#parrot Parrot 2.2.0 "Like Clockwork" Released! | parrot.org/ | Channel log: irclog.perlgeek.de/parrot/today | Priority: fix line number annotations | Finish GSoC applications
Set by moderator on 6 April 2010.
darbelo plobsing: It's not a matter of reconfiguration. 00:03
gen::config_pm - Record configuration data for later retrieval...Use of uninitialized value in printf at config/gen/config_pm.pm line 121, <$IN> line 19.
Use of uninitialized value in printf at config/gen/config_pm.pm line 121, <$IN> line 19.
00:05 tcurtis joined
darbelo It fails for me on a clean chekout. 00:05
cotto_work config_lib.pir needs to depend on datatypes.pasm 00:11
darbelo Putting it as a dep of config.fpmc should be enough. 00:14
cotto_work sure 00:17
I should have said that 00:18
seen dukeleto 00:27
purl dukeleto was last seen on #parrot 3 hours, 3 minutes and 32 seconds ago, saying: Whiteknight: pong, with a one foot out the door
cotto_work I guess he managed to get the other foot out too
00:30 davidfetter joined
plobsing how did I let that slip by? 00:41
darbelo A working Makefile contains at least 15% of luck. 00:42
cotto_work Do you usually do a parallel build? That covered up the problem when I tried it the first time. 00:43
plobsing I always use parallel make - so impatient 00:45
dalek rtcl-nqp: 10a88cc | Coke++ | t/cmd_linsert.t:
the parser error here was trumping the index check
 00:53
rtcl-nqp: 9eaaf7a | Coke++ | (4 files):
created <quoted_word> to be shared by <list>

  - untodo all related tests
rtcl-nqp: 294435d | Coke++ | (3 files):
When constructing lists, \\{ are special only when counting braces,
00:54 abqar joined 00:58 ascent joined 00:59 khairul joined
dalek rrot: r45624 | plobsing++ | trunk/config/gen (2 files):
fix build
 00:59
Coke fights with NQP more. 01:28
01:38 bacek_at_work joined 02:06 theory joined
dukeleto cotto_work: pong 02:43
02:48 janus joined 02:50 mariano joined
cotto dukeleto, could you make me the mentor for khairul's proposal and make whiteknight the mentor for the nfg proposal? 02:53
dukeleto cotto: ok 02:54
cotto thanks 02:55
dukeleto cotto: done 03:00
cotto Could you take a gander at my response to the Perl 6/dlr proposal and say if you think it's fine by gsoc's rules?
sorear dlr? 03:01
purl dlr is, like, fuuuuuuuugly or Microsoft's Dynamic Language Runtime - that on which IronPython and IronRuby are based or Docklands Light Railway
04:08 Andy joined 04:12 bubaflub joined 04:34 JimmyZ joined
dalek rtcl-nqp: 00d24f3 | Coke++ | (3 files):
These test files complete - run them, TODO failing tests.
 04:35
rtcl-nqp: 3fa5cd7 | Coke++ | src/Partcl/commands/main.pm:
add very basic args check to [apply]
rtcl-nqp: 07b1f3f | Coke++ | src/Partcl/commands/main.pm:
make [split] always return a list.
rtcl-nqp: 290c08d | Coke++ | (11 files):
All these tests at least parse now, run them all, TODO'ing where needed.
Coke msg austin - I think I just created some LHF in partcl-nqp - (ack TODO t/*.t) 04:38
purl Message for austin stored.
04:48 snarkyboojum joined
dalek rtcl-nqp: 88b3e3c | Coke++ | (4 files):
split all list-related commands into a new file.
 05:03
rtcl-nqp: c3b8270 | Coke++ | (4 files):
move all commands that generate control exceptions into a separate NQP file.
05:09 davidfetter joined 05:11 szabgab joined 05:46 fperrad joined 06:00 fperrad_ joined 06:01 uniejo joined
fperrad ping bacek 06:08
06:25 aukjan joined 06:29 payload joined 06:49 payload1 joined 06:53 chromatic joined, tcurtis joined 07:01 payload joined
mikehh make smoke/smolder_test has a requirement for config_pasm which has been removed 07:26
extra_file config_lib.pasm does not exist! at t/harness line 78 07:27
07:31 iblechbot joined 08:15 khairul joined 08:24 clinton joined
dalek rrot: r45625 | mikehh++ | trunk/t/harness:
change extra file sent to smolder to be config_lib.pir not config_lib.pasm
 08:27
purl dalek: that doesn't look right
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#33143), fulltest) at r45625 - Ubuntu 10.04 beta amd64 (g++ with --optimize) 08:30
chromatic msg plobsing My measurements show startup as 5.711% faster at r45626, with 2.785% of that r45626 and 2.926% of that r45623. Nice work. 08:35
purl Message for plobsing stored.
mikehh ha - forgot to change the copyright again :-{ 08:41
szbalint those 0.95**n's are starting to stack up :) 08:43
moritz purl: 0.95**4 08:44
purl 0.81450625
dalek rrot: r45626 | chromatic++ | trunk (6 files):
[encoding] Added ->hash member to ENCODING struct. This allows encodings to

as they know the width of the characters and don't have to dance with string iterators. This speeds up startup by 2.785% and improves runtime performance. Note that currently only the fixed_8 encoding has a hashing function; other encodings could use them for similar speedups.
chromatic Still 5.36 times slower than Perl 5.10.0's "Hello, world!" but ....
09:04 sorear joined
moritz I've skimmed the "instrumentation tools for parrot" gsoc proposal 09:36
and so far I'm not quite sure what this achieves for us that that the profiling run core can not
could anybody please explain that?
chromatic After I sleep I can.
moritz ok, no hurry :-) 09:37
dalek kudo: cb45216 | (Solomon Foster)++ | src/core/Real.pm:
Fix Real.abs and Real.sign to use named invocants.
 09:59
rrot: r45627 | fperrad++ | trunk (3 files):
[TAP] refactor with iterator
 10:06
bacek fperrad, pong 10:14
(o hai)
fperrad hi bacek, this morning TT#1550 was a problem for me, but now I find a workaround 10:16
bacek fperrad, interesting. Sub.clone looks pretty ok 10:17
fperrad, can you create todoed test for it? 10:24
fperrad bacek, ok 10:25
bacek fperrad, sigh... Problem is: we resolve "lexical chain" (from inner to outer) using pointers. And looks like we are filing to resolve it properly for cloned subs. 10:30
yeah. 10:31
darbelo bacek: ping
bacek fperrad, Sub.lex_info created lazily on first invoke. That's why it's not cloned properly 10:32
darbelo, what do ya want? :)
darbelo bacek: Does nopaste.snit.ch/20255 look right to you? It's ucs2 version of r45626 10:33
bacek darbelo, DRY failure makes me cry... 10:35
darbelo It isn't my fault we have a godzillin encodings.
bacek it's your fault to not submit GSoC project to get rid of current strings encoding/charset nonsense! :) 10:37
darbelo I did, it's right there at the end. Just after adding another encoding ;) 10:38
bacek (otherwise looks ok)
dalek rrot: r45628 | bacek++ | trunk/src/ops/var.ops:
Properly check for NULL result in find_lex op.
 10:39
bacek If you can create C++ templa^W^W "plain C macro" for calculating string's hash value it will be great. Preferably based on murmur hash.
darbelo Won't a 'plain C macro' FAIL with utf8's multy-byte nonsense^Wshit^Wstuff? 10:41
bacek no. Why? 10:43
purl no. Why are you quoting it at all?
darbelo We can't just *(pos++) in utf8 some chars take up two slots. 10:44
bacek partial template specialization^W^W^W 10:45
shit
yes, you have to supply "pos_increment_func" macro arg ( or something like this) 10:46
10:48 ruoso joined
darbelo bacek: I'm seeing failures in t/op/lexicals.t was that you? 11:11
bacek darbelo, erm... May be
darbelo nopaste.snit.ch/20256 11:12
bacek darbelo, it's definitely caused by my commit. But it's just exposed other problems. (Most likely bugs in tests) 11:15
darbelo The "doesn't match '/Null PMC access/" thing makes the tests suspicious.
I ussualy read those things as "OH NOES, OUR FAIL IS LESS EPIC LIKE THIS! TEH HORRORS!" 11:17
bacek :) 11:18
#8 still has other bug 11:19
darbelo The test 8 failure looks a bit more legit.
11:26 szabgab joined 11:36 allison joined
dalek rrot: r45629 | fperrad++ | trunk/t/op/lexicals.t:
add a todoed test for TT#1550
 11:44
11:58 whiteknight joined
dalek rrot: r45630 | darbelo++ | trunk/src/string/encoding/ucs2.c:
Add a hashing function for ucs2 strings. People with a DRY reflex whould not look at this and r45626 at the same time.
 12:00
whiteknight good morning, #parrot 12:03
mikehh hi there whiteknight
whiteknight hello mikehh, how are you today?
mikehh doin' good and you
whiteknight pretty good, though it's still early 12:05
12:05 payload joined 12:08 Mokurai joined 12:39 tetragon joined
mikehh bacek, darbelo - t/op/lexicals.t has some horrible tests 13:05
13:06 Andy joined
mikehh expecting /Null PMC access/ in test 32 and 39 is not good and now FAILS 13:08
as well as test 8
darbelo #8 is probably a legit failure. The others look like the error message changed. 13:10
Parrot_io_make_string(), I hate you. 13:11
13:20 snarkyboojum joined, muixirt joined
allison msg Coke you should have admin rights on /languages, since your user is part of the "admin" group. But, there were some permission problems with the auth files for /languages on the osuosl server. I've fixed them now (hopefully) so could you try it out and make sure? 13:32
purl Message for coke stored.
allison msg Coke I like the new twitter box on parrot.org, it fits in with the whole page nicely. 13:33
purl Message for coke stored.
13:39 Andy joined 13:52 davidfetter joined 13:57 iblechbot joined 13:58 patspam joined
Coke ponders bumping people off the mailing list with full mailboxes. 13:59
13:59 payload joined, plobsing joined
Coke allison: I can get into the admin screens now. 14:07
and apparantly I have a commit bit on languages.
darbelo Why is the code in src/io/ constructing it's own strings? 14:09
whiteknight darbelo: because it knew you were coming 14:11
darbelo: I think we all know by now that the strings code is sub-optimal
darbelo But this is *io*. I mean I know what to expect when I see src/strings/ in the path to a file. 14:13
whiteknight oh, shoot. I misread that.
that's weird. What is the IO system doing creating it's own strings?
Coke src/string/api.c:1455: warning: ignoring return value of ?Parrot_str_length?, declared with attribute warn_unused_result 14:14
If we are shooting for a warnings free build, it would be nice if we didn't have so many warnings. =-)
whiteknight Coke: maybe we need to lower our expectations 14:15
new goal: an error-free build
...and success!! 14:16
darbelo Ocasionally.
Coke: I have other changes in my working copy, but nopaste.snit.ch/20257 should fix that warning. 14:18
davidfetter oh hai
14:19 theory joined
darbelo Heck, it's even an optimization. 14:20
dalek tracwiki: v20 | Jack9++ | JITRewrite 14:21
tracwiki: trac.parrot.org/parrot/wiki/JITRewr...ction=diff
darbelo Oh look! WikiSpam.
14:21 Andy joined 14:25 atrodo joined
Coke killing his account... 14:27
whiteknight Coke: already done
deleted the edit and killed the account
Coke whiteknight: did you add his email to the list of spammers? 14:28
whiteknight no. There's a list?
Coke trac.parrot.org/parrot/wiki/TracSpammers
whiteknight done and done 14:29
14:33 bubaflub joined
dalek tracwiki: v17 | Jack9++ | BuildTaskList 14:37
tracwiki: trac.parrot.org/parrot/wiki/BuildTa...ction=diff
tracwiki: v9 | whiteknight++ | TracSpammers
tracwiki: trac.parrot.org/parrot/wiki/TracSpa...ction=diff
particle Jack9-- 14:38
dalek rrot: r45631 | darbelo++ | trunk/src/string/api.c:
Quiet a warning.
particle Jack9--
rrot: r45632 | darbelo++ | trunk/src/io/utf8.c:
Replace some direct manipulation of STRING structures with a function call from the API.
particle gotta get the karma right....
darbelo particle++
:)
Coke karma Jack9 14:52
purl jack9 has neutral karma
Coke Jack9--
dalek TT #1555 created by coke++: t/op/lexicals.t failures 15:10
TT #1555: trac.parrot.org/parrot/ticket/1555
rrot: r45633 | fperrad++ | trunk/runtime/parrot/library/TAP (2 files):
[TAP] handle args/options
 15:11
rrot: r45634 | fperrad++ | trunk (4 files):
[TAP] add an initial Tapir, an executable over TAP/Harness (like prove)
Andy Installing gcc 4.4.3 15:13
hoping It'll turn up more fun arnings
warnings
purl rumour has it warnings is just a magical bitmask, iirc
15:19 whiteknight joined
whiteknight goddamn windows just gave me a bluescreen and immediately rebooted 15:20
do not pass go, do not save your work, just shut the hell down immediately
Coke I'd mock you but that happened to me on OS X a few weeks ago. 15:25
of course, the screen wasn't blue, it was a lovely shade of gray with a very nice font. 15:26
dukeleto: ping 15:27
15:30 whiteknight joined, atrodo joined, payload joined, davidfetter joined, snarkyboojum joined, allison joined, szabgab joined, clinton joined, janus joined, ascent joined, preflex joined, dngor joined, elmex joined, tewk joined, darbelo joined, nopaste joined, mikehh joined, eternaleye joined, Chandon joined, eiro_ joined, knewt joined, s1n joined, magnachef joined, PerlJam joined, mj41_ joined, Khisanth joined, integral joined, pjcj joined, jjore joined, pmichaud joined, Util joined, slavorg joined, sjn joined, jhelwig joined, ttbot joined, zibri joined, moritz joined
moritz sounds like a painful experience 15:30
dalek kudo: 3b5a790 | moritz++ | src/core/Any-list.pm:
make the grep() arguments slurpy
 15:31
15:32 ruoso joined 15:50 whiteknight joined, payload joined, davidfetter joined, snarkyboojum joined, szabgab joined, preflex joined, elmex joined, tewk joined, darbelo joined, Chandon joined, eiro_ joined
darbelo Hm, do we have any io experts? 15:58
16:07 aukjan1 joined
Andy Sorry, my expertise is on Ganymede and Callisto. 16:09
dalek rrot: r45635 | darbelo++ | trunk/src/io/api.c:
There is little point to instantiating a string indirectly through a function filled with conditional aiming to determine information we laready know. More cleanup than optimization, but still worth it.
 16:17
Coke darbelo: do you still need the bufused there? 16:20
16:20 ash_ joined
Coke or does str_new_noinit take care of that? 16:20
perl 5.12 is out. 16:28
darbelo Coke: It failed a test when I left it out. There's a lot of stuff under the cover.
Basically, that bufused is wrong. It shouldn't be set at all. But other code deeper indide io is (ab)using it as a substitute for the buffer length. 16:29
s/indide/inside/ 16:30
And that plays into the 'fake' string passed as a buffer to PIO_READ()
And *that* one... 16:31
16:32 tcurtis joined
darbelo ... blah... blah... encapsulation... amputate and set on fire... etc. 16:32
allison Coke: great! (on getting into the admin screens now) 16:33
Coke darbelo: ok. be mindful of the big release next week. major hacking on that should wait until post 2.3
Coke builds 5.12 and will attempt to build parrot shortly. 16:34
(though ISTR it did ok in the RCs)
16:34 whiteknight joined
bubaflub i need to get setup with perlbrew to play with 5.12 16:34
whiteknight seriously, my computer has bluescreened twice today now
darbelo Oh, I'm not doing any major hacking. This is just small-ish cleanup to allow other work on strings.
Oh, and I'd have to do it on three platforms, since io is split into unix/portable/windows. Not gonna happen this week ;) 16:35
But... I can kill the fake string now. 16:37
cotto_work good mornings
purl good mornings are when all my stocks are up in after hours trading
darbelo Can someone please test nopaste.snit.ch/20259 on windows and... whatever system we're supporting with src/io/portable.c 16:46
16:47 allison joined 16:55 ash_ joined
Coke the lexicals failure is the likely culprit for my rakudo build being borked. 16:55
dalek kudo: fb38dbf | quester++ | src/core/ (3 files):
Fix Parcel.sort and add Hash.sort.

  cloud.github.com/downloads/perl6/bo...010-04.pdf
See also: rt.perl.org/rt3/Ticket/Display.html?id=74334 Signed-off-by: Moritz Lenz <moritz@faui2k3.org>		 16:58
Coke ah, I blame bacek. testing... 16:59
darbelo Coke: It was bacek. r45628
But only #8 is a real failure. 17:02
Andy Seems to me that clone_interpreter doesn't belong in the parrotinterpreter PMC 17:03
darbelo asks without looking.
Why?
Andy maybe in src/interp/inter_create.c 17:04
it's a standalone function, not part of the PMC 17:05
darbelo Is it used outside of the PMC?
Andy yes
:q
ack -w clone_interpreter
Three hits. 17:06
Really, there should be no visisble funcitons in any PMC
darbelo Yep.
Andy If there's a visible function in the PMC, either the function should be hidden with a static, or it should be moved elsewhere. 17:07
darbelo +1 on the move.
Andy tanks 17:08
whiteknight and planes!
Andy I'm trying to get everything headerizerd properly
darbelo The only argument for keeping it in would have been that it was unused outside.
Andy And I would have made it static 17:09
because I'm smart like that.
darbelo Andy++
Coke dalek? 17:19
purl well, dalek is #parrot's spammy little rss bot or (see: dalek plugins)
17:21 chromatic joined
dalek TT #1555 closed by coke++: t/op/lexicals.t failures 17:23
TT #1555: trac.parrot.org/parrot/ticket/1555
rrot: r45636 | fperrad++ | trunk/runtime/parrot/library/TAP/Formatter.pir:
[TAP] fix verbose output
rrot: r45637 | coke++ | trunk/src/ops/var.ops:
Revert r45628, which fails t/op/lexicals.t and breaks rakudo's build.
17:54 dngor joined
dalek parrot: 83c9200 | (Joshua Tolley)++ | t/sql/test.sql:
Add a test for syntax error handling. This doesn't actually test anything, or output tap, because... I don't think anyone knows how exactly syntax error reporting should be handled. Right now it just crashes, which is not what we want.
 18:06
parrot: 7a785b7 | (Joshua Tolley)++ | t/sql/test.sql:
Revert "Add a test for syntax error handling. This doesn't actually test anything, or output tap, because... I don't think anyone knows how exactly syntax error reporting should be handled. Right now it just crashes, which is not what we want."

We want this in a branch so master keeps passing tests		 18:18
AndyLunch Hmm, clone_interpreter needs ot see Parrot_Class_attributes 18:19
Andy so much for that idear. 18:21
dalek rrot: r45638 | darbelo++ | trunk/src/io/buffer.c:
Replace a fake STRING with a real one created via the STRING api.
 18:30
rrot: r45639 | darbelo++ | trunk/src/io/buffer.c:
Fold two separate conditionals that did piecemeal string construction into a single API call.
 18:47
18:53 ingy joined 19:10 joeri joined
dalek TT #1556 created by tcurtis++: Wrong path and description of mk_language_shell.pl in PCT Book 19:21
TT #1556: trac.parrot.org/parrot/ticket/1556
19:24 jrtayloriv joined
cotto_work chromatic: ping 19:33
japhb #ps is an hour from now, yes?
darbelo ps?
purl somebody said ps was postscript or process status or see "parrotsketch" or non-vector?! or annoying.
cotto_work yes
darbelo parrotsketch?
purl parrotsketch is probably a status meeting for parrot core committers held every Tuesday at 20:30 UTC in #parrotsketch
darbelo clock?
purl darbelo: LAX: Tue 12:33pm PDT / CHI: Tue 2:33pm CDT / NYC: Tue 3:33pm EDT / LON: Tue 8:33pm BST / BER: Tue 9:33pm CEST / IND: Wed 1:03am IST / TOK: Wed 4:33am JST / SYD: Wed 5:33am EST /
japhb cotto_work, thx 19:34
chromatic pong 19:42
moritz dukeleto, particle: any news from the slot allocation front? 19:43
Coke nothing on the mentors list.
particle no news 19:44
moritz thanks 19:45
Coke no, no news is <reply>No news is good news with garry gnu.
purl okay, Coke.
nopaste "jrtayloriv" at 192.168.1.3 pasted "Why "Method 'ast' not found for invocant of class 'ResizablePMCArray'" ??" (187 lines) at nopaste.snit.ch/20261 19:49
moritz jrtayloriv: what's the grammar? 19:51
purl rumour has it the grammar is "The Deluxe Transitive Vampire" Mmm. Good book
moritz jrtayloriv: is <unsigned_real_10> quantified? with a question mark, maybe? 19:52
nopaste "jrtayloriv" at 192.168.1.3 pasted "grammar" (248 lines) at nopaste.snit.ch/20262
jrtayloriv moritz, no, but unsigned_int_10 is with a '+' 19:53
moritz jrtayloriv: that's why it's an ResizablePMCArray, not a Match object
jrtayloriv: and only Match objects have an .ast method, not RPAs
jrtayloriv: also 'token unsigned_real_10' looks fishy to me 19:54
NQP doesn't do longest token matching
so the first alternative always wins against the second
jrtayloriv hmm .. OK -- I'll fix that. Thanks. 19:55
moritz second, if you have two named captures of the same name, they'll end up as an array, too
you can rename them as <nominator=.unsigned_int_10> '/' <denominator=.unsigned_int_10>
19:56 cotto_work joined
moritz then they are all single match objects, not arrays 19:57
jrtayloriv I see. Thanks again.
moritz you're welcome 19:58
20:00 NotFound joined
NotFound hi 20:00
moritz hi 404 :-) 20:01
tcurtis Is there an emacs mode for PGE Grammars? 20:07
moritz tcurtis: in the pugs repo there's a Perl 6 file for emacs, don't know how good its grammar support is 20:08
tcurtis: svn.pugscode.org/pugs/util/cperl-mode.el 20:09
since PGE and NQP are meant to be a subset of Perl 6, it might work for you
20:10 dngor joined
dalek rrot: r45640 | petdance++ | trunk/src/pmc.c:
Undid a local variable shadow naming. Removed an extra test, and added some parens for clarity.
 20:10
20:15 cognominal joined
tcurtis Thanks, moritz, that mostly works. 20:16
NotFound WTF? make: *** No rule to make target `tapir.pbc', needed by `installable_tapir'. Stop. 20:25
nopaste "jrtayloriv" at 192.168.1.3 pasted "grammar/actions/parse etc." (78 lines) at nopaste.snit.ch/20263 20:26
jrtayloriv Moritz: Sorry to bother you again, but I can't figure out how I would make these into match objects, or what I should do. It's not just unsigned_int_10 that is a ResizablePMCArray, but also real_10 and digit_10 for example. (see incoming paste)
dalek rrot: r45641 | petdance++ | trunk (2 files):
Remove an extra math op. Make Parrot_pmc_is_null() use PMC_IS_NULL()
 20:27
cotto_work NotFound: tapir uses distutils. It looks like you just run setup.pir
jrtayloriv I guess I'm not understanding why real_10, unsigned_int_10, etc. are Arrays instead of match objects ...
NotFound cotto_work: I just run: make ; make install
Andy crap
20:29 dngor_ joined
NotFound If I "make tapir" first, it works. 20:29
jrtayloriv Oh, wait ... I think I understand ... nevermind
NotFound Looks like someone forgot a rule. 20:30
mikehh time for #ps
chromatic allison, did you check the tt_389 branch?
allison chromatic: didn't see it, was it mentioned in the ticket? 20:31
chromatic Should have been, but it sounds like you reached the same point Whiteknight and I did. 20:32
allison chromatic: it wasn't in the ticket (just checked)
chromatic: the fix is insanely simple
and, I just found the PGE problem too!
chromatic Not the PGE fix.
allison I mean, the fix for not storing methods and vtable functions in the namespace is simple 20:33
chromatic: we need a flag for NSENTRY marked subs like we have for methods
chromatic I thought we had that. 20:34
dalek rrot: r45642 | petdance++ | trunk/src/pmc.c:
We can't use PMC_IS_NULL() because that calls us here in some cases
 20:43
rrot: r45643 | NotFound++ | trunk/config/gen/makefiles/root.in:
Makefile rule for tapir.pbc
20:44 TiMBuS joined
moderator #parrot Parrot 2.2.0 "Like Clockwork" Released! | parrot.org/ | Channel log: irclog.perlgeek.de/parrot/today | Priority: Documentation sprint for 2.3, fix line number annotations | Review and vote GSoC applications 20:56
moderator #parrot Parrot 2.2.0 "Like Clockwork" Released! | parrot.org/ | Channel log: irclog.perlgeek.de/parrot/today | Priority: documentation sprint and pre-release testing for 2.3, fix line number annotations | Review and vote GSoC applications 20:58
dalek rrot: r45644 | NotFound++ | trunk/src/string/encoding/fixed_8.c:
cast required for C++ build
 21:00
21:07 dngor joined 21:11 allison joined
bacek morning 21:16
mikehh hi bacek 21:17
bacek hi mikehh
21:27 plobsing joined 21:28 estrabd_ joined
GeJ Goor morning everyone. 21:28
bacek G'Day GeJ 21:29
21:29 estrabd joined
GeJ Heya bacek. How's life treating you these days? 21:30
bacek GeJ, not so bad 21:32
dalek rrot: r45645 | mikehh++ | trunk/src/io/utf8.c:
fic codetest failure - trailing spaces
21:33 Whiteknight joined
mikehh ha that shouls have been s/fic/fix/ 21:34
21:34 AndyA joined
Whiteknight dukeleto: ping 21:36
dalek tracwiki: v4 | chromatic++ | PerformanceImprovements
tracwiki: trac.parrot.org/parrot/wiki/Perform...ction=diff
tracwiki: v5 | chromatic++ | PerformanceImprovements
tracwiki: ordered by rough progress
tracwiki: trac.parrot.org/parrot/wiki/Perform...ction=diff
21:41 snarkyboojum joined 21:43 cotto_work joined
bacek chromatic, any progress with rakudo on immutable strings? 21:54
cheap karma for someone who will remove pcc_mergecells branch 21:57
mikehh bacek; that's to go? 22:03
chromatic bacek, no progress. I don't know if Jonathan has been available at all today.
bacek mikehh, yes 22:05
plobsing 38869/57921 (~67%) bytes in parrot_nci_thunk_gen.pbc are null for me. is that representative of PBCs in general? 22:06
mikehh All tests PASS (pre/post-config, make corevm/make coretest, smoke (#33169), fulltest) at r45645 - Ubuntu 10.04 beta amd64 (g++ with --optimize)
that's with perl 5.12 ;-} 22:07
had to add a bunch of modules :-}
chromatic Four-byte INTVALs, most of them numbers < 255 -- sounds reasonable, sadly. 22:08
cotto_work plobsing: I think so. They compress well.
chromatic If Parrot had gzip support, fakecutables could... I'll get my hat. 22:09
plobsing I'm just looking at the xxd of it. row after row after row of what looks to be mostly garbage 22:12
cotto_work Wow. parrot-nqp.pbc compresses from 1.4M to 104K.
chromatic Even a quick RLE in the fakecutable would help. 22:13
cotto_work orle? 22:14
plobsing what's the cost of decompression vs disk read? 22:16
chromatic I'm thinking of compilation time of the executable, which is a little silly.
Coke RLE? 22:22
purl RLE is Run Length Encoding or Read Learn Evolve
chromatic Run-length encoding
Coke all orgs have initial GSOC slots assigned. 22:23
cotto_work only 10 for tpf 22:24
chromatic How many proposals did we have? 22:25
cotto_work 24
Coke most of TPF's proposals are currently slated for rejection. 22:26
chromatic The good news is that TPF does have quarterly grants, so some of them can go that direction.
cotto_work is it kosher to say which ones looks like they'll make it?
Coke also some perl6 stuff in there that could hit the Hague, mebbe. 22:27
cotto_work: Not here, I think.
cotto_work ok
Coke (all the rejection ones are probably just lacking assigned mentors)
cotto_work They're all assigned mentors.
Coke ohohoh. sorry, those are just below the 10th slot. 22:28
22:30 wagle_ joined
plobsing If I wanted to add support for compression in parrot images, what would be required other than showing it improved performance? Is a third party library sufficient? Do I have to rewrite in something that runs on top of parrot? 22:35
22:36 allison joined
dalek rrot: r45646 | mikehh++ | branches/pcc_megrecells:
branch completed
 22:38
Coke if it's a 3rd party library, there are 3 issues: cross-platform support (if we ever get back to cross platform support in PBC), licensing, and support on the main 3 platforms. (this last is more important than the first, though they're related.) 22:50
(would presumably also need a deprecation notice about a planned change to the bytecode.) 22:51
plobsing bytecode modification deprecation is on the books since 1.0 IIRC 22:52
plobsing checks 22:53
guess not 22:54
Whiteknight 10 slots really doesn't seem like a lot 23:03
but looking at the scores, many of the Parrot ones are up in the top 10
so that makes me feel better 23:04
23:05 snarkyboojum joined
cotto_work I'd be happier with more but I don't think we'll lose any of the best proposals. 23:05
Whiteknight we only had 5 possible students, and it looks like 4 of ours could be in the top 10 pretty easily 23:06
cotto_work btw Whiteknight, did you say that allison had dibs on mentoring darbelo's nfg proposal?
23:06 tcurtis joined
Whiteknight cotto_work: I don't know about "dibs", but she had expressed interest in it. 23:06
that's the one that looked to me like it needed the most architectural design work 23:07
hello tcurtis
tcurtis Hello, Whiteknight.
Whiteknight so when do we need to make final decisions on all these, the 21st? 23:08
cotto_work istr so 23:09
Coke obra++ 23:18
cotto_work I somehow got allison's post to parrot-dev almost an hour after Whiteknight's reply. 23:26
23:30 AndyA joined
allison cotto: I don't need to mentor the NFG proposal, but will be involved in it whether I mentor or not 23:31
cotto_work decides to let dukeleto handle any gsoc issues that don't affect him directly. 23:32
23:33 Mokurai joined 23:34 mariano joined
plobsing wonders what random text in the form of "(0x123,0x345,...)" in the middle of PBC is 23:55
cotto_work It's a sekrit message: "delete me" 23:56
chromatic "Don't forget to drink your Parrotine" 23:57
cotto_work plobsing: what does pbc_disassemble say it is? 23:58
Whiteknight moc tod enofeivom

Powered by Cro and the Raku® Programming Language.

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

Thank you!