Parrot 3.6.0 "P�jaros del Caribe" released | parrot.org | Log: irclog.perlgeek.de/parrot/today | #parrotsketch meeting Tuesday 19:30 UTC
Set by moderator on 5 August 2011.
00:20 kid51 joined
kid51 Coke++ for restarting (or fixing) Smolder 00:31
dalek rrot: c3f359c | jkeenan++ | docs/book/pct/ch04_pge.pod:
Rebreak lines so as to demonstrate smart whitespace matching better. See: trac.parrot.org/parrot/ticket/2161.
 00:39
Coke msg kid51 was a permissions problem, fixed by osuosl once we pinged them about it. 00:44
aloha OK. I'll deliver the message.
00:49 cotto_work joined 01:03 whiteknight joined, daniel-s joined
whiteknight good evening, #parrot 01:05
bubaflub evening whiteknight 01:08
whiteknight hello bubaflub, how are you doing? 01:09
bubaflub whiteknight: not bad. working on my last paper for seminary. itching to just get it finished. 01:10
whiteknight: any news on the house?
whiteknight bubaflub: we went out looking last night with our old realtor, the good one, and we're feeling hopeful again 01:11
bubaflub whiteknight: that's good - i heard your other place fell through.
whiteknight yeah. 01:12
our realtor at the time was not so great, and the sellers were not being realistic with their asking price, and it just didn't work 01:13
NotFound: ping 01:15
sorear whereabouts in the world are you? 01:16
01:16 woosley joined
whiteknight sorear: who? 01:17
01:17 soh_cah_toa joined
dalek nxed/multi_syntax: 95ce9fe | Whiteknight++ | winxedst1.winxed:
Add a new 'multi' declarator to create multis.
 01:35
nxed/multi_syntax: 912bbfa | Whiteknight++ | t/advanced/10multi.t:
+tests for multi. Can define them now, still can't call them elegantly.
nxed/verbose_diagnostics: f89664f | Whiteknight++ | winxed.winxed:
Add a --verbose option to the compiler driver. --verbose=winxed outputs a winxed backtrace on failure. --verbose=pir prints out a PIR backtrace on failure. Omitting it displays no backtrace
 01:36
whiteknight msg NotFound take a look at github.com/Whiteknight/winxed/commit/95ce9feedc It does basic :multi with built-in types. I have a few changes to make. I might try to merge MultiFunctionStatement into FunctionStatement and promote to multi if we have functions with the same name 01:39
aloha OK. I'll deliver the message.
kid51 seen bacek? 01:53
aloha bacek was last seen in #parrot 70 days 15 hours ago joining the channel.
kid51 seen bacek_at_work?
aloha bacek_at_work was last seen in msg 16 days 23 hours ago <private message>.
cotto ~~ 01:58
kid51 Y'know, I have to say that I found all those smolder reports showing fails on NetBSD really annoying. 02:00
We don't have anyone who actually develops on that system, so we have no one who can diagnose/fix the failures. 02:01
But this box on the compile farm generates more Smolders than any other source.
bubaflub kid51: can NetBSD fit into a VM like VirtualBox? maybe we could coax someone into installing it and trying it out 02:05
s/fit into/install in/ 02:06
kid51 bubaflub: I suspect so, but I have no personal experience with Virtualbox.
So don't trust anything I say about that ;-)
My feeling is that if we know a particular machine is going to report failures, and if we can't either work on that machine or its equivalent, then maybe we only need one report per 24 hours from that box -- rather than one each commit (which is what it seems like). 02:08
dalek kudo/nom: 0b90bc4 | Coke++ | t/spectest.data:
track failure modes
 02:11
Coke kid51: just TODO the failing tests for that particular OS. 02:12
that seems to be the appropriate thing if no one has the tuits to fix the test.
kid51 The failures are only coming when it's configured with "--without-threads" 02:29
The same box passes without that option.
Well, too late tonight for me to think clearly about this any more. 02:30
dalek rrot/soh-cah-toa/hbdb: 5626ad9 | soh_cah_toa++ | t/tools/hbdb/cmds.t:
Added test for one argument form of 'break' command.
 02:47
rrot/soh-cah-toa/hbdb: a8277f2 | soh_cah_toa++ | src/runcore/cores.c:
Commented out statement in runcore that displayed PC value since it's just a stub.
rrot/soh-cah-toa/hbdb: e3e3d9a | soh_cah_toa++ | t/tools/hbdb/cmds.t:
Added test for no argument form of 'run' command.
rrot/soh-cah-toa/hbdb: 43817e3 | soh_cah_toa++ | t/tools/hbdb/cmds.t:
Added test for no argument form of 'step' command.
kudo/nom: a8f8222 | Coke++ | t/spectest.data:
run fudged tests.
 02:51
03:54 theory joined
soh_cah_toa msg cotto i got the tests done but some commands like 'continue' can't have a test b/c i can still only run one command at a time. i see what you were talking about now. :\\ anyway, now i need your help on how to approach watchpoints. we'll talk tomorrow 04:00
aloha OK. I'll deliver the message.
soh_cah_toa msg whiteknight if you got some time tomorrow, i could use your help on designing how watchpoints should work. thanks :)
aloha OK. I'll deliver the message.
cotto d'oh 04:07
04:24 worr joined 04:55 preflex_ joined 05:24 SHODAN joined 06:04 SHODAN joined 06:56 jsut joined 07:04 Eclesia joined
Eclesia hi 07:04
07:05 woosley1 joined 08:08 woosley joined 08:16 woosley joined 09:14 woosley joined 10:14 M_o_C joined 10:59 mj41 joined
dalek kudo/nom: cd81901 | moritz++ | / (5 files):
Bring back Perl 6 level stack traces

There are still a lot of quirks to be worked out, for example some routines magically not appearing in the stack trace, and other showing up that I would prefer to keep out (like &die), but I believe it is a net improvement. The old PIR backtraces can be obtain with the --ll-exception command line argument.		 11:23
kudo/nom: d06de50 | moritz++ | NOMMAP.markdown:
update NOMMAP
11:58 lucian joined 12:07 contingencyplan joined 12:18 jsut_ joined 12:23 whiteknight joined
whiteknight good morning #parrot 12:23
moritz good morning whiteknight
whiteknight hello moritz 12:27
12:31 Kulag joined 12:49 JimmyZ joined
dalek nxed: 05d6699 | NotFound++ | winxedst1.winxed:
change casting to var semantic when symbols are know at compile time
 12:49
12:53 bluescreen joined
whiteknight NotFound: ping 12:53
NotFound whiteknight: pong
12:53 JimmyZ joined
whiteknight NotFound: I've been working on the multi patch for winxed yesterday and this morning. I will probably open a pull request for it tonight 12:53
it's much nicer than what I msg'd you last nigh 12:54
NotFound whiteknight: nice, but I'll probably don't have enough time to review and test it in time for the release. 12:55
bubaflub ~
whiteknight NotFound: that's okay. No rush
12:55 JimmyZ joined
whiteknight oh wow, the release is this tuesday, isn't it? 12:56
I didn't know it was so close
12:56 JimmyZ joined 13:01 JimmyZ joined
NotFound Yes, and this weekend I'll be on Fiesta! 13:02
whiteknight Fiesta? Sounds fun
NotFound I hope so.
whiteknight I could have several pull requests waiting for you when you get back :) 13:05
I think some will start very good discussions 13:06
github.com/Whiteknight/winxed/network Lots of activity here!
13:07 JimmyZ joined
dalek kudo/nom: ab5fbf9 | moritz++ | src/Perl6/ (2 files):
fix s///
 13:07
NotFound Nice graphs 13:12
whiteknight I have a few other features I want to play with too, and things on my laptop from this morning that I haven't pushed 13:14
tadzik going to yapc::eu?
aloha: going to yapc::eu?
aloha tadzik: going to yapc::eu is cotto or jnthn or szabgab or Maddingue or TimToady or rurban or pmichaud or tadzik
whiteknight aloha going to yapc::na?
aloha whiteknight: going to yapc::na is cotto or dukeleto or atrodo or Util or kid51 or benabik or soh_cah_toa
whiteknight er, eu 13:15
moritz yapceurope.lv/ye2011/talk/3388 has an interesting selection of people going to YAPC::EU :-)
NotFound I've been a bit dsitracted from parrot, writing a media player for Android, 13:16
whiteknight NotFound: oh, cool. Is it open-source?
NotFound Probably, for a now is just in my devel environment and in my tablet for testing. 13:17
The idea is being easy to use while driving without distracting. 13:18
Two easy gestures for next/previous, two other for volume up/down. 13:19
whiteknight oh cool 13:20
NotFound Playing is ready, now it needs form to prepare playlist et al. 13:21
whiteknight what is the name of the player? "Media Not Found!" 13:40
NotFound Vidente 13:49
whiteknight that's good too, but not as funny 13:51
JimmyZ Media Player is Not Found! 13:57
13:58 JimmyZ left, JimmyZ joined
dalek rrot: 6a0599e | Whiteknight++ | src/pmc/exception.pmc:
+docs for Exception.backtrace_strings. TT #2175
 14:21
14:21 JimmyZ joined
dalek TT #2175 closed by whiteknight++: method backtrace_strings in PMC Exception not documented 14:24
TT #2175: trac.parrot.org/parrot/ticket/2175
Coke (release) Uuuuugh. I thought the 3rd tuesday was later also.
So, yes, release is next tuesday. 14:27
whiteknight Coke: you're grandmaster of those proceedings, right? 14:31
Coke Call me flash. 14:32
moritz flash!
Coke <glare> 14:33
atrodo Ah aah 14:38
moritz tests nom on latest parrot
atrodo (Looks like Queen songs don't translate well)
moritz calls Coke the saviour of the universe 14:39
Coke will roll with the queen theme song to flash if you acknowledge grandmaster flash. :P 14:47
atrodo All hail grandmaster flash Coke! 14:51
Coke great, now I just feel old.
anyway, release is tuesday. I want those test failures from netbsd to be todo'd before then.
(they've been failing for so long, we're obviously not going to be able to fix them just yet.)\\ 14:52
meetings.
moritz who's the release manager?
14:52 JimmyZ joined 14:54 M_o_C joined
moritz so far the nom spectest on newest parrot look good. 14:54
though only just in S04, a bit early to celebrate :-)
Coke moritz: for the next release, it's me 15:04
dalek TT #2156 closed by coke++: Relocate Python Scripts in "blib" Directory 15:12
TT #2156: trac.parrot.org/parrot/ticket/2156
15:24 Eclesia left
cotto ~~ 15:26
15:34 not_gerd joined
not_gerd hello #parrot 15:34
cotto: have you seen gerdr.github.com/on-parrot/rethinking-m0.html ? 15:37
15:38 alester joined
benabik o/ 15:38
cotto not_gerd, I will very soon have seen it. ;) 15:39
cotto sets xchat to blink on "m0" 15:40
15:44 dmalcolm joined
moritz not_gerd: if m0+ doesn't have strings, we need *another* low level-ish format that has strings 15:52
not_gerd: I'm afraid we might be stacking up too many low level formats 15:53
not_gerd moritz: no assembly language I know of has the comcepts of strings
whiteknight not_gerd: Parrot isn't like any other hardware machine 15:54
moritz not_gerd: it's not parrots job to provide an assembly language; it's parrot's job to provide a platform for easily implementing dyanmic languages 15:55
cotto I see some charm in having a layer that's closer to hardware. The idea of m0+ is a pretty significant change, but I've been wondering what m0 would look like if I was more familiar with assembly.
moritz not_gerd: an assembly language is just a tool for that job
not_gerd moritz: that's what PIR/Mole is for - users don't write m0
moritz not_gerd: and what does compilers emit? 15:56
not_gerd moritz: I see what you're getting at... 15:57
however, you can have string ops without having a string type
benabik Correct me if I'm wrong, but I thought M0 is replacing C not PIR. M0 is what we'll use for opcodes and internals. PIR is what our HLL compilers will continue to emit.
And in M0 strings are yet another PMC.
not_gerd benabik: for efficiency reasons, HLLs will probably want to output bytecode files themselves 15:58
without going through PIT, that is
^PIR
cotto not_gerd, what kind of complexity would be needed for the translation from m0+ to m0-? 15:59
benabik Yes, but PBC bytecode, not M0 bytecode. HLLs will want to continue to use the wide variety of opcodes we have at our current level. 16:00
not_gerd cotto: if you don't care for optimizations, translation should be easy (for common ops) 16:01
16:01 theory joined
not_gerd I choose two m0- registers for each type so each maps to one m0+ opßerand 16:02
cotto not_gerd, are you going to yapc::eu by chance? 16:03
not_gerd cotto: no, sorry
cotto not_gerd, how far along is your concept of m0+/-? 16:06
not_gerd what you just read is basically it 16:08
I really like the idea of bootstrapping via LLVM, but that may not be feasible
cotto not_gerd, can you expand on the problems with M0 that your post tries to address? 16:09
not_gerd as it is now, m0 is not suited for either interpretation or jitting (more) 16:11
you can easily implement interpreters for minimal instruction sets, but no fast interpreters
you can easily implement jitters for a minimal instruction set (in particular, one that is lower-level than the target instruction set), but you won't produce good code without an optimizer 16:12
expanding a high-level instruction to several lower-level ones is in general easier than joining them again
cotto not_gerd, thanks for articulating that. Those are valid concerns. 16:14
not_gerd perhaps another point to clarify why I introduced m0- in addition to m0+: 16:17
16:18 jsut joined
not_gerd implementing an m0+ interpreter would probably lead to a lot of duplicate code, which will probably mess with cache performance 16:18
I'm no expert on that, though
the reason why I singled out LLVM as a targetinstruction set is because you can do this: gist.github.com/1142401 16:25
ie writing fast interpreters and still be portable 16:26
if that goal interferes too heavily with the other features, another approach would be a Mole to LLVM IR compiler 16:27
cotto efficient execution of M0 is important 16:37
Coke hey, does anyone want to take over the release this tuesday? 16:48
I would definitely appreciate it. 16:49
16:56 ambs joined 17:02 theory joined
whiteknight Coke: I can probably do it. You busy? 17:02
not_gerd moritz: restrings in m0+, see github.com/gerdr/on-parrot/commit/...ent-528683 17:28
17:47 JimmyZ joined 17:58 worr joined
not_gerd bye, #parrot 18:03
atrodo aloha: msg not_gerd Before m0 was my prototype github.com/atrodo/lorito . If you had the time, I'm curious to chat with you about it 18:04
aloha atrodo: OK. I'll deliver the message.
Coke whiteknight: I wasn anticipating a difficult which didn't seem to materialize. Nevermind for now. Thanks. 18:31
*was
dalek kudo/nom: 6e22b91 | moritz++ | src/core/BackTrace.pm:
refactor BackTrace a bit, hide non-Routines in the setting by default
 18:32
18:38 theory joined
dalek kudo/nom: 537bd91 | Coke++ | tools/update_passing_test_data.pl:
remove unused import, var.
 18:45
18:48 whiteknight joined
cotto goes to yurp 19:02
whiteknight goes derp 19:16
19:27 soh_cah_toa joined
soh_cah_toa whiteknight: hey hey 19:29
whiteknight soh_cah_toa: what the dill, yo?
soh_cah_toa whiteknight: i'm pretty much done picking out bits and pieces from dwarf. now i could use some help in super reviewing the spec to see how an implementation would be possible 19:30
whiteknight soh_cah_toa: okay, I can take a look at it tonightish 19:31
soh_cah_toa yeah sure
whiteknight What I saw earlier was looking pretty good 19:32
soh_cah_toa i'm really excited about it. i think this could really open up some new doors and possibilities for analysis tools for parrot
profiling as well
19:33 Eclesia joined
Eclesia hi 19:33
soh_cah_toa we could integreate it into "hbdb2" w/ a 'profile' command. something like 'profile sample' for pc sampling or 'profile instrument' for runtime instrumentation
Eclesia: howdy
19:38 dmalcolm_ joined 19:45 dmalcolm__ joined 20:24 theory joined, jsut_ joined
dalek kudo/nom: d3fbdcb | Coke++ | t/spectest.data:
run this fudged test.
 20:28
Tene_ [A 20:40
20:40 bluescreen joined
Tene_ Anyone here going to PPW this year? 20:40
soh_cah_toa Tene_: pittsburgh perl workshop? 20:41
Tene_ Yes.
soh_cah_toa Tene_: yeah, i probably will be there
dalek Heuristic branch merge: pushed 20 commits to lalrskate by ekiru 20:49
tcurtis ~~ 20:52
Eclesia ppw ? 20:55
Tene_ Eclesia: Pittsburgh Perl workshop 20:56
Eclesia hm 20:58
20:59 bluescreen joined 21:13 Eclesia joined 21:44 sjn left 21:45 Psyche^ joined 22:06 lucian_ joined 22:18 cotto joined
cotto ~~ 22:22
22:22 kid51 joined 22:25 Kulag joined 22:47 Limbic_Region joined 22:56 whiteknight joined 22:59 cotto joined
whiteknight good evening, #parrot 23:00
dalek nxed/multi_syntax: 69759bb | Whiteknight++ | winxedst1.winxed:
Start merging MultiFunctionStatement into FunctionStatement
 23:07
nxed/multi_syntax: 0d59163 | Whiteknight++ | winxedst1.winxed:
Remove the multi keyword, and fold multi logic into FunctionStatement. Right now, only works in namespaces
nxed/multi_syntax: 250bbe7 | Whiteknight++ | winxedst1.winxed:
Add multi funcitionality to class methods.
whiteknight With that patch, we have basic multidispatch in Winxed
well, once it merges. I need tests first
23:21 jsut joined
Tene_ nice 23:22
dalek nxed/multi_syntax: b239fab | Whiteknight++ | t/advanced/10multi.t:
update multi tests
 23:27
23:32 jsut_ joined 23:34 Coke joined 23:46 worr joined
dalek nxed/verbose_diagnostics: d79419f | Whiteknight++ | winxed.winxed:
Use load_bytecode_p_s instead of the deprecated load_bytecode_s
 23:52
nxed/verbose_diagnostics: e28a288 | Whiteknight++ | winxed.winxed:
rename --verbose to --err, to be clearer and more similar to other options
 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!