Parrot 3.3.0 released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Trac accounts should be working properly again; talk to cotto, coke or whiteknight if you have trouble
Set by moderator on 13 May 2011.
00:00 rofl0r joined, lucian left
rofl0r i hope rohit_nsit08 won't fail at the gsoc as well 00:01
since he can't even correctly copy/paste an URL
www.google-melange.com/gsoc/proposa...t_nsit08/1 <- this would be the right url 00:02
(i'm referring to parrot.org's homepage, 3rd entry to be exact) 00:04
00:17 davidfetter left 00:29 kid51 joined
moderator Parrot 3.3.0 released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Parrot Developer Summit, Sat May 14 2100 UTC this channel 00:37
whiteknight rofl0r: we accepted him because of his coding skills, not his blogging skills 00:43
00:44 bacek joined
rofl0r well, a good coder should know how to copy paste an url instead of typing it by hand, which is pretty error prone :V 00:45
kid51 rofl0r: ... and your contributions to the Parrot project have been ... ? 00:46
rofl0r actually i'm telling you this so that someone can fix the link 00:48
00:48 whiteknight left
kid51 bacek: will you be able to make PDS? 00:49
bacek kid51, when is it? Tonight?
aloha, pds?
aloha bacek: pds is now, if anybody forgot
rofl0r i had preferred to post a comment, but i don't feel like registering on yet another webpage
bacek aloha, parrot developer summit 00:50
aloha, parrot developer summit?
aloha bacek: I have no idea.
kid51 aloha: PDS is Sat May 14 2011, 2100 hours
aloha kid51: ... but PDS is now, if anybody forgot ...
kid51 aloha: forget pds
aloha kid51: I forgot about pds.
kid51 aloha: PDS is Sat May 14 2011, 2100 hours
aloha kid51: Okay.
00:50 whiteknight joined
kid51 pds? 00:50
aloha, pds?
aloha kid51: pds is Sat May 14 2011, 2100 hours
kid51 good bot
rofl0r whiteknight, you missed this <rofl0r> actually i'm telling you this so that someone can fix the link 00:51
kid51 aloha: Parrot Developer Summit is PDS
aloha kid51: Okay.
rofl0r and that <rofl0r> i had preferred to post a comment, but i don't feel like registering on yet another webpage
bacek kid51, it's 7am tomorrow for me. I'll be little bit later. May be 7:30
kid51 bacek: Yes, it's doomed to be at a bad time for someone. Look forward to seeing you. 00:52
well, seeing your keystrokes
00:54 user_3491 joined 00:55 user_3491 left, kid51 is now known as kid51_at_dinner 00:56 whiteknight left, whiteknight joined
Tene rofl0r: Yes, there have been several times that I've wanted to tell someone about an error, but their blog required registration, so I hunted them down on IRC or sent email instead. I hate that. 00:59
rofl0r the email method seems the most effective 01:00
seems nobody cares about bugs reported in IRC or even bugtrackers 01:01
Tene whenever I've reported an error about a blog post on IRC, it's been fixed. I must read more-conscientious bloggers than you do. ;) 01:02
rofl0r yeah, it's my first "blog-bug" i reported. 01:03
Tene rofl0r: What's your interest in Parrot? I'm always curious about what draws in different people.
rofl0r usually its software bugs
uhm, i was just surfing around perldoc.org and then clicked on the parrot link, to see if its slowly getting somewhere 01:04
after the 3rd redesign it may finally get usable...
Tene Yeah, I've been pretty hopeful about recent work on Parrot. Seems like things are finally starting to come together well. :) 01:05
rofl0r hopefully 01:07
whiteknight you know, things move faster and go better if more people help with the coding 01:08
Tene :0 01:10
:)
whiteknight Tene: I need to figure out what your favorite baked good is, and then bribe you with a plate of them :) 01:11
Tene whiteknight: I'm not a big fan of baked goods in general, and what appetite I do have for them, my girlfriend provides in excess. She's always looking for people to give away baked goods to. 01:13
whiteknight damnit! There goes all my leverage
01:14 redicaps joined
Tene hehe :) 01:14
bacek
.oO( "Backed goods"? Can you drink them? )
 01:15
seen cotto
aloha cotto was last seen in #parrot 9 hours 47 mins ago saying "so I do".
bacek seen cotto_work
aloha cotto_work was last seen in #parrot 1 hours 48 mins ago saying "winxed is nice like that".
Tene as a reminder, there's an open invitation to everyone to do whatever you'd like to try to get me working on Parrot again. I really enjoyed working on Parrot, back when I was contributing regularly. I seem to be recovering from burn-out pretty well, but somehow haven't been able to get it together enough to start working on Parrot again yet. 01:16
dalek rrot: 8f34b7f | jkeenan++ | MANIFEST:
Someone forgot to update MANIFEST after adding new files.
rrot: b8b412c | bacek++ | src/gc/gc_gms.c:
Fix braino in gc_gms_finalize
 01:19
nopaste "plobsing" at 192.168.1.3 pasted "[PATCH] zavolaj string conversion (BROKEN)" (97 lines) at nopaste.snit.ch/45257 01:21
ttbot Parrot 8f34b7fc MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/5049 01:26
Parrot b8b412c0 MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/5076 01:35
01:44 rofl0r left 01:52 wagle left
dalek rrot: 724cdbb | bacek++ | src/gc/gc_gms.c:
DRY in gc_is_pmc_ptr
 01:52
rrot: 9d4b5dd | bacek++ | src/gc/gc_gms.c:
Temporary code to investigate win32 GC failure
01:55 wagle joined
kid51_at_dinner bacek: At commit b8b412c0bf on darwin/ppc, I got seg fault at: ./parrot-nqp --target=pir --output=compilers/opsc/gen/Ops/Op.pir compilers/opsc/src/Ops/Op.pm 01:55
01:55 kid51_at_dinner is now known as kid51
cotto bacek, hio 01:56
bacek kid51, it's... impossible.
cotto, aloha. I'm trying to "blindly" fix GC failure on win32...
cotto bacek, I'm pretty sure there's an achievement for that.
kid51 pulls and retries 'make gc_tests' 01:57
cotto actually, for any two of those
01:57 jsut_ joined
cotto I can't help much until Monday 01:57
bacek cotto, at least I've indication which commit broke it. Just don't understand why. 01:59
cotto it's a start 02:00
ttbot Parrot 9d4b5dd5 MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/5174 02:01
02:02 jsut left
bacek still failing 02:02
nopaste "kid51" at 192.168.1.3 pasted "cotto: proposed PDS agenda" (19 lines) at nopaste.snit.ch/45279 02:07
kid51 cotto: can you review paste?
darwin/ppc: parrot once again builds at commit 9d4b5dd5; running make gc_tests now 02:08
cotto kid51, thanks. will do 02:12
bubaflub seen dukeleto 02:21
aloha dukeleto was last seen in #parrot 4 hours 28 mins ago saying "cotto_work: my freebsd box is hosed for now. can't test on it".
bubaflub msg dukeleto i finished all my finals. i walk on Tuesday; let's figure out a time to skype and i'll write up my introductory blog post on parrot.org 02:22
aloha OK. I'll deliver the message.
dalek rrot: 1dc5e21 | jkeenan++ | src/platform/ (2 files):
[codingstd] Need at least one space between C keyword 'sizeof' and open parenthesis.
 02:24
whiteknight msg NotFound I'm making a Winxed REPL. "using" functions is a big pain to do every time. Is it possible to either get a syntax for calling function with the namespace (foo.bar()) or else I would like to insert a custom sub lookup routine into the compiler
aloha OK. I'll deliver the message.
cotto How feasible would it be to revert the removal of "t" from nci sigs without undoing anything else? 02:29
need to go afk 02:31
will backscroll
dalek sella: a17e37a | Whiteknight++ | src/multicast/Function (2 files):
start converting NQP files to winxed for multicast library
 02:32
sella: 8dee9c4 | Whiteknight++ | src/string/ (2 files):
fixes to the string library so it builds and parts of it work for some tests
sella: 0c44b92 | Whiteknight++ | src/multicast/FunctionCompositor.winxed:
some fixes/changes to function compositor
sella: f4e6bf9 | Whiteknight++ | / (4 files):
Add REPL functionality to the winxed library. Add in a quick driver program to run it
kid51 darwin/PPC: parrot builds at 9d4b5dd568 and PASSes make gc_tests and make coretest 02:36
linux/i386 make test PASS at 1dc5e215 02:37
cotto: Okay to post that agenda to parrot-dev?
whiteknight msg NotFound or a global way to set which namespaces to search would be nice 02:39
aloha OK. I'll deliver the message.
02:40 whiteknight left
ttbot Parrot 1dc5e215 MSWin32-x86-multi-thread make error tt.taptinder.org/cmdinfo/5306 02:41
02:44 redicaps left 02:45 bubaflub left 02:57 kid51 left 03:01 Andy joined
dalek rrot: 15fdf60 | bacek++ | src/gc/gc_gms.c:
Revert most part of 6f0cfa8 because I can't figure out why it broke win32.
 03:03
rrot: 2ff3c93 | petdance++ | / (2 files):
adding headerizer args
 03:07
bacek sigh... 15fdf60abf did fix win32. Don't quite understand why it was broken in first place. 03:14
plobsing bacek++ 03:21
atrodo dukeleto: ping 03:26
dukeleto: unping, found my answer 03:30
plobsing dukeleto: ping 03:31
03:40 Andy left
plobsing msg dukeleto (re: jitterbug.leto.net:3000/api/build/p...l-v5.10.1) this is a cascading fail starting *very* early in the config. is this the first commit that has that error? this commit doesn't touch any config-related stuff. perhaps you've hit your disk-quota or something similar? 03:45
aloha OK. I'll deliver the message.
04:40 theory left 04:48 pmichaud_ left 04:57 bubaflub joined 05:02 pmichaud joined 05:12 bubaflub left 05:40 Anxuiz left
cotto pmichaud, the first thing on my plate tomorrow morning is to dig into the nci issue and (1) revert the offending commit and (2) provide a very clear upgrade path, barring a technical reason why the revert wouldn't be possible. 06:31
Whether the fault is lack of perspective on our part, a hole in our policy or just ignoring what policy we have, it's become a cluster. I don't intend on letting it get worse.
06:57 contingencyplan joined 07:39 dodathome joined 07:48 rurban_ joined 07:49 rurban left 07:50 rurban_ is now known as rurban 08:36 ShaneC joined 08:37 mj41 joined 09:02 mj41 left 09:30 contingencyplan left
dalek p: 372e310 | jonathan++ | src/6model/reprs/P6opaque.c:
Fix crash when we have a typed attribute but the meta-attribute doesn't expose a .box_target.
 10:29
10:30 redicaps joined 10:39 redicaps left 10:42 woosley joined 11:02 Psyche^ joined 11:07 Patterner left, Psyche^ is now known as Patterner 11:09 whiteknight joined
whiteknight good morning, #parrot 11:14
tadzik good morning whiteknight 11:19
whiteknight hello tadzik. How are you doing? 11:23
tadzik: I never heard the final diagnosis on your arm 11:24
tadzik whiteknight: I haven't either, that's the fun part. They took of my cast, didn't even make an x-ray, the doctor just examined it for a while and said "oh, you're fine" 11:25
It was probably twisted, but I'll never know
whiteknight oh
tadzik now I'm struggling with my GSoC work 11:26
whiteknight what's funny is that it probably cost several hundred dollars for that informative meeting with a doctor
tadzik actually not, this is public healthcare 11:27
now if it was private healthcare, doctors will be smiling, gentle, asking if I feel right, and what can they do for me. Also, I wouldn't be standing like a bum in front of the door on 7 AM with a dozen of other people who are ought to be diagnosed today, but the nearest free slot is a month away 11:28
whiteknight oh 11:40
well, at least the shoddy treatment is free 11:41
tadzik yeah
as long as you have insurance, but that's quite normal 11:43
11:47 ShaneC left
dalek p: b12c862 | jonathan++ | src/NQP/ (2 files):
Add a way to stub in a type without doing anything with it other than new_type and with an explicitly specified meta-object type.
 11:49
p: d0f50b8 | jonathan++ | src/NQP/Actions.pm:
Ensure we build proper AST node in newly added action method.
p: 3b5eec5 | jonathan++ | src/NQP/SymbolTable.pm:
Rather than dying when trying to run code we didn't compile yet, instead try to compile it on-demand and run it. Missing various environmental bits just yet, but essentially works.
bacek ~~ 11:53
jnthn__, ping
dalek p: a555ce0 | bacek++ | src/6model/reprs/P6opaque.c:
Add awful, terrible block of GC
 12:18
bacek aloha, seen tadzik 12:22
aloha bacek: tadzik was last seen in #perl6 23 seconds ago saying "heh, I'll upload my NLPW pics too, for the fun".
bacek it works :)
tadzik (: 12:23
12:28 ambs joined
bacek msg jnthn Looks like my last commit fixed nqp on GenGC. At least it didn't crashed yet during compilation in 256K (yes, "K") available memory. 12:33
aloha OK. I'll deliver the message.
12:34 perlite left, perlite joined 12:38 fperrad joined 13:00 woosley left 13:12 lucian joined 13:20 JimmyZ joined
pmichaud github.com/pmichaud/rakbench/blob/...140653.txt # latest rakbench results 13:28
github.com/pmichaud/rakbench/blob/...140651.txt # latest rakbench results, #2
13:53 redicaps joined
jnthn__ bacek: It can compile itself using 256K available memory?! 14:05
14:10 dodathome left 14:19 SHODAN joined 14:48 woosley joined 14:51 redicaps left 14:54 kid51 joined 15:02 contingencyplan joined 15:06 bubaflub joined 15:09 theory joined 15:15 mj41 joined 15:18 JimmyZ left 15:19 jsut joined 15:24 jsut_ left, whiteknight left
dalek sella/winxed_constructors: ae3de80 | Whiteknight++ | src/winxed/Repl (2 files):
some refactors on the repl
 15:25
sella/winxed_constructors: 1e35764 | Whiteknight++ | src/winxed/Repl.winxed:
add better results printing to the repl
sella/winxed_constructors: af9f96e | Whiteknight++ | s (3 files):
break CClassTokenizer up. Add a Tokenizer abstract parent class
sella/winxed_constructors: 34da27a | Whiteknight++ | src/core/ (2 files):
update the core library to use constructors instead of BUILD
sella/winxed_constructors: a1b7b92 | Whiteknight++ | src/ (10 files):
update action and container to use proper constructors
sella/winxed_constructors: c2df478 | Whiteknight++ | t/ (41 files):
convert build to construct in t/
sella/winxed_constructors: 44fa98a | Whiteknight++ | / (4 files):
fix action types
sella/winxed_constructors: 08806cf | Whiteknight++ | src/ (35 files):
mass-convert many other files
sella/winxed_constructors: 7060c57 | Whiteknight++ | / (6 files):
fixes and cleanups
15:27 whiteknight joined 15:39 bluescreen joined 15:40 mariano joined 15:41 bluescreen_ joined 15:45 bluescreen__ joined 15:46 bluescreen__ left, mariano left, bluescreen__ joined 15:47 rurban_ joined 15:50 dodathome joined, rurban left 15:51 rurban_ is now known as rurban 15:53 NotFound_b joined
NotFound_b whiteknight: ping 15:54
whiteknight pong
NotFound_b whiteknight: the scope search for function calls is already available, but not in the installable version,
whiteknight NotFound_b: okay, what does that do? 15:55
NotFound_b I'm waiting for r1000 to update it.
whiteknight I'm trying to find functions defined in other namespaces
cotto ~~
NotFound_b whiteknight: it search for functions in namespaces in scope and its parents. 15:56
whiteknight NotFound_b: okay, what is "in scope"?
NotFound_b And does the same as "using static" whenn found.
whiteknight: for example, a function in namespace a uses b.something(); it searchs in a.b and in b 15:57
something() search in a and in the root namespace 15:58
whiteknight okay, so if I am in namespace Rosella, it will search in Rosella.String?
NotFound_b Scopes in current compilation unit, at compile time.
whiteknight: no, that will need "using namespace", and that is not implemented yet. 15:59
whiteknight ok
I'm probably going to add a --ns= commandline argument to the repl
I created it to do quick testing with Rosella 16:00
cotto contemplates the sound of one chainsaw clapping
NotFound_b whiteknight: I'll probably do r1000 today updating the installable compiler with the recent changes, but I'm not sure, I don't have much time available. 16:01
whiteknight NotFound_b: okay.
NotFound_b SYL
16:02 NotFound_b left 16:06 theory left
dalek sella: ae3de80 | Whiteknight++ | src/winxed/Repl (2 files):
some refactors on the repl
 16:06
sella: 1e35764 | Whiteknight++ | src/winxed/Repl.winxed:
add better results printing to the repl
sella/winxed_constructors: cf1b1d9 | Whiteknight++ | src/ (4 files):
fixes so that we build, run the test suite, and pass most tests again
sella/winxed_constructors: ea3e4c9 | Whiteknight++ | src/mockobject/ (7 files):
fix constructors in mockobject, reclaim more tests
sella/winxed_constructors: 765d889 | Whiteknight++ | src/ (8 files):
last few fixes. All tests pass
16:07 lucian_ joined 16:12 lucian left 16:14 lucian joined, lucian_ left
dalek sella: af9f96e | Whiteknight++ | s (3 files):
break CClassTokenizer up. Add a Tokenizer abstract parent class
 16:14
sella: 34da27a | Whiteknight++ | src/core/ (2 files):
update the core library to use constructors instead of BUILD
sella: a1b7b92 | Whiteknight++ | src/ (10 files):
update action and container to use proper constructors
sella: c2df478 | Whiteknight++ | t/ (41 files):
convert build to construct in t/
sella: 44fa98a | Whiteknight++ | / (4 files):
fix action types
sella: 08806cf | Whiteknight++ | src/ (35 files):
mass-convert many other files
sella: 7060c57 | Whiteknight++ | / (6 files):
fixes and cleanups
sella: cf1b1d9 | Whiteknight++ | src/ (4 files):
fixes so that we build, run the test suite, and pass most tests again
sella: ea3e4c9 | Whiteknight++ | src/mockobject/ (7 files):
fix constructors in mockobject, reclaim more tests
sella: 765d889 | Whiteknight++ | src/ (8 files):
last few fixes. All tests pass
 16:15
16:18 lucian_ joined, lucian left
cotto pmichaud, ping 16:18
16:21 bluescreen__ left 16:22 bluescreen__ joined 16:23 bluescreen left, bluescreen_ left, bluescreen__ left
cotto pmichaud, is it just 't' that's causing all the pain or are there other nci types that need to be resurrected? 16:25
16:27 woosley left
kid51 cotto: Do you want that agenda I pasted yesterday to be posted to parrot-dev? 16:30
dalek Heuristic branch merge: pushed 17 commits to Rosella/path_refactor by Whiteknight 16:31
cotto kid51, yes, thank. I haven't come up with anything that's missing. 16:32
*thanks
dalek sella/gh-pages: db43892 | Whiteknight++ | libraries/core.md:
updates for core documentation
 16:41
16:45 fperrad left 16:48 fperrad joined 16:55 dodathome left
dalek rrot/headerizer-work: ccd2ffd | petdance++ | / (55 files):
Don't show SHIMs in the function declarations
 16:57
rrot/headerizer-work: 12bed22 | petdance++ | / (31 files):
Don't putted SHIMmed interps in the declarations
17:03 kid51 left, dodathome joined 17:05 contingencyplan left
dukeleto ~~ 17:12
cotto hio dukeleto
dukeleto cotto: it is just 't' that is causing all this ruckus 17:13
cotto: hola
is there a PDS today?
cotto dukeleto, good to know
yup
see /topic 17:14
dukeleto indeed 17:16
cotto: should we be preparing something about M0 for the PDS?
cotto dukeleto, that would be a sensible course of action 17:18
atm I'm trying to get 't' working
dukeleto cotto: how do you feel about pmichaud++'s comments about the recent NCI "upgrade path"? 17:20
17:20 mj41 left
cotto dukeleto, that it's insufficiently documented? 17:20
dukeleto cotto: yeah 17:21
cotto rings true to me
dukeleto cotto: StructView also throws lots of warnings in the build, which does not make me feel warm and fuzzy
cotto I like Ptr* et al but they're not very approachable
dukeleto cotto: yeah, we need friendlier documentation 17:22
cotto and more examples
dukeleto cotto: i think, in general, we need some kind of "documentation team" that kicks peoples ass about having approachable and friendly docs
cotto: yes, examples as well
cotto dukeleto, we need someone to take on that role 17:25
If it's just getting after people, one person can do that fine.
dukeleto cotto: just sent an email to parrot-dev about it 17:31
17:31 mj41 joined
cotto dukeleto, thanks 17:31
dalek sella: 76be71f | Whiteknight++ | VERSION:
bump version numbers, since we've made backwards-incompatible changes to constructor behavior
sella: adeb72f | Whiteknight++ | / (5 files):
version information is now stored in Rosella.Version objects.
sella: b747808 | Whiteknight++ | src/compose/test.nqp:
remove the compose directory, not using it
sella: e73c0cc | Whiteknight++ | VERSION:
+version info for two new libraries
sella: 3bc194d | Whiteknight++ | src/benchmark/Benchmark.winxed:
add in some stub logic for a new benchmarking library
sella: ea21764 | Whiteknight++ | VERSION:
benchmark is 0/alpha
rrot/shims: 74661cf | petdance++ | / (60 files):
We no longer put SHIMs in the function declarations
 17:35
17:51 jsut_ joined
cotto plobsing, ping 17:55
17:55 jsut left
plobsing cotto: pong 18:00
cotto plobsing, can you re-add the 't' nci type? 18:04
dukeleto, I'm not seeing anything from you on parrot-dev
18:05 mj41 left
plobsing cotto: it isn't that easy. the infrastructure has been changed since and that would need to be rolled back as well. that entails rolling back all of the improvements made since 18:06
dalek rrot: 2eb9fb1 | dukeleto++ | NEWS:
Improve NEWS
 18:07
cotto plobsing, is nci now incompatible with 't'?
I saw that there were some significant changes.
plobsing cotto: effectively, yes
dukeleto cotto: can you add a NEWS item for your get-entropy branch?
cotto well dang
dukeleto, sure 18:08
plobsing further, I am not convinced that is necessary. zavolaj can be updated as my preliminary patch shows 18:09
dukeleto plobsing: i missed your patch. where is it? 18:10
plobsing nopaste.snit.ch/45257
dukeleto plobsing: i can attempt to help unbreak zavolaj, if you need testing on different systems, or a second set of eyes on something
plobsing dukeleto: it is just a matter of rakudos complex referency things that I can never get straight 18:11
that really needs a rakudo core dev. I talked to pmichaud, who defered to jnthn, who I msged.
dukeleto plobsing: looks like you can unify the pir:: code to another function, in that patch 18:12
cotto plobsing, thanks for working on that
dukeleto plobsing: what does the zavolaj test suite do on that patch?
zavolaj?
plobsing dukeleto: probably. I was more concerned with getting it working than making it good. it also leaks memory
dalek rrot: d5b49dd | cotto++ | NEWS:
Merge branch 'master' of github.com:parrot/parrot
dukeleto plobsing: mmmm, memory.
dalek rrot: 8a1a19e | cotto++ | / (16 files):
Merge branch 'compiletime-git-describe'
dukeleto plobsing: because you are never free'ing the allocated c-strings?
plobsing: sometimes I take what I can get :)
plobsing dukeleto: zavolaj lacks a meaningful testsuite. it is best to pick an example that you think will work on your machine and run that. 18:14
moritz MiniDBI has a test suite and uses zavolaj
you can use that
plobsing dukeleto: it is a known problem with the current patch. I have ideas on how to fix it. It will be fixed after the core of the patch is working. One thing at a time.
dukeleto: did you ever manage to figure out that failure you msged me with? 18:17
ttbot Parrot 5a96a78f i386-linux-thread-multi make error tt.taptinder.org/cmdinfo/5969 18:18
dalek rrot: 41a226f | cotto++ | NEWS:
add entropy to NEWS
cotto awesome 18:19
cotto tries to fix the build
doesn't seem to fail for me 18:22
plobsing cotto: looks like a missing dependancy on datatypes.pasm 18:23
nopes, the dependancy is stated in the makefile. wtf.
cotto except it's there
It does look like it might have something to do with the parrot_version.pir file that merge added 18:25
*the merge
odd. I'm tempted to ignore it until it shows up again. 18:27
dukeleto plobsing: yes, it was a problem with the cached git repo jitterbug was using, not a problem with your branch. 18:32
plobsing: parrot is the first C-based project to get tested by jitterbug, so it is shaking out some bugs :)
cotto: did you get this email yet? lists.parrot.org/pipermail/parrot-d...05881.html 18:33
cotto dukeleto, nope 18:34
dukeleto cotto: welcome to the past
cotto: might want to tell your email provider to hire faster passenger pidgeons
cotto guess su 18:35
so
dukeleto moritz: i see that rakudo has a GD library that uses zavolaj directly 18:36
moritz: perhaps that is a good canary
plobsing I've been trying with the mysqlclient example 18:37
dalek rrot/shims: 9d4a194 | petdance++ | frontend/pbc_merge/main.c:
Remove unused ct argument
 18:42
rrot/shims: 181a959 | petdance++ | / (4 files):
fix shim arguments
rrot/shims: f8e277c | petdance++ | / (5 files):
fixing function annotations
dukeleto plobsing: what is the status of github.com/plobsing/parrot-libjit-fb ? zavolaj's README links to it.
plobsing dukeleto: it was working up until the latest refactor. it is currently broken, but the fix is straightforward.
I'd like to keep it around since it is more efficient than libffi. It could be useful for implementing a full JIT for Parrot. 18:44
I'd like to see parrot's JIT provide NCI thunks and a replacement for structview. 18:45
dukeleto plobsing: ok, sounds good. Just wanted to understand the status of it. That clears it up.
18:45 davidfetter joined
dukeleto plobsing: does structview throw lots of warnings when compiling on your machine? 18:45
plobsing dukeleto: yeah. stupid stuff it shouldn't warn about. I've looked at them and they are not a problem. 18:46
dukeleto man, parrot finds all kinds of bugs in jitterbug. Caching git repo's is not simple. I need to nuke the actual directory every time and copy a pristine git repo to a temp dir, evidently.
plobsing dukeleto: can't you 'make realclean' at the end of your tests and keep it around for next time? 18:47
dukeleto plobsing: i currently do "git clean -fdx" 18:48
plobsing: "make realclean" can be broken, for any given commit to the repo 18:49
plobsing: the build could fail, etc
plobsing: but still, i run into issues
whiteknight oh awesome, the parrot-dev mailing list was randomly selected to win on million dollars from the british telecomlottery 18:51
one million dollars *
would anybody mind if I reply to them with a list of all your real names, mailing addresses, and government ID numbers?
cotto My government id number is 3 18:53
It's a very small government. 18:54
dukeleto whiteknight: my SSN is "Inf"
davidfetter wonders whether +NaN is an acceptable SSN 18:55
plobsing whiteknight: are those british dollars? 18:57
SHODAN my name is Robert');DROP TABLE Winners;--
whiteknight plobsing: I didn't think they used dollars over there, but the email looks legit. They must have switched
cotto odd. I sent myself a test message just fine but still haven't seen dukeleto's message to parrot-dev. 18:59
19:01 soh_cah_toa joined
pmichaud PDS is on #parrot or #parrotsketch or ... ? 19:24
sorear #parrotsketch
cotto dukeleto, did the agenda ever get sent? I don't seem to be getting stuff sent to parrot-dev. 19:35
pmichaud I didn't see it on parrot-dev 19:40
but it's in the archive: lists.parrot.org/pipermail/parrot-d.../date.html 19:41
so somebody's mailer is slow
plobsing I have not recieved that mail either
pmichaud so far I've gotten the "Release Managers Needed!" email but nothing after that
19:43 luben left, luben joined
tadzik same here 19:44
cotto dukeleto, can you poke parrot-dev with a stick? Others are having the same problem I am of not seeing anything after "Release Managers Needed!". 19:48
19:50 SHODAN left
cotto The nice thing about having an upcoming release is that I can look for release names like this: wins.failblog.org/2011/05/11/epic-w...ral-lakes/ 19:51
20:04 mj41 joined
nopaste "plobsing" at 192.168.1.3 pasted "[PATCH] more working-ish zavolaj upgrade" (109 lines) at nopaste.snit.ch/45500 20:08
cotto PDS in 35 20:25
How does it work that we have stupid technical issues right around PDS? 20:35
pds agenda (since parrot-dev is wonky) nopaste.snit.ch/45279 20:39
20:46 soh_cah_toa left 20:49 soh_cah_toa joined
cotto hio sorear 20:50
er, soh_cah_toa
soh_cah_toa cotto: hey :)
20:51 dodathome left
dukeleto ~~ 20:52
cotto hi dukeleto
dukeleto cotto: howdy. just got back from a delightbul bike ride 20:53
cotto: i don't know that I have a stick to poke parrot-dev with
cotto: i am alergic to maintaing mailing lists
plobsing: it seems that it would be very beneficial to have a test to go along with that fix in zavolaj 20:55
plobsing: a canary in the coal mine, if you will
cotto dukeleto, rough allergy 20:56
20:56 kid51 joined
pmichaud I've posted my comments to gist.github.com/972625 20:57
moderator Parrot 3.3.0 released | parrot.org | Log: irclog.perlgeek.de/parrot/today | Parrot Developer Summit, Sat May 14 2100 UTC #parrotsketch 20:57
dukeleto PDS now? 21:00
tadzik aloha: clock?
aloha tadzik: LAX: Sat, 14:00 PDT / CHI: Sat, 16:00 CDT / NYC: Sat, 17:00 EDT / UTC: Sat, 21:00 UTC / LON: Sat, 22:00 BST / BER: Sat, 23:00 CEST / TOK: Sun, 06:00 JST / SYD: Sun, 07:00 EST
tadzik yeah, seems so
dalek nxed: r1000 | NotFound++ | trunk/winxedst1.winxed:
fix mistake in Modifier.optimize
 21:07
nxed: r1001 | NotFound++ | trunk/ (3 files):
update installable files
tadzik millenium, edi... ah, too late! 21:08
21:11 mj41 left 21:25 jsut joined 21:27 fperrad left 21:30 ambs left, jsut_ left 21:43 mikehh left
kid51 has opened support ticket with OSU OSL re delays in distribution to our mailing lists 21:44
pmichaud kid51++
cotto kid51++
dalek rrot: 63df102 | bacek++ | src/gc/gc_gms.c:
Get back some performance of is_(pmc|string)_ptr. This invariant should always hold
 21:51
21:54 mikehh joined
dalek sella/path_refactor: e03b533 | Whiteknight++ | / (2 files):
Fix the path algorithm so it's properly recursive and passes all tests
 21:54
sella/path_refactor: 63d8469 | Whiteknight++ | src/path/Path.winxed:
Fix path so it passes all tests
sella/path_refactor: b8f921e | Whiteknight++ | src/path/Path.winxed:
remove debugging statement
sella/path_refactor: 01ccfd2 | Whiteknight++ | src/path/Path.winxed:
cleanup and remove cruft
sella: e03b533 | Whiteknight++ | / (2 files):
Fix the path algorithm so it's properly recursive and passes all tests
 21:57
sella: 63d8469 | Whiteknight++ | src/path/Path.winxed:
Fix path so it passes all tests
sella: b8f921e | Whiteknight++ | src/path/Path.winxed:
remove debugging statement
sella: 01ccfd2 | Whiteknight++ | src/path/Path.winxed:
cleanup and remove cruft
sella: b575817 | Whiteknight++ | / (7 files):
Merge branch 'path_refactor'
22:22 bluescreen joined 22:25 lucian_ is now known as lucian
nopaste "kid51" at 192.168.1.3 pasted "Draft: PDS policy statements" (15 lines) at nopaste.snit.ch/45564 23:16
dukeleto cotto: ping 23:24
cotto dukeleto, pong
dukeleto cotto: how do I represent "I37" or "S42" in m0 bytecode? 23:25
cotto: i am looking thru to spec 23:27
cotto dukeleto, Ix = 8+x, Nx = 70 + x, Sx = 132 + x, Px = 194 + x
dukeleto cotto: is that in the m0-spec? or only in your implementation? 23:28
cotto dukeleto, it's indirectly in the spec
Register Types and Context Structure
dukeleto cotto: ah, there it is 23:29
23:31 contingencyplan joined
dukeleto cotto: what is the M0 assembler supposed to do when it sees I62/N62/etc ? 23:31
cotto pmichaud, I talked with plobsing and he said that 't' would be difficult to add back after some nci refactors that have happened since the removal. He's got a patch for zavolaj that he's been working on to work around the lack of 't'. 23:32
dukeleto, unicode sad face
pmichaud cotto: I think that's the wrong approach entirely (more)
if 't' is hard to do, it's not made any easier by making every HLL developer do it (every time the HLL needs to pass a cstring to a C function) 23:33
dalek rrot/m0-spec: 4421ed4 | dukeleto++ | docs/pdds/draft/pdd32_m0.pod:
Fix formatting of segment numbers
23:33 kid51 left
pmichaud I'm sorry, but I can't imagine doing the "make a cstring / free the cstring" dance every time I want to make a C function call that expects a cstring 23:34
especially if the dance is as difficult as it's proving to be for zavolaj (see patch)
cotto pmichaud, I see where you're going. If it's hard in nci, it'll still only be hard once. 23:35
pmichaud right
it just feels to me like we're making life difficult for N HLL developers
and, in the case of zavolaj, that's just one NCI library
passing a c string to a C func is absolutely one of the most common use cases I can come up with. 23:36
it just seems bizarre that it would be so difficult to do from a user perspective
cotto pmichaud, yes. It needs to be a supported use case. 23:37
pmichaud and supported along the lines of what we can already do for int, float, and other structs
i.e., with a simple letter, not with a complex "oh, call these other things you obtained via dlfunc..." bit
fwiw, I'm fine if the the simple is not perfectly robust. 23:38
i.e., it's okay to say "you can use 't', but beware if you do anything outside of X-Y-Z boundaries because your string might disappear/leak/whatever
in other words, 't' doesn't have to be the "handle every string being passed in every instance" -- it just has to be good enough to handle the common instances. if a caller is doing something a bit more uncommon, then it's okay to say "you have to do some setup/takedown on your own to make this work" 23:39
I think I can argue that what zavolaj is doing (pass a simple string to a C function) is a very common use case :-) 23:40
cotto easy argument
pmichaud it may be that parrot's nci is unable to handle that case, in which case I'd posit that nci is failing its audience
but we've already seen that it's possible to do (from existing UnManagedStruct), and before throwing out the 't' option entirely I'd like to see a more detailed discussion of why it's absolutely necessary 23:41
23:41 contingencyplan left
pmichaud anyway, that's all I needed to add. I'll let you guide it from there with plobsing? If you need more input, ask jnthn or I, but I will say that I think the patches that have been suggested so far ought to be non-starters for discussion. 23:42
cotto Thanks.
23:47 rurban_ joined 23:49 rurban left, rurban_ is now known as rurban
dalek rrot/m0-prototype: 20a1d88 | dukeleto++ | src/m0/m0_assembler.pl:
Teach the M0 assembler how to convert symbolic register names into their binary representation
 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!