Parrot 4.0.0 "Hyperstasis" | parrot.org | Log: irclog.perlgeek.de/parrot | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 12 February 2012.
dalek kudo/macros2: 39ff0b2 | masak++ | src/ (4 files):
partial commit

This is really close to working -- the code in this commit sets the OUTER of the incarnated quasi to the correct context.
But this code still doesn't work, and we really expect it to:
  ./perl6 -e 'macro foo { my $a = "OH HAI"; quasi { say $a } }; foo'		 00:19
00:52 lucian joined 01:36 Maddingue joined 01:55 Maddingue joined 04:14 cogno joined 05:29 Kulag joined 06:55 jsut joined 08:17 fperrad joined 09:44 jsut_ joined 10:02 mj41 joined 10:39 lucian joined 11:07 lateau joined
dalek nxed: 15955f0 | NotFound++ | .travis.yml:
delete non-working travis file
 11:11
nxed: 3fda23c | NotFound++ | config/ (2 files):
delete unused old configure experiment
 11:21
nxed: bde46b5 | NotFound++ | t/advanced/06defaultvalues.t:
minimal testing of default values
 11:29
11:35 lucian joined 11:59 nopaste joined 12:25 jsut joined 12:33 whiteknight joined 12:50 Util joined
whiteknight good morning, #parrot 12:52
dalek kudo/macros2: e4552e2 | masak++ | src/Perl6/Actions.pm:
[Perl6::Actions] fix lexical issues

  ...this is the other half of the partial commit from yesterday.
Fixes two problems:
  - Arranges so that the injected code doesn't get a static outer
   set at compile time, overriding the one we want for it.
  - Fixes a strange discrepancy where ordinary variables were
   recognized from inside the quasi block, but macro parameters
   weren't.		 12:53
kudo/nom: ac5d99f | jnthn++ | docs/ChangeLog:
A few more ChangeLog entries.
 13:05
nine Good morning, whiteknight! 13:22
whiteknight hello nine. How are you doing today?
nine Apart from some weird feeling in my throat fine. Trying to get some pages done today. You? 13:23
Btw. niner.name/Hybrid_Threads_for_the_Parrot_VM.pdf if you want to have a peek. 13:28
whiteknight: you PMed me some time ago about concurrency-enabled streams and feeds. I'd very much like to hear about that.
whiteknight I can't remember who, one of the p6 guys was pushing that idea 13:30
the general idea being that we have a stream type which contains a data buffer and a generator thread. When the buffer is empty or has space, the generator is activated to fill it 13:31
when the buffer is full, the generator is put on hold
nine message said ruoso was it
whiteknight Ah, ruoso. That's the one
nine Sounds like all that's needed on the parrot side is a flag on the Task to allow to exclude it from scheduling. Which sounds like a good idea anyway. 13:33
13:35 PacoAir joined
tadzik nine: you're doing your bachelor work right now; bachelor is after how many years of studying? 13:48
nine tadzik: three years
tadzik so that's quite equivalent to Polish «engineer»
which I should probably start to look forward to
nine tadzik: how far are you into your studies? 13:53
NotFound whiteknight: Do you want to update the winxed snapshot for the release? Aren't we in feature freezing? 13:55
tadzik nine: I'm now starting my third year 14:00
nine So it would be about time to pick a topic 14:03
tadzik slowly, yes. I should be done with my works after 3.5 years
so 3 semesters to go 14:04
masak places a callback on tadzik: { say "we're hiring!" }
:) 14:05
tadzik yes, I remember about you :)
whiteknight NotFound: I would like the snapshot updated, yes 14:06
tadzik I'll see what Monday will change though :) 14:07
whiteknight It's only a feature freeze for core. You are in control of Winxed freezes
tadzik masak: also, I'd probably stay for another 1.5 years for the magister degree
masak makes sense.
NotFound whiteknight: ok, I'll try to do it.
tadzik nine: typo, "Parrto compiler toolkit" 14:08
nine fixed 14:10
My biggest problem with this whole paper is that I only have a very vague idea what a reader would be interested in. So since part of my motivation is the hope that much of what I write can be included into Parrot's documentation, I'd like to ask: what do you as Parrot users/developers want to read about? 14:29
14:40 lateau__ joined 15:44 Psyche^ joined 16:19 lucian joined 18:24 lucian joined 18:40 contingencyplan joined, contingencyplan_ joined
dalek p/bs: 11f08d2 | jnthn++ | src/core/NQPRoutine.pm:
Update dispatcher derevation so it clones the things it should.
 19:23
p/bs: 7500b8a | jnthn++ | src/ (3 files):
Ensure we do the required attachment of code objects to code refs.
p/bs: 3169880 | jnthn++ | src/ (2 files):
Last couple of fixes needed to get new multi-method stuff passing tests again.
p/bs: 985e7d1 | jnthn++ | src/NQP/World.pm:
Fix a thinko; gets us much further into deserializing NQPCORE.setting.
p/bs: 9cc8b4a | jnthn++ | src/6model/serialization.c:
Fix some issues in type object deserialization.
p/bs: 1b41ba7 | jnthn++ | src/NQP/World.pm:
Add a missing $NEW_SER check, which gets us a little further into the build.
19:35 jsut_ joined 20:03 mj41 joined 20:22 perlite joined
dalek kudo/sink: c25fcb8 | moritz++ | /:
Merge branch 'sink', remote branch 'origin/sink' into sink
 20:22
kudo/sink: 8274182 | moritz++ | src/Perl6/Sinker.pm:
[sinker] recurse into nested blocks
p/bs: ba5e046 | jnthn++ | src/6model/serialization.c:
Improve an error to aid debugging.
 20:32
p/bs: 7876241 | jnthn++ | src/Regex.pir:
Pop a couple of things from the PIR-based Regex engine into an SC, so we don't try and serialize things from here.
p/bs: 7c96cef | jnthn++ | src/6model/serialization.c:
Another error tweak.
p/bs: 8768530 | jnthn++ | src/6model/serialization.c:
On deserialization, make sure an SCs static code refs are all correctly annotated.
p/bs: bacf8fb | jnthn++ | src/Regex (4 files):
More SC fakery for the old PIR-based regex engine, so the serializer doesn't get upset over it (it's all throwaway once we bootstrap with QRegex anyway). With this, we now complete the build of NQP, with serialization/deserialization of compilation units along the way. The final executable doesn't actually work just yet, though the immediate cause is a known NYI. \\o/
kudo/sink: 5705951 | moritz++ | src/core/traits.pm:
remove a workaround in "is export" trait. Now we can load Test.pm, but the underlying problem is not fixed
 20:39
20:49 zby_home_ joined 20:52 zby_home joined
whiteknight After the last firefox update, it appears the keyboard shortcuts don't work anymore 21:05
Ctrl+t, Ctrl+w, Ctrl+PgUp, Ctrl+PgDn are essential to my productivity and my mental well-being
21:08 plobsing joined 21:44 awwaiid joined
dalek nxed/fix_installed_driver: ee3bbc1 | Whiteknight++ | winxed_installed.winxed:
Fix installed driver after update to use PackfileView PMC in .compile
 21:50
p/bs: 83d190d | jnthn++ | src/6model/serialization.c:
Serialize mode flags.
 21:52
whiteknight msg NotFound github.com/NotFound/winxed/pull/19 It's an emergency fix
dalek p/bs: 96892f6 | jnthn++ | src/NQP/World.pm:
We don't want to serialize the Parrot-specific vtable mapping stuff, so make sure we fix it up after deserialization.
aloha OK. I'll deliver the message.
p/bs: 067d904 | jnthn++ | src/6model/serialization.c:
Serialize/deserialize boolification spec. Now we get an NQP executable that basically works.
p/bs: 25e15e6 | jnthn++ | src/NQP/World.pm:
Turn on serializer by default now. Some tests fail...triage time.
21:59 plobsing joined, mj41 joined, Patterner joined, nopaste joined, benabik joined, TonyC joined, wagle joined, jtpalmer joined, cxreg joined, cotto joined, kshannon joined, allison joined, atrodo joined, hernan joined, arnsholt joined, Timbus joined, Tene joined, janus joined, knewt joined, GeJ joined, athomason joined, ingy joined
dalek p/bs: 08757d6 | jnthn++ | src/NQP/World.pm:
Various fixes to handling of code refs. NQP with the serializer now passes all of t\\nqp. \\o/ Doesn't mean were there yet, mind; role serialization is still busted, meaning that QRegex can't be compiled. But...closer.
 22:15
p/bs: bbe4085 | jnthn++ | src/NQP/World.pm:
Remove option to use the serializer (it's always on now) and all old deserialization code gen for NQP, which muchly clears up World.pm.
 22:28
p/bs: 790fc9a | jnthn++ | src/6model/serialization.c:
Use a seen hash to avoid duplicate strings in the string heap.
 22:56

Powered by Cro and the Raku® Programming Language.

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

Thank you!