weekly meeting at 18:30 UTC, Tuesday | logged at irclog.perlgeek.de/parrotsketch/today
Set by moderator on 30 September 2008.
01:58 particle1 joined 15:40 davidfetter joined 16:05 pmichaud joined 17:02 davidfetter joined 17:16 kj joined 18:03 moritz joined 18:09 masak joined 18:11 barney joined 18:15 chromatic joined 18:20 ILyuha joined 18:21 Wknight8111 joined
tewk_ get_global should go 18:25
<@chromatic> ... except there's no get_hll_global PMC STR STR 18:26
18:26 cotto joined
pmichaud (get_global) -- it's been long settled that find_global is deprecated and get_*_global remains. 18:29
chromatic We can't remove find_global until get_*_global does everything that find_global does, however. 18:30
18:30 jhorwitz joined
moritz detects that it's time[tm] 18:30
chromatic Hello, all. 18:31
barney hi
davidfetter OH HAI
masak y0
particle ZOMG
jhorwitz ~~
Wknight8111 hello
18:32 jonathan joined
kj hi 18:32
chromatic Let's go in reverse order. Everyone can be excited and confused.
Wknight8111?
Wknight8111 * A few small fixes here and there, nothing major
18:32 bacek joined
Wknight8111 * Spent a week on vacation, so slowly getting caught up 18:32
EOR
chromatic And congratulations in order for the WHY of vacation. 18:33
tewk_?
Wknight8111 thank you (I got hitched, for those who don't know what he's talking about)
chromatic pmichaud? 18:34
jhorwitz saves a beer for Wknight8111
pmichaud == Parrot stuff
: continued working on lexicals implementation
: wrote a long message to the mailing list describing push_eh/pop_eh usage
: updated many parts of trunk to be consistent with correct push_eh/pop_eh
: still some blockers in exception handling (see mailing list for description)
== Rakudo stuff
: provided pointers and suggestions to others applying patches
EOR
chromatic particle? 18:35
particle ~ refactoring rakudo, added an 'isa' method to PCT::Node and converted rakudo to use it instead of the incorrect .WHAT or (more correct) .HOW.isa calls
~ preparing for pds2008. need to get some info, (e.g. hotel details) announced
~ i think jonathan++ may have unblocked my import patch for rakudo, will know more shortly
.end
chromatic moritz?
moritz * fixed a few issues with evalbots 18:36
* thinking about setting up a svn mirror that the IRC bots can poll as often as they want, without upsetting the svn.perl.org admins
.end
chromatic masak?
masak * Implemented one small Rakudo reature (.end)
* Fixed one small Rakudo bug (inf. loop in Object.list)
* Caused one small debate on #perl6 (due to faulty commits, tests and
behaviour in connection with junctions. still partly ongoing)
* Still looking into why r32189 breaks my Parrot.
EOR
chromatic kj?
kj == work on PIRC
* added PASM parsing capability, which was quite easy
* PIRC now can parse PIR into PASM, and sugar PASM into PIR (except for all the fancy stuff).
* implemented a register allocation optimizer, which was quite easy too :-)
* refactored some datastructures; not so easy.
eor
chromatic jonathan? 18:37
jonathan * Back from vacation, starting to dig in again
* Managed to give Coke++ a useful hint that led to a bug fix
* Today been spending some time on the Rakudo RT queue, have fixed various things and we're passing more spectests now too (by adding/unfudging as well as writing some new). Nothing in particular to speak of, mostly minor but useful fixes.
* Got flights booked for attending Parrot Developer Summit.
EOR
chromatic jhorwitz?
jhorwitz a week of debugging mod_perl6 issues proved fruitful. finally have a working implementation of Perl6Module so we can preload handlers, which speeds things up nicely.
did a bit of hacking with Tene while he was in town. mod_cardinal was born, though just a proof-of-concept.
EOR 18:38
chromatic cotto?
cotto eor
chromatic I fixed some bugs. 18:39
I also want to remove deprecated opcodes, but we need to add variants to get_*_global to do so. I have a question about that.
I also added temporary_pmc_new/temporary_pmc_free, with big hairy scary warnings on them. If you want to know when to use them, ask me. They do relieve some GC pressure in a few cases, but they have other drawbacks, like stack-allocated variables in C. 18:40
barney
?
barney Started with Pipp tests written in PHP.
Added a 'make smolder_test' target to the Pipp Makefile.
.eor
chromatic bacek?
bacek chromatic: I'm not commiter. Just rumbling around :) 18:41
chromatic I'm an includer, not a divider.
It's question time.
particle please let's officially drop gcc 2.* support, if it's not already explicit in our docs
chromatic That's not a question, but I support it. Any opposition?
pmichaud +1 18:42
Wknight8111 +1
chromatic Alright.
jhorwitz +1
chromatic Next question: any objection if I add the apparently missing get_hll_global PMC STR STR opcode? 18:43
pmichaud do we need it?
chromatic We have find_global P S S, but it's deprecated.
particle +1
pmichaud the only variant needed would be get_hll_global_p_s_s/sc
Wknight8111 +1
pmichaud currently one can do the same with a single-element array
chromatic I could add yet another rewrite in IMCC to the supported version.
... 18:44
How do the Manga say this? ^_^'
pmichaud would we also get get_global_p_s_s and get_root_global_p_s_s ?
chromatic I expect so.
pmichaud is it common enough to deserve its own opcode?
chromatic find_global P S S exists. 18:45
pmichaud sure, but that doesn't mean it's used frequently, or that it should exist.
also, find_global_p_s_s was created long before we had hll namespaces 18:46
I'm simply thinking there aren't many instances of $P0 = find_global $S1, $S2
chromatic I'm comfortable removing it if there's consensus. I just wanted to bring up the fact that it's a syntax change for PIR/PASM.
pmichaud it's currently built into imcc directly? 18:47
chromatic Not IMCC; just an op.
src/ops/var.ops:430
18:48 allison joined
pmichaud I prefer to drop it altogether, and see what breaks. But a quick check of "ack find_global" doesn't show any find_global_p_s_s variants yet 18:49
kj isn't the first _s the namespace?
find_global_p_>s<_s
chromatic Yes.
pmichaud Yes.
kj so, namespaces are represented as keys, no?
pmichaud that's one way, yes.
chromatic You'd have to ack PIR/PASM to see if anything uses it. 18:50
kj so instead of the _s version, a version with a key would be better than just a single string I think
pmichaud we already have the keyed variants of get_*_global 18:51
kj so then if we have the keyed version, we don't need the _s version right ? :-)
if we just put in the single string into [ and ]
pmichaud there's a difference between _s and _sc
kj ah yes, you're right in that 18:52
pmichaud it's easy to convert find_global_p_sc_s to its equivalent get_hll_global form.
chromatic Yes, find_global_p_s_s is harder.
pmichaud get_hll_global doesn't have an equivalent of find_global_p_s_s. However, I can't find any places where find_global_p_s_s is actually used.
jhorwitz i did this for mod_parrot with some fun regexes
pmichaud so, rather than add a new opcode (with six variants), I'd rather just eliminate find_global_p_s_s altogether until we can prove that we need it. 18:53
and since there's a relatively easy workaround for find_global_p_s_s, we should use that if we do find only one or two instances.
kj well, if it's not used (at all?) currently, then using a few ops, 2 or 3 to get the same effect would be fine I guess 18:54
chromatic I'll convert the easy ops and remove the deprecated ones, then report on how the changes look for the hard one.
pmichaud I really think find_global_p_s_s is extremely rare. I'm willing to be proven wrong, though.
kj chromatic++ # sounds like a good plan to me. 18:55
pmichaud yes, converting the existing usages of find_global is a necessary first step at any rate.
chromatic allison?
allison - Nearly finished with the core implementation of the new I/O. 18:56
- Since the old and new peacefully co-exist, I'll merge in the branch as soon as that's done, before I start converting old code to use the new system.
- Worked on MMD tickets.
- Announced BBC joining Parrot's advisory board today.
EOR
kj BBC?
ah no bot
allison kj: British Broadcasting Corporation 18:57
Wknight8111 where was that announcement?
allison www.parrot.org/news/bbc-joins-parro...sory-board
kj allison: yeah that's how I know it; I just wondered what broadcasting has to do with virtual machines...
pmichaud also on front page of www.parrot.org
kj: think media retargeting :-)
chromatic Other questions? 18:58
allison kj: they're huge users of open source programming languages
kj ah I see
particle read the press release, it should make sense then
chromatic If there's nothing else, let's adjourn until next week. 19:02
CLOSE THOSE TICKETS
19:03 jhorwitz left, pmichaud left, jonathan left, moritz left 19:05 bacek left, kj left 19:14 chromatic left 19:16 Wknight8111 left 19:24 cotto left 19:41 Coke joined
Coke belated report: fixed "./parrot -t4" this week, in an attempt to help figure out where partcl is spending its time. 19:42
19:42 Coke left 21:07 PacoLinux joined 21:09 PacoLinux left 21:31 davidfetter left 22:15 Whiteknight joined 23:05 particle left 23:27 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!