pugscode.org/ | nopaste: sial.org/pbot/perl6 | ?eval [~] <m oo se> | We do Haskell, too | > reverse . show $ foldl1 (*) [1..4] | irclog: irc.pugscode.org/
Set by diakopter on 11 July 2007.
00:05 theorbtwo joined 00:09 lichtkind left 00:27 thoughtpolice joined 00:52 fglock left 01:01 cognominal_ joined 01:05 fglock joined 01:26 weinig is now known as weinig|away 01:29 fglock left 01:44 jonathanturner joined 02:02 thoughtpolice left 02:04 Limbic_Region left 03:01 kanru left 03:02 kanru joined 03:07 elmex_ joined 03:21 elmex left 03:24 zakame_ joined 03:39 AzureStone joined 03:42 r0bby_ joined 03:46 r0bby_ left 04:25 amnesiac joined 04:46 zakame_ left 04:57 solomon joined 05:19 zakame_ joined 05:31 ofer left 05:41 penk joined 05:43 zakame_ is now known as zakame 06:05 kanru left 06:29 kanru joined 06:40 dj_goku joined 06:42 zakame left
araujo reading PP 07:03
TimToady, bunch of good quotes there
:-)
07:19 drrho joined 07:36 ruz left 07:46 weinig|away left 07:57 franck___ joined 08:00 Aankhen`` joined 08:13 masak joined 08:34 zakame joined 08:43 jisom joined 08:51 kane__ joined 08:54 iblechbot joined 08:56 masak left 09:00 zakame left 09:05 gongyiliao joined
rafl obra: pong 09:16
09:24 cognominal_ left 09:52 drupek12 left 10:10 zakame joined 10:13 penk left 10:15 chris2 joined 10:20 Khisanth left 10:27 iblechbot left 10:41 iblechbot joined 10:42 drupek12 joined 10:43 zakame left 10:48 zakame joined 10:49 drbean joined 10:56 gongyiliao left 11:03 drbean left, drbean joined 11:16 jisom left 11:17 polettix joined 11:40 achilles14 joined
achilles14 hi all, I am new to parrot. how to complie perl6 script to parrot bytecode? 11:40
11:43 achilles14 left 11:44 achilles14 joined
achilles14 which command to use to compile script into parrot bytecode? 11:44
11:48 achilles14 left 11:56 AzureStone left 12:03 drbean_ joined 12:06 drbean left 12:10 ebassi joined 12:14 chris2 left 12:38 jonathanturner left 12:43 weinig joined 12:48 weinig left 12:49 weinig joined 12:50 drrho left 13:03 lichtkind joined 13:04 ebassi is now known as ebassi|lunch 13:07 ludan joined
ludan hi 13:08
13:13 cognominal_ joined 13:25 Khisanth joined 13:34 cmarcelo joined 13:39 laye joined 13:40 ebassi|lunch is now known as ebassi, ebassi left 13:41 ebassi joined 14:07 drbean_ left, drbean joined 14:08 rindolf joined 14:15 drbean left, drbean joined 14:18 c9s left 14:19 masak joined, hcchien left 14:21 solomon left 14:30 kanru left 14:32 drupek12 left 14:35 pmurias joined 14:38 cognominal_ left 14:41 cognominal_ joined 14:42 pbuetow left, drbean left 14:43 drbean joined 14:44 drupek12 joined 14:46 TJCRI joined 14:48 c9s_ joined, c9s_ left 14:49 c9s joined 14:51 drrho joined 14:56 drbean left, drbean joined, gongyiliao joined, gongyiliao left 14:57 gongyiliao joined 14:59 pbuetow joined 15:00 pbuetow left, pbuetow joined 15:05 penk joined 15:09 gongyiliao left 15:13 Psyche^ joined 15:23 polettix left 15:24 rindolf left 15:30 Patterner left, Psyche^ is now known as Patterner 15:31 zakame_ joined 15:33 stevan_ joined 15:35 TJCRI left 15:40 TJCRI joined 15:44 ilogger2 joined 15:47 Casan_ joined 15:48 ofer joined 15:50 baest joined 16:08 Lorn joined
obra hey rafl. one of my minions has been hacking Net::SSLeay for segfaults we're hitting in prod. I was curious if his patch (on rt.cpan) looked sane 16:20
rafl obra: I didn't have a chance to give ssleay some love recently. you could ask [email@hidden.address] the co-maintainer. 16:21
16:22 Southen joined
obra ok. will do. thanks 16:22
16:24 cls_bsd left 16:36 Psyche^ joined 16:40 kanru joined 16:41 dlocaus joined
dlocaus Hi everyone! :) 16:43
16:53 ilogger2 joined 17:18 justatheory joined 17:29 justatheory left 17:30 justatheory joined 17:42 TJCRI joined 17:45 rindolf joined 17:53 perlbot joined 18:00 justatheory left 18:09 fglock joined
fglock dlocaus: hi :) 18:10
18:11 jjore-w joined
fglock I went to Tomar (Portugal) to give a talk on Perl, not much hacking today 18:12
I rewrote the .BIND code yesterday, but it still has some errors
dlocaus cool
fglock I guess I'll move the failing tests to t/todo/ and commit 18:13
dlocaus make forcerecompile; make test; breaks the current code
fglock hmm
dlocaus I'm trying to find the revision that broke it
fglock looking
dlocaus its either dlo or pmurius that broke it
fglock not necessarily :P 18:14
pasteling "dlocaus" at 63.192.210.11 pasted "hunting down bad commit" (74 lines, 2.3K) at sial.org/pbot/28757
dlocaus fglock: I was wondering, since we have a good working test case (0 errors), could we convince lambabot to do a "make forcerecompile; make test" after every commit? 18:15
after the shell script btw, is some "workspace"
18:16 rindolf left
fglock or, we could have a dedicated bot to do that and then make a smoke report 18:17
we can even reuse the pugs infrastructure
but I don't know much about how the pugs smoke works
still compiling...
dlocaus agh! I'm the guilty party
fglock "guilty partly" - collaborative work++ 18:19
dlocaus 18886
that's the bad revision
oh wait a minute, that's not me.
dev.pugscode.org/changeset/18886 18:20
fglock about my talk, people are much interested in the Perl 6 multi processor capabilities 18:21
pasteling "dlocaus" at 63.192.210.11 pasted "Revision 18886 broke the test cases when doing a forcerecompile" (53 lines, 2.6K) at sial.org/pbot/28758
fglock and the thing they want most is an IDE with some kind of "best practices" plugin 18:22
dlocaus Who is "they"? One of the things I REALLY hate is when my IDE (ultraedit) attempts to correct my perl. It drives me nuts. 18:23
(auto-correct keywords).
Don Canonway's PBP? 18:24
fglock it's a university audience, this what they expect from a "first language"
dlocaus or by best practices do you mean Perl::Critic?
fglock they want a limited syntax 18:25
which is possible with Perl 6
dlocaus a limited syntax?
fglock like, just one way to do each thing
but this is just for beginners
dlocaus oh I see. 18:26
fglock they do want all the other things, later
and a nice beautiful colorful IDE
:)
dlocaus It is an interesting question... do it 1 way (java) or DWIW (perl) 18:27
fglock: how do I got about reverting out pmurias's changes? 18:28
Should I inform him and wait for a reply? and if he doesn't respond "reasonbly" soon, just do it and @tell him? 18:29
pmurias: ping?
pmurias: ping? 18:30
fglock dlocaus: what kind of changes? if you are fixing a mistake, just do it
dlocaus fglock: dev.pugscode.org/changeset/18886
lambdabot Title: Changeset 18886 - Pugs - Trac
dlocaus If I knew it was a mistake I would do it.
fglock my local copy is doing an infinite loop after make forcerecompile - I guess I'll not commit
looking
dlocaus But I'm really just acting the part of a machine. Do a binary search, find out which commit broke the code, revert that code. 18:31
and I didn't even have the good graces to do a binary search, I only did part of one. 18:32
fglock dlocaus: if you find out that the deleted code was necessary, I think it's better to add a test to demonstrate the need
than just put the code back
dlocaus fglock: I would, if I knew what was going on.
I'm only good enough (right now) to know that the code broke the test cases. 18:33
fglock this code is working around a grammar bug:
the grammar thinks 'self' is actually 'self()'
so it parses to self.APPLY 18:34
that code reverts that at the emitter
dlocaus fglock: this part? src/KindaPerl6/Ast.pm
oh, wait, nm 18:35
fglock dlocaus: it's fixable at Grammar::Term
dlocaus fglock: could you please fix it? I do not think I have the skills (yet) to fix it. 18:36
fglock but the workaround is ok for now, since all the grammar will be rewritten anyway
dlocaus workaround? but the test cases break
fglock ok 18:37
you mean, the tests break because the workaround was removed?
dlocaus well, when pmurias removed the code, the test cases broke. 18:38
removed the code in 18886
yeah.. that's what I mean.
fglock you can just paste the code back 18:39
dlocaus ok
fglock and add a comment explaining the workaround
dlocaus humm ok. 18:40
fglock recompiling...
18:40 pjcj joined 18:41 drbean joined
fglock we do need a fast compiler 18:41
dlocaus :)
You know, I was thinking about the _isa, and I was wondering if it would be possible to write some sort of "compression" hack, that would check out an ::DISPATCH mess, and then compress it. 18:42
something to "flatten" it.
fglock the thing the audience liked most today was: time perl -e ' print "hello, World\n" '
dlocaus oh, you know I came across something interesting about perl
fglock dlocaus: yes, it's possible 18:43
dlocaus perl '' test.pl
fglock it's a method cache, I think perl5 implements it
dlocaus forces perl to read from the source code from STDIN and not from test.pl
I've been wondering if I should make of note of that to the perl5.10 people to see if it should be fixed 5.10?
fglock heh
dlocaus: do you believe it's an error? 18:44
dlocaus well, it is a bit odd. On one side of the coin, its not expected behavior. 18:45
on the other hand, it was caused/found by ahemm.. you.
fglock the problem with the flatten hack, is that we have "open" classes, so you need to track when a class is changed and recalculate the objects
dlocaus using TAP::Harness (Test::Harness 3.0) it called an open3 call, which when given an empty $extra_lib string in script/run_tests.pl caused the end arguments passed to perl to be "perl '' test.pl" 18:46
fglock :)
dlocaus when the invocation could have been "perl -I(libs) test.pl"
so, I'm not sure if it is a bug or not, but definitely unexpected.
I've been thinking about checking out 5.10 and seeing what changes need to be done to get kp6 working on it. 18:48
It would be good for the community... :)
fglock interesting bug, but I think it it's more like an "undocumented behaviour"
sure it would
dlocaus :)
I'll see if this bug works on perl5.10 and if it does, I'll send in a note to the development team about it. 18:49
fglock I don't think 5.10 will help much with kp6, because the kp6 runtime is so "low level" 18:50
it mostly just uses array, hash, and closure - none of the 5.10 cool hacks
dlocaus well, it will be a good test to do, just to "make sure". 18:52
fglock dlocaus: make forcerecompile / make test - passes all tests here
dlocaus on 18890?
or after you add the hack back in? 18:53
18:54 weinig_ joined
fglock without modifications - but I'm compiling again 18:54
ah - a second compilation fails
I'll try with the 'self' thing now 18:55
re second compilation, maybe the Makefile is compiling the Runtime::Perl6 stuff before recompiling kp6 proper 18:58
dlocaus I've always had grief with the Makefile...
esp. that make -j option. 18:59
:)
fglock I've only 10min left - compilation eats a lot of battery :P 19:00
dlocaus :(
feel free to commit, I'll update and run cmds for you. 19:01
fglock k
19:02 hanekomu joined
dlocaus I undid pmurias 18886 and put it in the 18890 revision, but there was massive breakage 19:03
@seen pmurias 19:04
lambdabot pmurias is in #perl6. I don't know when pmurias last spoke.
fglock I committed 18891, but I'm still testing it 19:05
now recompiling the Perl6 modules...
19:07 TJCRI left
fglock dlocaus: looks good here - I'm committing the recompiled modules 19:07
dlocaus ok
fglock done - 18892
dlocaus if it passes on my side, I'll annotate the re-hack 19:08
if you don't mind my asking, how are you running out of battery, but still web-enabled? 19:09
fglock 7min battery - bye bye :)
1st class trains have 220V power plugs
19:09 rindolf joined
fglock this is the cheap one :) 19:09
dlocaus if you are on a train, how can you be internet connected?
does the train have WIFI?
fglock I've got mobile internet (hsdpa modem) 19:10
dlocaus oh... neat :)
fglock very neat :)
did it work? 19:11
dlocaus 3.2GHZ amd is only so fast.... :(
I've gotten to the mp6.pl part now 19:12
fglock I've reverted my local changes - I need to split BIND into small refactorings
.BIND is everywhere in the code
dlocaus while forloops and such may be great for the mutli-cpu stuff, I wonder how the parrot people will handle it 19:13
it apperas to work, the snaity tests came back ok
!yea!
avar wonders how kp6 is doing these days
fglock avar: I'm working on the bootstrap,
dlocaus is doing janitorial duties, so it looks much better :)
avar fglock: :) 19:14
fglock which means to fix several workarounds
dlocaus is about to try and re-straighten out the perltidy -st option.
fglock which work in Perl5, but break in Perl6
dlocaus All tests successful.
Files=85, Tests=200, 81 wallclock secs ( 0.25 usr 0.08 sys + 74.72 cusr 3.78 csys = 78.83 CPU)
Result: PASS
fglock dlocaus++
avar: I rewrote much of the container code, including hash and array 19:15
1min!
dlocaus says a prayer for fglock's battery... 19:16
19:16 IllvilJa joined 19:18 drbean_ joined
dlocaus thinks his prayer worked... 19:20
avar billions of people worldwide share your little fallacy:)
avar waits for bootstrap to try lisp emitting again 19:21
dlocaus avar: :)
avar: well, it makes for an interesting test case. perl -e ' print "1..1\nok\n" prayer.pl 19:22
avar You can test whether the system runs but not whether all the lights are on?:) 19:23
dlocaus with todays webcams that shouldn't be a problem :)
I wonder if I could patent that?
Using a webcam to monitor blinking lights?
if image.x.y = red then warn "oppps!" 19:24
I hope that line of code did not just eliminate someone's job. 19:25
19:26 Limbic_Region joined 19:27 drbean left 19:29 drbean joined
Limbic_Region salutations all 19:32
dlocaus waves 19:33
19:36 drbean_ left 19:38 fglock left 19:39 hanekomu left, drbean left
pmurias dlocaus: ping 19:39
well pong
19:39 drbean joined
dlocaus pmurias: hiya 19:42
pmurias hi
dlocaus pmurias: I just wanted to let you know that fglock and I put back in some code you removed 19:43
commits 18891-18893 : dev.pugscode.org/changeset/18893
lambdabot Title: Changeset 18893 - Pugs - Trac
pmurias seen it
dlocaus ok. that's it.
pmurias the hack i removed $.code has a .name which is not always the case 19:45
dlocaus pmurias: I'm sorry, I'm not that far along in knowing how kp6 works, to understand what you said. 19:46
pmurias sorry 19:47
i meant to type the hack i removed assumed
that
$.code being the code attribiute of Apply 19:48
which is assumed by the hack to be a Var
is it now clearer ? ;) 19:49
dlocaus yeah, self vs self()
yes.
pmurias more like foo() vs $hash<key>.() 19:50
dlocaus I don't even understand perl6 all that well... ;)
pmurias foo() vs ($hash->{key})->() in perl5 terms ;) 19:52
dlocaus yep! I got that.
pmurias althought the later gets turned into a Call in kp6
19:53 jisom joined
pmurias Apply could be replaced by Call, as it's just the case with method attribiute being 'postcircumfix:<( )>' 19:54
dlocaus: if you want a refactoring task you could do it 19:55
dlocaus Right now, I want to streighten out the util/perltidyrc which forces an "-st" # output to stdout
pmurias ok 19:56
dlocaus I'm also looking at the cost of doing perltidy.
pmurias it's significant ;)
dlocaus Yes, I'm thinking it might be worhtwhile to put an in ENV{ kp6perltidy } to turn it on or off. 19:57
I'm running a test now..
not to mention, I think putting -st in perltidyrc is bad practice.. since it forces output to stdout, and prevents usage of -o and -b. 19:58
pmurias you can change it
dlocaus yes, but then the code in script/kp6, script/mp6 and Makefile all use it.
pmurias the is a CLI option to kp6 for not using perltidy
dlocaus yes.
I thought I'd run a test.. ah
its done, one second
19:59 rindolf left
dlocaus by wall clock seconds... 20:00
Running: make forcerecompile ; make test W/O perltidy
costs: 4m and 14seconds
Running: make forcerecompile ; make test W/ perltidy
costs: 5m and 28seconds
a difference of 102 seconds or about 1minutes and 42 seconds.
Assuming a developer forcerecompiles as a test 10 times a day, that's 20 minutes 20:01
or assuming there are 5 developers who forcerecompiles 10 times a day that is a loss of 85 minutes of developer time 20:02
pmurias: do you think it is worth putting in a on/off global perltidy switch? 20:03
pmurias i don't know
dlocaus ok, well, I'll let it be then for now.
20:03 Limbic_Region left
dlocaus if someone else has a thought on this, then they can @tell dlocaus "message" me. 20:04
20:04 weinig_ left 20:05 weinig joined 20:08 ruz joined 20:12 drbean left 20:16 Lightop joined 20:20 Lightop left 20:23 kanru joined 20:24 drbean joined 20:32 jisom left 20:38 bsb joined 20:44 mj41 joined 20:56 TJCRI joined 21:09 Psyche^ joined 21:10 Psyche^ is now known as Patterner 21:24 thoughtpolice joined 21:43 TJCRI left 21:45 Limbic_Region joined 22:11 weinig is now known as weinig|away 22:15 thoughtpolice left 22:52 drupek12 joined 23:18 polettix joined 23:59 ggoebel 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!