl use the same # in the filename as for the original pprof output? so that they are paired up? 18:59
mikehh Coke: 'k will see what I can do about that 19:00
atrodo NotFound++ # Preaching to the choir
Coke mikehh++ # I knew if I complained enough someone would try to shut me up! =-) 19:01
chromatic Makes sense. 19:02
mikehh Coke: BTW any progress on html_cleanup branch? 19:09
Coke mikehh: *hang head*
Got distracted this week.
mikehh Coke: going to look at it a bit later today/tomorrow
Coke ok, will hack on that before profiling.
ooh, even better.
NotFound =item B<new>(out STR, in INT) - Allocate a new empty string of length $2 19:11
Does this make some sense in a immutable strings world?
chromatic Can't see how. 19:12
mikehh NotFound: not to me
Coke OOOH, what if you want to leak memory? 19:13
NotFound Coke: let's create a Leak pmc
Coke that probably should have been ripped out with the bacek chagnes.
I wonder if this falls under cotto's demonstrably broken criteria. 19:14
dalek kudo: dfefdd2 | (Solomon Foster)++ | src/core/Range.pm:
Add much faster range iterator for infinite Int ranges.
kudo: c7ad36f | (Solomon Foster)++ | t/spectest.data:
Turn on new S03-operators/range-int.t test file (as stress test).
Coke I vote yes. recommend new usage: create a stringbuilder and just get a string from it when you're done.
darbelo Anyone who uses that, is doing it wrong. Let's help them make it better :) 19:15
mikehh darbelo: btw ran make splint again earlier - a few lines less - still far too many 19:19
19:22 mtk joined 19:29 mtk joined
darbelo mikehh: Yeah, I know. But you have to admit, 180+ warnings killed by a 2-line diff isn't too bad :) 19:39
What I was suggesting is tryinto identify the warings that stem from mis- or un-annotated generated code. Big bang for very few bucks. 19:41
19:44 LoganLK joined
mikehh darbelo: one of the worrying items is double compare == - ok for == 0.0 but others BAD 19:48
darbelo Hmm. I didn't see that one. What file was it in? 19:49
mikehh let me look
All tests PASS (pre/post-config, make corevm/make coretest, smoke (#35183), fulltest) at r48231 - Ubuntu 10.04 amd64 (g++) 19:52
darbelo: quite a few in src/ops/core_ops.c 19:56
Coke wonders if fixing these splint warnings does anything other than... quiet splint down. 19:57
darbelo Heh, that's generated code :) 19:58
Coke: A big part of the churn is missing ARGWHATEVER_COWSOK() in our code, which I would argue are good to have anyway. 19:59
Then there's the bugs that we aren't seeing because they hide in the churn. 20:00
splint claims that somewhere in the *big* pile of warings it's handing us are 15 actual bugs. 20:01
mikehh darbelo: src/debug.c:1954:50: Dangerous equality comparison involving double types: k == l and next line k != l (that's lc L)
Coke darbelo: seems like we could ask splint to just tell us about the errors, neh?
I mean, I like clean code and all, but I don't want to focus on the wrong end of the camel here. 20:02
chromatic onion, please
purl onion is a joke magazine, fyi. or good for health
mikehh Coke: one of the biggest mistakes ever made in the C compiler world was separating lint from the c compiler 20:03
chromatic One of the other biggest mistakes ever made in the C compiler world was not adapting to machines that aren't PDPs.
mikehh keeping are the kruft from that yes 20:04
backward compatability and all 20:05
NotFound You can always switch to freepascal ;) 20:06
mikehh or go
NotFound Go where? 20:07
purl i heard Go was gobase.org/ www.gokgs.com www.samarkand.net/ games.yahoo.com/ or called "igo" in Japanese (U56F2 U7881) members.lycos.co.uk/xuy/a2z/weiqi.htm www.dragongoserver.net or a programming language from Google golang.org or use #weiqi to speak about go on twitter
mikehh golang.org - even got built in gc and coroutinesbuilt in gc
20:08 GodFather joined
mikehh don't know how that happened 20:08
NotFound code.google.com/p/go/issues/detail?id=9
mikehh saw that, but is that really an issue :-} 20:11
NotFound I just refuse to talk about that thinh because of that. 20:12
They don't aknowledge the issue, I don't acknowledge the project. 20:13
mikehh saw an article saying that they were going to integrate into chrome as NativeClient - NaCl - maybe should be 'Salt@ 20:14
NotFound Yeah, and they can even put wave in the standard library X-)
darbelo bacek_at_work: ping 20:34
20:40 AndyA joined
mikehh darbelo: <another_bacek> this one in SIngapore on corporate laptop :) (about 10 hours ago) 20:54
20:55 whiteknight joined 20:56 macroron joined
cotto_work NotFound: is that deprecation removal properly documented? 20:59
darbelo hears chainsaws in the background. 21:01
NotFound cotto_work: long time ago 21:02
cotto_work I mean on the wiki
NotFound cotto_work: mmm... maybe unnecessary, but will be nice to add a note, yes.
21:03 particle joined
NotFound "The following functions will be deprecated in the future" What does tha mean? 21:04
trac.parrot.org/parrot/wiki/PCCMigrationNotes
I read that as "will be removed", because they are already deprecated. 21:05
cotto_work ParrotDeprecationsFor2.9
NotFound TT #443: All API function names that don't conform to this standard are deprecated and will be renamed (or removed) before 1.4 or 2.0. 21:06
cotto_work (currently doesn't exist, use ParrotDeprecationsFor2.6 as a template)
NotFound I think I'm going to delete also Parrot_pcc_invoke_sub_from_c_args, then. 21:07
cotto_work You can summarize on ParrotDeprecationsFor2.9 and link to the other page. My concern is that people looking at that page will see that something's changed and will know where to look to get enough information to update their code. 21:08
NotFound cotto_work: Were are these pages linked form?
from
cotto_work ParrotDeprecations is the main page
NotFound I sense some chaos in our wiki 21:10
cotto_work your sense is accurate
darbelo NotFound: Parrot_pcc_whatever *does* conform to the standard :)
NotFound darbelo: yes, but is grouped together with PCCINVOKE in trac.parrot.org/parrot/wiki/PCCMigrationNotes 21:11
darbelo Oh, missed that. Axe away! 21:13
NotFound Uh... Parrot_pcc_invoke_sub_from_c_args is used a lot. 21:30
21:31 donaldh_ joined
dalek tracwiki: v1 | NotFound++ | ParrotDeprecationsFor2.9 21:32
tracwiki: trac.parrot.org/parrot/wiki/ParrotD...ction=diff
tracwiki: v7 | NotFound++ | ParrotDeprecations
tracwiki: trac.parrot.org/parrot/wiki/ParrotD...ction=diff
cotto_work NotFound: thanks 21:33
NotFound cotto_work: Fear the chainsaw ;) 21:34
cotto_work I'm glad I can put it back in its case. 21:35
21:36 donaldh joined 21:37 AndyA joined
Austin seen tene? 21:39
purl tene was last seen on #parrot 1 days, 23 hours, 46 minutes and 21 seconds ago, saying: cotto_work: I've never had a problem with it... just check in and catch up occasionally. [Jul 28 21:52:54 2010]
cotto_work NotFound: is Parrot_pcc_invoke_method_from_sig_args a drop-in replacement for Parrot_PCCINVOKE?
Austin msg tene Got a wierd packfile error trying to use current nqprx. I'll try rolling back to your commit, see if that works.
purl Message for tene stored.
Austin msg Tene (^^ this is re: "our method") 21:40
purl Message for tene stored.
cotto_work i.e. is s/Parrot_PCCINVOKE/Parrot_pcc_invoke_method_from_sig_args/g sufficient to update code?
dalek izkost: 38e5ef7 | AZawawi++ | Configure.pl:
Configure.pl now displays the correct parrot_config "make" name
 21:41
izkost: e215f09 | AZawawi++ | src/pmc/p5interpreter.pmc:
Newx should be called after perl construction. This fixes the silent exit before
izkost: a221ca8 | AZawawi++ | README:
Added Debian libperl-dev installation note
NotFound cotto_work: can't prove that, don't have any use case. 21:42
Tene Austin_away: pong
NotFound cotto_work: the PCCMigrationNotes page says so.
cotto_work wfm 21:43
21:46 davidfetter joined
darbelo 'Parrot_pcc_invoke_sub_from_c_args' is exactly the same as 'Parrot_ext_call', but with a worse name, so use 'Parrot_ext_call' instead. 21:47
dalek rrot: r48232 | NotFound++ | trunk (3 files):
replace Parrot_pcc_invoke_sub_from_c_args with Parrot_ext_call in PMC usages
tracwiki: v8 | cotto++ | ParrotDeprecations 21:48
tracwiki: trac.parrot.org/parrot/wiki/ParrotD...ction=diff
21:50 AndyA joined 21:51 cognominal joined 21:56 tommyd joined 21:58 contingencyplan joined
dalek rrot: r48233 | NotFound++ | trunk/src (6 files):
replace Parrot_pcc_invoke_sub_from_c_args with Parrot_ext_call in several places
 22:03
tracwiki: v2 | cotto++ | ParrotDeprecationsFor2.9 22:05
tracwiki: elaborate a bit and add examples
tracwiki: trac.parrot.org/parrot/wiki/ParrotD...ction=diff
tracwiki: v3 | darbelo++ | ParrotDeprecationsFor2.9
tracwiki: Typo police.
tracwiki: trac.parrot.org/parrot/wiki/ParrotD...ction=diff
Austin tene ? 22:06
purl tene is, like, Stephen Weeks or a madman or blogs.gurulabs.com/stephen/
Tene 15:39 <@Austin> seen tene?
I replied before reading the rest of what you had to say. 22:07
Austin ok
Tene where are you seeing packfile errors, exactly? 22:08
cotto_work darbelo: thnaks
nopaste "Austin" at 192.168.1.3 pasted "nqprx packfile errors for Tene" (4 lines) at nopaste.snit.ch/22480
cotto_work wonders, if goto is Lorito's primary control flow primitive, does a Lorito-level call op make sense (apart from calling into C functions)? 22:09
Tene Austin: you need to realclean parrot and rebuild. There was some kind of problem with bytecode versioning around the time of the last release. 22:10
22:11 hercynium joined
Austin It's not a release built parrot - I can't get the last few released versions to build. (oom error) That said, is there a particular revision that fixes this, or what? 22:11
Tene Austin: you're trying to load bytecode somewhere that's too much older than the version of parrot you're using. 22:12
There was a bump in the bytecode compatibility version number mumble.
Austin Is that me loading the bytecode, or you?
Tene I wish it would *say* which pbc it was trying to read, though 22:13
you could probably strace it.
Austin It's not me.
cotto_work It's not you, it's me. 22:14
Austin strace: command not found 22:15
Tene Austin: Somewhere, parrot is trying to load a .pbc that was compiled with an older version of parrot, so has an incompatible bytecode version.
It's a build issue.
Austin okay, so I'll make realclean, make install, see what happens... 22:16
Tene Ah, you haven't used strace before?
Yeah, pretty much do that. 22:17
chromatic cotto_work, probably not 22:24
22:25 bacek joined 22:42 donaldh joined
dalek website: darbelo++ | The extra pointer has to go 22:42
website: www.parrot.org/content/extra-pointer-has-go
22:52 simcop2387 joined
whiteknight good evening, Parrot 23:00
23:00 ruoso joined
cotto_work hio, whiteknight 23:02
sorear NotFound: I personally recommend *against* writing Parrot 3 in Free Pascal 23:04
whiteknight helo cotto_work 23:05
darbelo But pascal is, like, totally better than C! 23:06
Nothing can go wrong if you write pascal. 23:07
23:08 simcop238 joined 23:09 kid51 joined 23:15 dduncan joined 23:20 dduncan left
NotFound sorear: What about Modula or Ada? ;) 23:21
23:22 snarkyboojum joined
chromatic ML would be okay. 23:22
sorear darbelo: Try feeding (x > y) ? (x + y / 2) : (y + x / 2) to both gcc and fpc with the optimizer set to max in both cases 23:23
gcc generates about 8 instructions, using integer shifts and cmov
fpc generates about 50 instructions, including FPU instructions to do the division 23:24
NotFound: Dunno, I've never had occasion to trawl through Modula or Ada compiler output
darbelo sorear: Fine, fine. Wel'' use the gcc pascal frontend. Geez. 23:25
NotFound I've a half written pascal compiler that outputs C++ 23:26
chromatic You should see someone about that. 23:27
NotFound kimba.euler.es/ -- In spanish 23:28
purl In spanish is, like, nicer :P
darbelo msg japhb I looked at moving plumage's Configure.nqp to use nqp-setting.nqp but the subst() in that lib doesn't handle a sub as 'replacement', which kinda complicates things. 23:42
purl Message for japhb stored.
darbelo msg japhb Should I go with that or is it better to try convince pmichaud to augment the nqp-setting subst() ? 23:43
purl Message for japhb stored.
darbelo goes afk for while.
japhb msg darbelo IMHO the nqp setting should support sub replacement for subst(). It's easy and makes a MASSIVE difference to utility. 23:50
purl Message for darbelo stored.
23:51 cognominal joined 23:58 Psyche^ 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!