Parrot 1.1.0 Released | parrot.org/ | 325 RTs left | Weekly Priority: Remove Deprecated Items
Set by moderator on 28 April 2009.
00:23 dmknopp joined, dmknopp left 00:35 bacek_ joined 00:51 dalek joined
dalek rrot: r38456 | Infinoid++ | trunk/CREDITS:
[CREDITS] Add commit-karma aliases for fperrad++ and moritz++.
 00:56
01:36 bacek joined
Infinoid What a beautiful day. Sadly I didn't have much time for parrot this weekend, I was hoping to get at least one thing from DEPRECATED done 02:02
02:19 keir joined 02:35 janus joined 02:36 keir left
darbelo ping cotto 02:41
purl I can't find cotto in the DNS.
cotto darbelo, pong! 03:31
darbelo I'm going to thrown together some prototypes for the different context policies that we discussed on the mailing list and had a few questions. 03:32
cotto cool 03:35
darbelo I figured it was the best way to decide. The code on svn follows per-PMC semantics, and some of the questions I have are related to implementing per-type semantics. 03:37
First, "$P0 = $P1 + $P2" (all DecNums) will dispatch to MULTI add(DecNum value /*P2*/, PMC* dest /*P0*/) right? 03:40
cotto I think so, but mmd is a little broken. 03:42
darbelo The thing is I have no type information for the destination. Most code unconditionally calls pmc_new(), but what if there is information I want to preserve on the destination PMC? 03:44
Stated another way: Is there a safe way to know if I got a destination of my own type? 03:47
03:50 petdance joined
darbelo I think "vtable->base_type" has the information I want, but I wanted to sanity-check the idea. 03:51
cotto that sounds like it'd work 03:54
you can also use vtbale_isa
darbelo Ok, thanks. That covers per-PMC for now. 03:58
For per-type I was thinking of a singleton PMC containig the context. 03:59
cotto yeah, that's what I was thinking too for that approach
04:01 Theory joined
darbelo I'm not very familiar with singleton PMCs. Should I use something like "ATTR PMC *ctx" or is there a better way? 04:03
cotto darbelo, it's preferable to use isa. istr that numerical type ids will eventually go away (although not soon) 04:04
That's the easiest way.
darbelo isa it is, then. Thanks. 04:05
cotto They'll all point to the same pmc, but that's fine.
darbelo I'll get that same pointer on every call to pmc_new(), from what I saw on pmc.c 04:07
cotto src/pmc/os.pmc and src/pmc/random.pmc are both singletons 04:08
if you need example code (there are also several others)
darbelo It's something I saw on random that prompted the next question. What is the best place for initialization? 04:12
random uses the init VTABLE to clear itself. That means every call to pmc_new() erases all changes. I can use pmc_new_no_init() to avoid that call internally. 04:14
but then "$P1 = new 'DecNumContext'" will undo all changes to the operating context. Is that something we want? 04:16
cotto Hmmm. I'll need to look into that, but singletons may have a defense against that.
If nothing else, you could have some logic in your init VTABLE function that dtrt. 04:17
bacek_ O HI 04:20
darbelo HAI bacek_ 04:21
cotto hi bacek_ 04:22
I've had a busy day and am just sitting down to hack on tt528_vtinit. 04:23
Thanks for the changes, btw.
bacek_++
darbelo Semi-unrelated question: Currently pmc2c won't allow unions (or even union *) as ATTRs, is that intentional or accidental? 04:25
cotto good #ps question
bacek_ cotto: you'll hit another problem with SUPER rewriting if you'll remove PARROT_EXPORT from generated files. I didn't manage to fix it. 04:36
cotto Oh. How fun.
Infinoid ATTR type compatibility is pretty horrid, but we're working on it 04:37
Whiteknight++ recently added special parsing so it could handle function pointers (r38353)
so it's accidental, not intentional. 04:38
cotto It'd be nice if a PMC could specify headers to be included in pmc_foo.h so that we don't need silly workarounds for non-built-in types. 04:39
Infinoid Yeah, you could use that to allow local typedefs to shine through 04:40
I hadn't thought of actual header.h headers before though 04:41
I was just thinking of some generic hackish sort of <pre> block
but that way lies madness (also known as XS).
cotto What about including anything that's #included before the pmclass line? 04:43
madness?
purl Welcome to the House of Fun!
Infinoid hmm, that seems reasonable (searching for #include lines)
one of the problems I ran into which #544 didn't address was, I had done "typedef void (*funcptr)(arg1, arg2)" but wasn't able to use "ATTR funcptr foo" because it ended up in the header without the typedef 04:44
sticking it in another header file would solve that, but it didn't occur to me at the time. 04:45
04:49 petdance joined
Infinoid And yes, madness. :) 04:49
Infinoid suffers from post-XS stress disorder, and thinks pmc2c's syntax is worlds better. 04:50
darbelo cotto: This reminds me, the "ATTR struct decContext *" warnings about incompatible pointer types can be solved with some very trivial patches to the library, so we're good to go for all non-union typedefs (ie. the ones we care about). 04:52
cotto I recommend that you keep a copy of the patch to decNumber in svn so it's easy to upgrade, and you probably should make more than trivial changes to the library, but that sounds good. 04:54
s/more/no more/
darbelo It's just a matetr of changing "typedef struct {/* ... */} decWhatever;" to "typedef struct decWhatever {/* ... */} decWhatever;" 04:56
cotto That definitely qualifies as trivial. 04:57
darbelo If requiring 'patch' to build is kosher I'll keep the patches in a separate dir and apply them as part of the build process.
cotto bacek_, does Parrot build with your patch?
s/patch/latest commit to tt528_vtinit/ 04:58
darbelo, I don't think that's necessary.
bacek_ cotto: yes.
cotto Hmm. Then it's my fault. 04:59
#1 0xb7bd2f5b in parrot_hash_put (interp=0x804a040, hash=0x806dbf8, key=0xc244489, value=0x0) at src/hash.c:1293
What's wrong with this picture.
s/./?/
bacek_ probably value=0x0 05:00
cotto hint: key
It's supposed to be a pointer.
darbelo Okay, I'm going to sleep now, I have to get up early tomorr^W today. I'll start commiting stiff somtime in the afternoon. 05:04
cotto night 05:05
darbelo Night, everyone.
cotto msg Infinoid does dalek follow Google Code hosted projects?
purl Message for infinoid stored.
05:06 darbelo left
Infinoid cotto: Yes, if they're listed on the Languages page on the trac wiki 05:06
cotto Hmmm. darbelo's project doesn't really qualify as a language.
Infinoid Ok, I'll need to add a specific hook for it then
I'm trying to make it autoconfigured as much as possible so I'm not a bottleneck, but we can add additional stuff as needed 05:07
cotto code.google.com/feeds/p/decnum-dynp...nges/basic
Infinoid its a gsoc project, right? (just so I get the POD right) 05:09
cotto yup
Infinoid++ 05:11
bacek_ cotto: looks like I didn't commit everything from my home computer. I;ve got a lot of warnings on clean checkout... 05:13
05:13 dalek joined
Infinoid cotto: github.com/Infinoid/dalek-plugins/b...cnumlog.pm 05:13
Lemme know if it actually works.
cotto Infinoid, will do. 05:16
bacek_ cotto: hooray. Try "rm src/pmc/*.[ch]" before "make" 05:27
cotto so make clean/realclean is broken 05:35
nm. clean/realclean should catch those 05:36
05:39 davidfetter joined
bacek_ cotto: actually pmc2c is broken. 05:40
cotto This is me not being surprised. 05:41
Infinoid Wow, that was convincing. :) 05:42
goodnight all
cotto Night Infinoid. 05:44
dalek rrot: r38457 | petdance++ | trunk/src/debug.c:
removed unused variable, and worked more towards single exit from get_cmd()
 05:45
05:48 szabgab joined
dalek rrot: r38458 | petdance++ | trunk/src/packfile.c:
removed unused variable filename
 05:58
06:01 iblechbot joined 06:04 uniejo joined
cotto bacek_, any changes incoming? 06:04
bacek_ cotto: not now... 06:08
cotto ok. I need to test, but I'll have some soon. 06:09
06:13 szabgab joined 06:52 mikehh joined
cotto bacek_, did you run make test 07:13
bacek_ cotto: I did it at home. 07:14
Running now on current box
It was 3 failing tests in t/tools/pmc2c AFAIK 07:15
AFAIR
cotto what about t/pmc/ro.t? 07:18
bacek_ ------------------- 07:20
t/src/extend (Wstat: 768 Tests: 17 Failed: 3)
Failed tests: 12-14
Non-zero exit status: 3
t/tools/pmc2c (Wstat: 768 Tests: 11 Failed: 3)
Failed tests: 3, 6-7
Non-zero exit status: 3
Failures in t/src/extnend looks really bad... 07:22
cotto The makefile dependencies are screwy. 07:35
For me, make test doesn't even get past t/op/annotate.t 07:36
It may be a while before this branch can be merged. 07:40
dalek kudo: r3c95282 | pmichaud++ | src/parser/actions.pm:
Refactor handling of nodes to !add_param, fixing implicit slurpy array.
 07:45
kudo: r020fd49 | pmichaud++ | src/setting/Any-list.pm:
Make sure .join flattens lists.
kudo: r0588e8c | pmichaud++ | src/setting/Any-list.pm:
Change .join flattening a bit.
purl dalek: that doesn't look right
cotto This code is too delicate. 07:46
Gah. If I just run make test after Configure.pl, the annotate test passes. 07:50
h8
Now I'm not seeing any test failures, except pmc2c. 07:57
and now it works. This bird is insane. 08:03
bacek_, I'm not seeing any reliable test failures except pmc2c. 08:09
bacek_ cotto: it's kinda weird... 08:20
cotto bacek, is line 833 a bug? 08:23
848
The code uses $extra_get_vtable and $get_vtable
It seems to work, but it looks suspicious. 08:24
08:27 clunker3 joined
cotto PMCEmitter.pm 08:27
bacek_? 08:28
bacek_ cotto: yeah. Classical copy-past bug 08:30
cotto The confusing thing is that it works.
bacek_ No-one check _ro_ variant...
cotto t/pmc/ro.t does, I think. 08:31
bacek_ Inherited _ro_ variant
08:31 riffraff joined
cotto and the build segfaults when I "fix" it. 08:34
bacek_ build segfaults???
cotto segfault during the build 08:35
bacek_ ouch
cotto OK. I see why it worked.
The ro variant needs to be added on top of the normal vtable.
which the code did by accident 08:36
bacek_ yes. We only have to build full _ro_ vtable for "default".
But default is special anyway
cotto There's no need for a ro default 08:37
it doesn't make sense since default isn't a real pmc anyway
bacek_ It is. All denerated _ro_ functions are same
trac.parrot.org/parrot/ticket/631 08:38
Item 2. Proposal 3
cotto That makes sense. 08:42
I'd like to clean up the code some before making any more functional changes, though. 08:44
bacek_ That makes sense :)
dalek rrot: r38459 | cotto++ | branches/tt528_vtinit/lib/Parrot/Pmc2c/PMCEmitter.pm:
[pmc2c] make accidentally correct code purposely correct, improve readability of generated code
 09:03
cotto bacek_, I hope you don't mind me rewriting your chained updates.
bacek_ cotto: of cause not 09:14
afk # decommute to home from office 09:15
cotto don't crash 09:16
(just in case you were so inclined)
09:33 bsdz joined
dalek rrot: r38460 | cotto++ | branches/tt528_vtinit/lib/Parrot/Pmc2c/PMCEmitter.pm:
[pmc2c] various minor formatting fixes and cleanups
 09:58
bacek hi again 10:08
purl oh, you're back!
bacek hm.. Less than one hour to drive home. Good result. And I didn't even crush :)
10:11 iblechbot joined
cotto Welcome back. I'm glad that you're not dead or injured. 10:11
(and if you are, I'm throughly impressed that you can still use irc) 10:12
bacek I'm not dead. Yet. 10:15
Need help with tt528 branch? 10:16
cotto Yeah. Implementing default ro vtable functions would be a help. 10:17
bacek Ok. I can do it. May be even today. 10:18
cotto bacek++
bacek not yet :) 10:19
cotto You get karma for taking a task, for working on it and for completing it. ;) 10:21
SUPER may prove interesting. 10:22
At least DYNSUPER isn't used. 10:23
bacek SUPER should be replace with "interp->vtables[enum_class_${super}]->$name(" 10:24
or something similar
cotto yeah
bacek s/replace/replaced/ 10:25
cotto actually, DYNSUPER appears to do just that,
.
bacek lemmecheck 10:26
O! Indeed.
cotto I'm checking what happens when SUPER does the same thing as DYNSUPER. If all tests pass, this'll be the easiest change ever. 10:27
bacek .SUPER should follow similar conventions
cotto .SUPER doesn't appear to be used. 10:29
bacek grepping
purl grepping the source is good for the soul
moritz indeed 10:30
bacek purl: good girl
purl thanks bacek :)
cotto botsnack
purl :)
bacek ok. .SUPER isn't used ATM. But we pretend to support multiple inheritance. So I vote to keep it and fix it 10:32
cotto Yes.
bsdz bacek, cotto: guys, just wondering if the work you're doing might fix vtable inheritance. really cool if it does. 10:34
bacek bsdz: ticket number? 10:35
bsdz one mo
cotto bsdz, what's broken that you want fixed?
If it's too long for irc, please file a tt and one or both of us will look at it. 10:36
bsdz i've created a ticket. basically if a I derive from a base PMC type like Integer. I can't use vtable methods like add/sub/... on my derived class
ticket is #562. should be renamed really. 10:37
cotto I think that's a mmd bug.
I've seen something similar.
bsdz I also wrote a newsgroup message with some investigation, groups.google.co.uk/group/parrot-de...91cf5ac3d#
jonathan I think it's because Parrot's MMD doesn't look up the inheritance heirarchy.
To find alternative candidates up there. 10:38
Or at least, it could be that.
IIRC, we had to implement doing that ourselves in Rakudo.
bsdz it's also a blocker for the matrixy hll.
cotto That needs to be fixed. It's the reason I stopped converting MULTIs to switching VTABLEs to try to speed up Parrot. 10:39
but of course, me saying that doesn't accomplish much 10:40
bsdz yes had considered following Rakudo's hard path but takes time away from doing some nice math programming :)
cotto bacek, making SUPER act like DYNSUPER causes the build to segfault when compiling config_lib.pasm. No super-easy fix for me. :( 10:41
Oh. It's trying to use default's freeze. 10:42
nm. That's not the bug. svn head calls Parrot_default_freeze too. 10:43
bacek try "rm src/pmc/*.[hc]" 10:44
It usually helps me a lot
cotto It's reproducible after make clean. 10:45
bacek even realclean? 10:46
cotto Of course! interp->vtables[default] == NULL 10:47
bacek Got segfault there as well
Oh shi...
cotto So it'll work if that vtable gets populated. 10:48
bacek Yes. We need class_init for default 10:49
For consistency sake
cotto but Parrot doesn't usually care about such a function. It'll need some other changes too.
stupid special cases 10:50
bacek So, let's kill this special case :) 10:53
cotto +a lot
although it'll still need the default ro vtable functions
bacek I'll make them 10:54
cotto Thanks. It's just about bedtime for me. 10:55
I don't have any pending changes, so go crazy. 11:06
bacek is crazy Russian :)
11:11 eternaleye joined
cotto is sleeping American 11:11
night
bacek night 11:26
12:06 ruoso joined 12:28 masak joined 12:31 iblechbot joined
dalek rtcl: r341 | wcoleda++ | trunk/README:
let people know what to expect.
 12:31
Coke msg Infinoid 08:31 <+dalek> partcl: r341 | wcoleda++ | trunk/README: 12:33
purl Message for infinoid stored.
dalek rrot: r38461 | NotFound++ | trunk/src/pmc/orderedhash.pmc:
[core] fix off-by-one in orderedhash TT #636 skids++
 12:47
Infinoid hmm 12:52
13:04 gryphon joined
Infinoid Coke: Thanks for testing. Dunno why that didn't work, seems to work fine in "make test". I made it a little more resistant to http timeouts just in case. 13:06
The bot needs better diagnostics, for some reason the logs stopped being useful a couple of months ago. I'll work on that next.
13:06 dalek joined
bacek purl: msg cotto I broke rt528 branch little bit. Have to figure out why find_method treated so specially in RO.pm 13:07
purl Message for cotto stored.
bacek And my commit message broke dalek again...
afk # sleep 13:08
Infinoid I restarted dalek to enact a code change
bacek++
bacek hey. What for? :) 13:09
Coke hurm. As someone on the grant committee, I would if someone who is not me could check out the parrot item listed news.perlfoundation.org/2009/05/200....html#more and here: news.perlfoundation.org/2009/05/200....html#more 13:10
(I am not entirely sure that is the case) 13:11
Infinoid Coke, you would *what*? 13:12
Coke er, like, appreciate, wonder?
Infinoid I think the "Benefits to the Perl Community" section could use some additional fleshing out. after all, we already have webmails and CMSes 13:16
so I'm not really sure what value they add
PerlJam concurs.
Coke Yah, feel free to comment on the proposals in general, obviously.
PerlJam both of those leave me with a kind of "huh?" feeling. 13:17
Coke there are 4 proposals this round.
news.perlfoundation.org/2009/05/200...osals.html
dalek rrot: r38466 | fperrad++ | trunk/tools/dev/fetch_languages.pl:
[languages] add XML
 13:29
13:32 gryphon_ joined 13:59 davidfetter joined 14:02 gryphon joined
Infinoid posted his comments (waiting for mod approval) 14:02
14:04 ilia joined 14:09 fperrad joined 14:36 particle joined 14:41 NotFound joined
NotFound hi 14:41
purl niihau, NotFound.
jonathan is pmc2c now copying attributes of a Parent class into child ones, or does that sitll need to be done manually when subclassing a PMC? 14:42
14:42 bsdz joined 14:50 Andy joined 14:56 rg1 joined 15:04 hudnix joined 15:15 Theory joined
Infinoid jonathan: I think it copies them now 15:24
at least, it does for UnManagedStruct/ManagedStruct 15:26
jonathan Infinoid: Yeah, I experimented a bit and it does.
15:32 davidfetter joined 16:01 japhb joined 16:04 barney joined, rdice joined 16:16 iblechbot joined 16:23 workbench joined 16:26 bsdz joined 16:36 cosimo joined, cosimo left 16:49 jsut joined 17:04 Whiteknight joined 17:08 fperrad joined 17:10 mst joined, mst left
Coke hurm? 17:38
purl Yes, Rorschach.
Coke geeks, "happy jedi day." 17:39
szbalint may the source be with you? 17:53
Coke Close. what's today's date? 17:54
particle quatro de mayo 17:56
szbalint mayo++ 17:57
davidfetter mmm...mayo :d 18:02
NotFound ..nesa 18:03
davidfetter made some of that just the other night 18:04
i'm sure glad i trust my egg source
particle opbots: trust chickens 18:05
clunker3 But I do not trust you particle
slavorg Ok
cotto hi 18:23
particle dumb bots. 18:24
hi cotto 18:25
davidfetter purl, have you ever retired a human by mistake? 18:28
purl i haven't a clue, davidfetter
rg1 purl, have you ever retired a human by mistake is <reply>I'm not a bladerunner 18:30
purl OK, rg1.
davidfetter purl, have you ever retired a human by mistake?
purl no idea, davidfetter
davidfetter hrm 18:31
szbalint botfail
rg1 ok, i guess that didn't work
davidfetter i guess libdwim isn't linked in
you didn't put a ? at the end
rg1 no, that's not it. not sure what i'm doing wrong, though. 18:34
rg ha! got my nick back.
cotto bacek++
davidfetter cotto, what'd he do this time? 18:35
cotto more awesomeness in the tt528_vtinit branch
It should be ready for testing pretty soon.
davidfetter um
cotto Can someone test the tt528_vtinit branch on windows after this next commit? 18:51
The fixes are committed. Any volunteers to test? 18:54
18:55 PacoLinux joined
dalek rrot: r38467 | cotto++ | branches/tt528_vtinit/lib/Parrot/Pmc2c/PMC (2 files):
[pmc2c] export Parrot_default_get_vtable, don't call Parrot_default_find_method directly
 18:56
18:56 bsdz joined
Infinoid cotto: I can test that on mingw 18:58
cotto Infinoid, thanks. 19:00
dalek kudo: r2e5607d | pmichaud++ | docs/spectest-progress.csv:
spectest-progress.csv update: 379 files, 10999 passing, 11 failing
 19:03
particle smokes vtinit with msvc-x86
19:04 darbelo joined
Infinoid This old mingw machine is still making a parrot checkout. particle++ will probably win this race :) 19:05
cotto sounds like a speed demon 19:06
Infinoid but it says "Athlon XP"! That means windows xp should run fast on it, right?
darbelo Faster than Vista, yes. 19:07
Infinoid I've got a special patch to make vista crash 5 times faster than xp
cotto I remember when I had an Athlon. 19:08
Infinoid I don't use this box much. Probably why it still has windows on it
Coke I've got a test going against strawberry. 19:11
smolder failed to post, but only t/tools/pmc2c.t failed. 19:12
cotto Great!
although I'm not particularly looking forward to working on pmc2c.t 19:13
I don't see the point of testing code that's only used one way for a particular flavor of brokenness
particle .\\src\\pmc\\default.c(2381) : error C2375: 'Parrot_default_get_vtable' : redefinition; different linkage 19:14
c:\\dev\\parrot\\vtinit\\src\\pmc\\pmc_default.h(263) : see declaration of 'Parrot_default_get_vtable'
Coke cotto: if it were me, I would ditch and let the cage cleaners fix it. 19:15
but I suck that way.
Infinoid cotto: Is this branch related to the pmc_pct stuff? 19:16
19:16 ilia joined
cotto ok. The header's goofy. 19:16
Infinoid, yes. Once it's working, it'll greatly simplify what pmc_pct needs to do.
The main idea is to build vtables at runtime instead of having pmc2c (or the pmc compiler) figure out which functions need to be used in which slots. 19:17
Infinoid rtfTT's 19:19
cotto TT #631 is relevant too 19:23
darbelo cotto: is dalek following decnum-dynpmc commits now? 19:25
19:25 allison joined
cotto darbelo, he should be 19:27
darbelo It hasn't reached RSS yet, but there is a new commit. So we'll find out soon then. 19:29
cotto I think dalek polls its rss feeds every 3 or 6 minutes. You should know by then. 19:30
Infinoid dalek polls its feeds about once every 5 minutes
the timings are perturbed a little bit to try to spread the traffic out
cotto good policy 19:31
Infinoid (and make the bot less likely to ping out)
darbelo Right now the delay is on google code. Commit went in, RSS unchanged. 19:32
RSS update. OMG I'm so ANXIOUS!!11 19:35
Infinoid darbelo++ # preemptive karma
Coke karma coke?
purl coke has karma of 2740
Coke karma wcoleda?
purl wcoleda has karma of 17
Coke wcoleda-- 19:36
wcoleda?
purl hmmm... wcoleda is Coke
Coke karma wcoleda?
purl wcoleda has neutral karma
cotto I sense some upcoming polyglotbot abuse
Coke purl?
purl yes, Coke?
rg cotto: there is no polyglotbit 19:37
*bot
Coke nah, I don't need to transfer the karma, just killing the alias.
cotto ok
Coke would have done it privately if I could have. =-)
Infinoid Ok, I think it's safe to say dalek is broken. This stuff would be a lot easier to debug if it had a working logfile; I'll fix that first 19:40
Coke wonders if anyone has any clues about TT #627
Infinoid 3*11*19 19:42
purl 627
darbelo 7 * 11 * 13 * 13 19:44
purl 13013
Coke 6 * 9?
purl it has been said that 6 * 9 is 42
Coke smart purl.
darbelo 42? 19:45
purl 42 is too great number for me :-)
Infinoid 6*9
purl 54
Infinoid you know, that's always bugged me.
rg purl: who said 6 * 9 is 42? 19:46
purl i haven't a clue, rg
rg typical :P
Infinoid (Hitchhikers Guide)++
rg thought they said 6*7
davidfetter 6*7 == 42base13 19:48
dalek nie: r61 | isop44++ | trunk/ (8 files):
Add Pynie;list class
nie: r62 | isop44++ | trunk/examples/ex-3-1-4-a.py:
Now we can print() a list, make example actually use this
Coke the base13 thing is not... canon. 19:55
dalek nie: r63 | isop44++ | trunk/TODO:
Remove TODO (contents are no longer relevant)
 19:58
Infinoid neither 20:03
dalek kudo: r1c0ac16 | pmichaud++ | perl6.pir:
Convert several internal Lists to Arrays.
 20:05
rrot: r38468 | cotto++ | branches/tt528_vtinit/lib/Parrot/Pmc2c/PMCEmitter.pm:
[pmc2c] make declarations of get_vtable and update_vtable PARROT_EXPORT
 20:07
cotto particle, can you re-test the tt528_vtinit branch with this latest commit
cotto also, which compiler are you using? 20:08
particle retests vtinit with msvc2008exp-x86 20:10
NotFound Coke: TT #627 may be related with linking against a static and a dynamic parrot lib at the same time. It has happened before.
Or even, linking with two different dynamic ones at the same time 20:12
Coke everything on OSX is dynamic. 20:14
(because the config currently hardcodes it so.)
s/config/platform hints/
nopaste "coke" at 72.228.52.192 pasted "faux" (11 lines) at nopaste.snit.ch/16471 20:15
Coke er, that was for you. 20:16
NotFound And dynlexpad links with the same libparrot? 20:18
Coke if I could find where dynlexpad was installed... 20:20
NotFound That will be a good first step ;)
nopaste "Coke" at 72.228.52.192 pasted "you were right." (11 lines) at nopaste.snit.ch/16472 20:21
Coke so the build process is borked, it would seem. 20:22
NotFound Now you know why I dislike the rpath thing ;) 20:23
Coke The macport installer rewrites the link deps on install; will have to add these to the list of rewrites in lieu of a real fix. 20:25
NotFound At least, to check if this is the real culprit of the problem. 20:26
Coke dammit. I just hit ^W on a trac page and it closed the window. 20:27
cotto I set tabmix plus to confirm closing multiple tabs for exactly that reason 20:30
nm. That's q, not w
although "undo close tab" might work 20:31
Coke chrome doesn't have "undo close tab"
be nice if there was JS on the trac page to say "you've edited data here, are you sure?"
Infinoid if js could prevent me from closing a window, spammers would abuse it immediately 20:32
NotFound Aren't the Macs supposed to be the most in user friendliness? ;) 20:33
Coke Infinoid: it can. 20:34
(having written code to catch the window close event for work apps in the past.) 20:35
NotFound: furthermore, the reason tcl_group.bundle isn't loading for partcl is... because it's getting the wrong path from the installed parrot.
(and linking against the non-existant build directory)
NotFound++ 20:36
cotto particle, did the build work? 20:40
20:46 PacoLinux joined
particle cotto: yes, failed 3 pmc2c tests 20:59
smolder upload failed, just like coke's 21:00
cotto particle++
rg particle: i think yours actually worked, only reported an error like it did for the last 2 weeks: smolder.plusthree.com/app/public_pr...ails/20585 21:03
particle seems rg++ is correct. however... without the branch name in the test report, tracking smolder reports is problematic. 21:05
Coke we can control that on our end, no? 21:06
rg i was wondering about that, and it should have been included. i guess there's a bug
Coke (just use 'svn info' to get the branch name and throw it in the list of variables.) 21:07
particle: did you save the error message you got?
particle it was a 500. i closed the console, though :(
rg particle: is it possible you don't have svn in your path? 21:08
particle don't know how i'd 'svn up' without it.
rg (or don't even have a svn command?)
maybe use tortoise svn or something ;) 21:09
Coke just not sending the branch in the absense of an svn command line util would be fine for now.
rg coke: see lib/Parrot/Harness/Smoke.pm line 170 21:10
Tene maybe instead of 'svn up' you accidentally typod "echo 'ERROR 500'"
i mean, I do that all the time.
Coke rg: ah. 21:11
rg so it *should* have been there
21:12 Tene joined
Tene :) 21:12
particle :) 21:13
cotto wb
rg uhm ... what's the name of the branch again?
cotto tt528_vtinit
rg that should have worked. it works for trunk, see e.g. one of my reports: smolder.plusthree.com/app/public_pr...ails/20582 21:14
particle btw 'make smoke' is rebuilding parrot for me, even though i just submitted a smoke report, and haven't 'svn up'ped
cotto make has been behaving like that for me too, although for different targets 21:15
rg hmm another smolder bug: the new reports don't have tags. maybe that's where the server error comes from 21:17
21:17 PacoLinux joined
Coke rg: Can you make sure we have tickets open that are tagged "smolder" ? 21:17
then we have a list we can give to mpeters.
rg none applicable. just one rfe 21:20
i guess i'll write one 21:21
Coke danke.
21:22 luca joined 21:28 Whiteknight joined
Whiteknight Who do I talk to about getting my blog added to Planet Parrot? 21:30
darbelo Cripes! Blog. I forgot I had to get one. 21:31
Whiteknight darbelo: Yes that would be good. I definitely want to post lots of links to you and your work!
by the way, I've been watching your commit logs and you're doing a great job so far
cotto darbelo, a question about using parrot.org (or parrotblog.org) is on my list for #ps tomorrow 21:32
If you want to write something, you can just stick it in a local file until you get a proper blog. 21:33
Tene I need to get planet parrot to be following the 'parrot' category of my blog instead of the whole thing.
darbelo Whiteknight: Slower than I'd like to go, but it's moving forward. 21:34
cotto: Isn't parrotblog.org an aggregator? 21:35
Whiteknight Tene: That's what I want to do for my blog too, just add the "Parrot" articles
cotto no idea
Whiteknight darbelo: No, parrotblog.org is just a regular blog
it's hosted at blogger
planet.parrotcode.org is the aggregator
allison: ping 21:39
Tene Whiteknight: that's what I wanted too, but I apparently gave the wrong url
allison Whiteknight: pong 21:40
bacek good morning 21:41
cotto hi bacek
bacek hi cotto
allison darbelo: parrotblog.org could change to an alias for the aggregator
bacek cotto: tt528 looks pretty good! 21:42
cotto The branch is almost ready, but it needs some tweaks to allow rakudo to build against it.
It's the ro_find_method function, which might be the same issue you were referring to. 21:43
bacek cotto: I'll check it.
cotto enum_class_Foo doesn't work for HLL PMCs, so something else needs to happen. 21:44
Thanks.
Whiteknight allison: who do we talk to to add new blogs to planet.parrot.org?
planet.parrotcode.org
purl planet.parrotcode.org is the aggregator
cotto bacek, the build does work with msvc 21:45
bacek cotto: erm. I don't have Win installation... 21:46
btw, t/tools/pmc2c failed because we start generating different code
cotto yeah. I'm not worrying about it for now. 21:47
allison Whiteknight: I don't know who set it up
Whiteknight allison: Okay, thanks anyway.
cotto I don't like that the test test that we're generating a certain kind of code rather than if the code works. 21:48
particle thinks coke was involved
dalek nie: r64 | isop44++ | wiki (2 files):
Created wiki page through web user interface.
allison Whiteknight: the planet.parrotcode.org domain name is an alias for x3.develooper.com, which means it's running on the Perl.org servers 21:49
Whiteknight right, that's where the "contact webmaster" link took me
are there any plans to host an aggregator at parrot.org?
allison Whiteknight: Ask and Robert may be the only people with access to add blogs
Whiteknight okay, thanks
I'll dig into i
it*
allison Whiteknight: mmm... bump that from "maybe" to "almost certainly" 21:50
particle so email admin@perl.org i suppose
allison Whiteknight: it says in the sidebar "maintained by Ask and Robert"
particle: yes, sensible 21:51
purl sensible is usually inappropriate, yes :)
particle webmaster@ is probably better...
Whiteknight urg, I must be retarted for missing that
Infinoid are there any plans to host an aggregator at parrot.org? 21:53
drupal does have a news aggregator plugin.
cotto allison, are there any plans to use parrot.org as a blog?
Whiteknight that would be nice, and would further our unification efforts
Tene oright, I'm in oregon next week. need to harass chromatic... 21:54
Whiteknight Tene: Which blog is yours? If I'm talking to the blog admins I can get your changes made as well 21:59
allison cotto: to the extent that anyone can post stories on parrot.org, it is a blog
Tene Whiteknight: blogs.gurulabs.com/stephen/
should be just blogs.gurulabs.com/stephen/parrot/
allison cotto: we haven't enabled the "blog" feature, which AFAIK gives each user their own page of posts made by them
Tene wait, that shoudl have a feed
lemme look... 22:00
cotto allison, should darbelo use that or just set up something on blogger (or wherever)
Tene blogs.gurulabs.com/cgi-bin/mt-searc...p;limit=20
allison cotto: happy to turn on "blog" if it encourages getting parrot happenings out
particle allison: i'd like to turn on blog only for known humans
cotto Ok. Thanks.
allison cotto: is he making status reports? 22:01
particle: yes, we can limit it to a specific permission group
particle there are thousands of users registered for parrot.org
Whiteknight Tene: Okay
cotto allison, he will be soon.
Whiteknight particle: Thousands? where the hell are they all? 22:02
allison particle: yes, but they don't automatically get permission to do anything
cotto non-spam users?
particle BEWARE THE BOT INVASION
allison: great, that's the way i'd like to keep it
allison particle: well, I guess they do, they get permission to post to the "Scratch" section (kind of like a wiki) 22:03
particle: but we can disable that
hmmm... do we want people to be able to post to the blog using XML-RPC? 22:04
cotto darbelo, it looks like you'll soon have a place to blog 22:05
allison there's a module for that too, installed by default, but enabled separately
bacek cotto: r38469. Rakudo successfully built on my box 22:07
cotto allison, when you enable blogs, could you also let darbelo post?
Whiteknight allison: No XML-RPC
that's my two ces anywy
allison okay, blogs enabled
Whiteknight w00t. allison++ 22:08
cotto bacek++
allison darbelo: is your www.parrot.org account name darbelo?
darbelo Yup.
dalek rrot: r38469 | bacek++ | branches/tt528_vtinit/lib/Parrot/Pmc2c/PMC/RO.pm:
Use pmc_type(Foo) instead of direct enum_class_Foo construction in RO.find_method handling.
darbelo Wait. is that the same as trac?
nie: r65 | blairuk++ | trunk/setup.py:
Allow setup.py to be run using Python 3.0; subunit won't work for now though :(
cotto darbelo, the front page is separate from trac
bacek cotto: ok. Branch is ready for review/merge. I can fix ro vtable handling in tt631 branch after merge 22:10
cotto I'm building/installing to verify... 22:11
allison darbelo: separate login from trac 22:12
darbelo Then no.
cotto bacek, I'll merge once I know if all parrot and rakudo tests pass. 22:13
bacek cotto: I'm running rakudo's spectest now
rg cotto: can you also run the t/spec/S09-subscript_slice/slice.rakudo spectest? that one currently dumps core for me on linux and freebsd. 22:14
jonathan It dumps core for me on Windows too.
It's been doing so for some days now.
rg ic
jonathan Didn't get chance to find a backtrace yet.
oh, or did I... 22:15
cotto I can test it, but it sounds like it shouldn't change whether I merge the branch.
bacek cotto: You'll probably have to merge trunk into branch before testing rakudo.
darbelo allison: It is now :) 22:16
allison darbelo: if you create an account on www.parrot.org, I can grant you posting access
rg cotto: i basically wanted to know if anyone else is seeing that, so jonathan++ already answered that. ooc you could still run the test to see if you accidentally fixed it ;)
darbelo allison: Just did it. 'darbelo'
cotto bacek, do you mean that the fix for the slice test is in trunk?
bacek cotto: yeas
rg thinks no 22:17
cotto is so confused
allison darbelo: (I saw your line just after entering mine)
darbelo: permission granted, you should now see a link to "Create content" in the sidebar
cotto is spectest now the same as the old spectest_regression? 22:18
bacek cotto: yes.
darbelo I see it. allison++
cotto Should I merge from trunk to branch then branch to trunk, or is branch to trunk sufficient? 22:26
bacek cotto: docs/project/branching_guide.pod :) 22:28
22:33 kid51 joined 22:34 Themeruta joined
cotto merge time. I'd appreciate it if people could hold off on parrot commits for a little bit. 22:38
Whiteknight darbelo: Let me know when you put up your first post, I want to link to you! 22:39
darbelo Will do. 22:41
purl hmmm... Will do. is there a more appropriate channel for me to ask question regarding trycatch in?
dalek rrot: r38470 | cotto++ | branches/tt528_vtinit (23 files):
bringing vtinit branch up-to-date with trunk
cotto all merged 22:47
dalek rrot: r38471 | cotto++ | trunk (60 files):
[pmc2c] merge the tt528_vtinit branch back into trunk
rg cotto: does that mean the pmc2c test will now fail or did you skip/todo it? 22:50
cotto Oh. That slipped my mind. It'll fail.
Whiteknight purl forget will do 22:54
purl Whiteknight, I didn't have anything matching will do
Whiteknight purl forget Will do.
purl Whiteknight, I didn't have anything matching will do
Whiteknight purl forget hmmm... Will do.
purl Whiteknight, I didn't have anything matching hmmm... will do
Whiteknight purl: you're retarded 22:55
purl Whiteknight: sorry...
Whiteknight damn straight you're sorry
cotto purl, do
purl i guess do is just a loop where something controls/constrains it to execute once
cotto purl, will do
purl cotto: excuse me?
cotto someone must have pm'd it away 22:56
nopaste "NotFound" at 213.96.228.50 pasted "Fix rakudo for c++ build" (37 lines) at nopaste.snit.ch/16474
NotFoundLaptop Can some rakudo guy take a look at this? 22:57
jonathan NotFoundLaptop: Looks sane. 23:00
NotFoundLaptop Maybe I must get a github account 23:03
23:03 Robrt joined
cotto rg, TODOs for pmc2c.t committed 23:03
rg++ for reminding me
dalek rrot: r38472 | cotto++ | trunk/t/tools/pmc2c.t:
[t] TODO some tests that need to be fixed after the tt528_vtinit merge
 23:04
Whiteknight Will do. 23:05
purl rumour has it Will do. is there a more appropriate channel for me to ask question regarding trycatch in
23:05 Robrt left
Whiteknight that bot is absolutely infuriating 23:05
rg cotto: thanks
Whiteknight purl, Will do.
purl well, Will do. is there a more appropriate channel for me to ask question regarding trycatch in?
Whiteknight purl, forget Will do. 23:06
purl Whiteknight, I didn't have anything matching will do
Whiteknight !?!?!
darbelo purl, do.
purl it has been said that do is just a loop where something controls/constrains it to execute once
Whiteknight needs a stiff drink
Infinoid Will do.
purl Will do. is there a more appropriate channel for me to ask question regarding trycatch in?
Infinoid forget Will do
purl Infinoid, I didn't have anything matching will do
darbelo purl, Will
purl Fire at Will!
Infinoid purl, Die in a fire.
purl HALP
dalek rrot: r38473 | Infinoid++ | trunk (2 files):
[cage] line length and trailing whitespace fixes to pass codingstd.
 23:07
cotto purl, forget "Will do." 23:08
purl cotto, I didn't have anything matching "will do."
cotto Will do.
purl it has been said that Will do. is there a more appropriate channel for me to ask question regarding trycatch in?
cotto no, Will do. is <reply>
purl okay, cotto.
cotto Will do.
Whiteknight cotto++ # for my sanity
cotto for mine too 23:09
darbelo Infinoid: Is there anything I can do on this end to help dalek track decnum-dynpmcs commits? 23:11
jonathan NotFoundLaptop: I can apply that patch for you... 23:12
jonathan hands Whiteknight a borovička - should be plenty stiff :-)
Infinoid darbelo: No, the module works fine under the test suite (which apparently simulates a botnix environment better than botnix itself does) 23:13
NotFoundLaptop jonathan: nice, thanks
jonathan no, thank you! :-)
Infinoid I just need to get better diagnostics out of botnix
And by "better", I mean "any at all"
darbelo Ah. Good luck then. 23:14
Infinoid Thanks. It's tonight's project. 23:15
23:15 tetragon joined
dalek kudo: r5ea5aad | jnthn++ | build/Makefile.in:
Add P6Invocation to makefile.
 23:17
jonathan NotFoundLaptop: Comitted, thanks!! 23:20
dalek kudo: rd86c014 | jnthn++ | src/ (2 files):
Fixes to let us build on C++. Patch courtesy of NotFound++.
 23:23
23:26 tetragon joined 23:45 janus joined 23:53 tetragon joined
Coke I have rights to add feeds to the aggregator. 23:58
In general "open a trac ticket" is always a good path.
robrt also doesn't like it when we bug him for stuff we can do ourselves. =-) 23:59

Powered by Cro and the Raku® Programming Language.

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

Thank you!