03:19 particle joined 07:08 particle joined 07:12 particle1 joined 12:26 Whiteknight joined 14:52 particle joined 16:17 particle1 joined 16:49 NotFound joined 17:45 pmichaud joined 17:51 cotto_work joined 18:02 Infinoid joined
Infinoid *lurk* 18:02
pmichaud I will probably miss #ps today. My report is 18:05
* Cleaned up Str.substr in rakudo.
* Updated some tests
* Rakudo is passing 2103 tests as of this morning
EOR 18:06
* Also fixed atan2 18:08
real_EOR
question/comment for today: Tene has submitted a patch for resumable exceptions -- Rakudo could really use this capability soon for lazy list operations (i.e., given/take). Any tuits that could be aimed at resumable exceptions would be ery helpful. 18:09
*very
Tene++
I'm still officially on vacation until Friday. Should have hacking time on the plane(s) though, if I'm not too tired
okay, gone for now. 18:10
18:12 jhorwitz joined 18:19 kj joined
kj early report (will go afk): 18:19
* some clean-ups for pirc/new 18:20
* updated DEPRECATED.pod after revision of pdd19
.eor 18:22
18:29 chromatic joined
tewk has short status today. 18:30
cotto_work has no status today.
18:30 rgrjr joined 18:32 jonathan joined
chromatic Hello, all. 18:33
NotFound Hello
jonathan ahoj
rgrjr hola. 18:34
chromatic Let's go in alphabetical order.
I helped Allison fix bugs and merge back the concurrency branch.
I have some segfaults to look at and will spend some time cleaning up PMCs. 18:35
next?
jhorwitz is late, but here
chromatic cotto_work?
jhorwitz i guess i'm next in alpha order...
NotFound j < n ;)
cotto_work nothing to report, just closing whichever bugs I can wrap my head around 18:36
Whiteknight is here too
chromatic japhb?
japhb chromatic: yes?
japhb looks at channel name ... 18:37
Oh, I see. Nothing to report here.
chromatic jhorwitz?
jhorwitz fixed some threading issues and tested mod_parrot with Apache's threaded worker MPM. it works. this week will add some thread-specific apache directives.
got jonathan up and running with mod_parrot and mod_perl6.
cleaned up some documentation, thinking about writing an HLL layer developer's guide, which would be useful for mod_perl6 hackers. 18:38
EOR
chromatic jonathan?
jonathan * Had Rakudo day last week. Mostly did various fixes, spent time reading the spec/STD.pm and getting my head straight on various things. Had some good discussion with pmichaud later on in the eveing.
* Didn't get my report written... :-( Will do that soon. 18:39
* Started doing some work on my MMD grant today - post on use.perl.org about that, will cross-post to Rakudo.org
* With lots of help from jhorwitz++, managed to get mod_parrot and mod_perl6 to run on my Linux box.
* Got one of my YAPC::EU talks written, almost. Kinda happy with it. Barely started the other one...want to finish it this week before I travel.
* Not sure which day to do Rakudo day this week...will certainly do more on the MMD grant, though.
.end
(OH, and hackathon-ing day before and day after YAPC::EU)
chromatic NotFound? 18:40
NotFound Working on parrot_debugger, adapting to last merge.
I have a question about that, 18:41
Eor
chromatic particle?
particle ~ working with osu osl on infrastructure setup of parrot.org
~ taking over mentorship of auzon, since moritz is on vacation for two weeks
still fighting my dvd drive drivers 18:42
.end
cotto_work queue 1 question
particle pmichaud's report is in scrollback, as is kj's
chromatic rgrjr?
rgrjr Started work on a proposal for a language interop PDD. 18:43
.end
chromatic tewk?
Whiteknight? 18:44
Whiteknight * More debugging work on GC
* Getting close, but no cigar yet.
* Some updates and clarifications to PDD09
EOR
chromatic NotFound, you had a question.
NotFound Main question: is fine to have a runloop specific for parrot_debugger? Secondary: someone is using parrot_debugger? 18:45
chromatic A specific runloop sounds reasonable. I have never used the debugger. Other opinions?
Whiteknight I like the idea. Adding a new runloop is cheap 18:46
Tene is here now.
NotFound The reason is to know if I can work freely in trunk or must I open a branch,
particle i'd like a runloop that can be jumped into and out of at will
NotFound particle: that is the intention. 18:47
particle fab
rgrjr I've tried using parrot_debugger, but it's pretty minimal. I'd rather see effort go into a PIR-level debugger.
particle has two questions
(pir-level debugger)++
NotFound I agree, but first we need a way to run code under debuger control. 18:48
rgrjr NotFound: Seems like these are orthogonal to me . . . but never mind.
NotFound pir-level looks to me just to know line numbers to step and to show. 18:49
Tene No allison today? 18:50
rgrjr I mean I want to be able *write* it in PIR. If I can do that, I can write it in an HLL as well.
particle allison's travelling between cambridge and copenhagen today iirc
NotFound rgrjr: ah, sorry, I don't understood.
My idea is to have some basic debugger functions, and be able to use them both from built-in debugger or an external one. 18:51
rgrjr In particular, I'd rather not write it in C. 18:52
NotFound And work on a pluggable interface, maybe a pmc.
rgrjr My idea of a "built-in" debugger is one that autoloads its PBC files when needed. 18:53
A PMC for contexts would probably be sufficient.
But this may be a distraction . . .
Infinoid (queueing a possibly related question) 18:54
NotFound rgrjr: is resaonable, but I'm currently trying to make the internals work.
spinclad [ pmichaud in scrollback queued a question/comment ]
rgrjr Go for it; I *am* a distraction then.
chromatic Tene, would you like to go? 18:55
Tene My contribution for the week is rt.perl.org/rt3/Ticket/Display.html?id=57610
NotFound Actually it does not work, and I think the runloop is the way to make it work.
particle one of my questions is repeating pmichaud's question, related to tene's work
Tene That's all.
Resumable exceptions patch. It touches on some stuff that I don't understand fully, so could use some review. 18:56
particle rgrjr: do resumable exceptions benefit kea-cl?
(while we wait) they definitely are required for rakudo's gather/take 18:57
Tene The PDDs didn't help at all, as they contained vague and possibly contradictory information.
rgrjr particle: No. CL has a different concept called "proceeding".
Tene Which I should also fix.
particle tene: any pdd review on this topic is encouraged 18:58
Tene Also, the behavior might not be right. It seems like the error handler is popped from the eh stack when it's executed. When I throw two exceptions in a row, I need to push_eh again before the second one if I resume straight from the error handler. 18:59
I don't know if it's supposed to work like that or not.
rgrjr Tene: My reading of S04 is that Parrot should stay in the error context until the handler decides to exit. 19:00
particle we may need a list of use cases to get the tests right
chromatic Agreed.
tewk, would you like to report now? 19:01
Tene My biggest question was if resumable exceptions were already implemented or not. Allison told me that they would be in the pdd25cx merge, but I couldn't find anything that looked similar at all.
tewk * x86 jit of nci stubs is in nci_gsoc branch.
* Working on x86_64 jit of nci stubs
* Question what is the status of x86_64 jit, who was the last one to work on it?
EOR 19:02
chromatic Probably Leo, but I'm sure it's been a while since anyone did anything other than "This may make it compile!" patching.
particle tewk: parrot fails to build for me in nci_gsoc branch with msvc atm
but let's take that bit to #parrot
chromatic cotto_work, you had a question too. 19:03
cotto_work background: I'm trying to get PHPArrays to play nice with Iterators.
One of the features of PHPArrays is that they are ordered and have an internal pointer that refers to the current bucket during iterantion
There are also next(), prev(), reset() and last() methods which move the internal pointer forward, backward, to first or last bucket, respectively.
question: Are the Iterator and Key PMCs flexible enough to rely *only* on the internal pointer (or some other internal of the iterated PMC)? 19:04
19:04 DietCoke joined
chromatic I think they all use vtable entries to iterate through the iterated PMC. 19:04
If they don't, it's a bug. 19:05
DietCoke belatedly arrives.
chromatic DietCoke, do you have a report? 19:06
cotto_work Can you me a high-level idea of how it should work?
question &
chromatic cotto_work, the Iterator should keep the current position in the iterable as a key (0, 1, 2 .. n) and use get_foo_keyed_integer or whatever to retrieve that item when iterated on. 19:07
Though I'm not sure if it pokes into the guts of hash-like PMCs directly.
Basically, try it and see what breaks, so we know what to fix. 19:08
cotto_work ok. I'll see if that's enough to go from and report back later.
DietCoke Here's my report: 19:09
no particular order.
- continued working on partcl
- started a flame^w discussion about branching on list. new svk sounds
spiffy.
- resolved a few old deprecated tickets.
- Working to resolve removal of the old 'getclass' opcode. Feeback on the patch
desired.
.EOR
chromatic Did I miss any questions? pmichaud's queued question? 19:10
cotto_work <pmichaud> question/comment for today: Tene has submitted a patch for resumable exceptions -- Rakudo could really use this capability soon for lazy list operations (i.e., given/take). Any tuits that could be aimed at resumable exceptions would be ery helpful.
particle s/given/gather/
i have a question for rgrjr 19:11
rgrjr: care to share your hll interop proposal (wiki) or do you prefer to work in isolation at this point?
Infinoid chromatic: 1 queued question here
chromatic Infinoid, let's give rgrjr one more minute to respond, and then go ahead. 19:12
rgrjr particle: I had hoped to post it by Sunday, but ran out of tuits. Since there's no way I can cover it all in detail, I had wanted at least to point out what's missing. But if you like, I can post what I have (tonight, probably). 19:13
particle no rush, but many hands light work etc 19:14
rgrjr Agreed. But, in light of the list conversation, my proposal has a definite slant, so I wanted to make a serious attempt at covering all the bases, at least in outline. 19:15
particle ok then, i await your finished draft for review :) 19:17
rgrjr s/finished/barely started/ ;-}
chromatic Alright, Infinoid please go.
Infinoid ok. is there anything blocked on having the ability to manipulate .pbc files from PIR (PDD13)? I'm only spending a day or two per month on this project, so I'm wondering if I should be coordinating with anyone else, or just keep doing my own thing. 19:18
chromatic Replacing IMCC with pirc, potentially.
Infinoid (that said, I have a vacation coming up next week which should include some good PDD13 hacking time.)
particle Infinoid: we'd like to generate pbc directly from post
chromatic++
boat, beer, bytecode? 19:19
sounds like a good vacation
jonathan chromatic: PDD13 doesn't specify a great deal that will make generating the actual bytecode stream (in terms of the ops) a load easier. 19:20
For the overall packfiles, yes, for sure, but in terms of generating a sequence of opcodes etc less so.
chromatic It's progress though.
jonathan Oh, for sure.
Infinoid so, for that, you need some pmcs that will allow you to specify fields and add stuff, and stringify the whole lot?
chromatic PMCs or functions or something better than compilers/imcc/pbc.c 19:21
jonathan :-)
Infinoid great, thanks.
particle yes, we need to convert to/from pasm/pbc
string vs binary representation of bytecode
jonathan erm 19:22
Infinoid hmm... converting between pasm and pbc is pirc's job, right?
I'm working on serialization/deserialization of the file format
jonathan Infinoid: Right. And the missing bit being discussed here is the data hold in one of the segments. 19:23
*held
Infinoid right
jonathan Or rather, code to generate it that isn't pbc.c
Infinoid if you can gimme an array of opcodes, I'll happily store them for you.
jonathan It's probably not crazily hard to do, but it's not especially easy either.
(I'd work on it, but my plate is a tad full right now...I just don't have extra tuits.) 19:24
Infinoid ok, thanks... end of question 19:25
chromatic Any other questions? 19:27
I think we're done then. 19:28
DietCoke One thing:
Please go through the tickets that you have assigned to you in RT and see if any are waiting for you to review something.
I noticed particle has at least one, perhaps there are some easy closers. 19:29
.
chromatic Same time next week. Thank you everyone.
particle ~~
19:30 DietCoke left, jhorwitz left, NotFound left, chromatic left 19:33 cotto_work left 19:44 Infinoid left 19:47 jonathan left 22:32 particle 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!