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.
09:14 ilogger2 joined 09:20 gongyiliao joined, amnesiac joined 09:22 iblechbot joined, amnesiac left 09:26 the_dormant joined 09:27 the_dormant left 09:31 polettix joined 09:35 the_dormant joined 09:48 the_dormant left 09:54 Azure-BOT joined 09:57 Azure-BOT left 10:11 ebassi joined 10:14 Aankh|Clone joined 10:15 jedai joined 10:30 gongyiliao left
moritz_ re 10:31
10:32 ruoso joined 10:41 elmex joined 10:59 pmurias joined
pmurias moritz_: re 10:59
lambdabot pmurias: You have 1 new message. '/msg lambdabot @messages' to read it.
11:01 gongyiliao joined 11:02 gongyiliao left 11:07 polettix left 11:11 devogon_ joined 11:19 the_dormant joined 11:22 the_dormant left 11:38 lichtkind joined 11:43 nipotan is now known as nipotaway, cai joined 12:00 rindolf joined 12:15 devogon_ left, devogon joined 12:16 devogon left 12:21 devogon joined 12:34 dalek joined 12:35 ebassi is now known as ebassi|lunch 12:47 |Jedai| joined 13:01 kanru joined 13:05 jedai left, ebassi|lunch is now known as ebassi, ebassi left 13:06 ebassi joined 13:29 baest joined 13:45 franck____ joined 13:48 franck___ joined, franck____ left 13:49 franck____ joined, franck___ left 13:50 franck____ left 13:51 Aankh|Clone left 14:01 Aankhen`` joined, TJCRI joined 14:04 drrho joined 14:06 franck___ joined, drbean joined, drbean left 14:07 drbean joined 14:21 gugod joined 14:25 justatheory joined 14:28 kanru left 14:31 Lunchy joined 14:42 kanru joined 14:46 justatheory left
[particle] yak shaving & 14:51
14:53 franck___ left 14:56 cai left
pugs_svn r18772 | agentz++ | [PCR] fixed the final Win32 test failure regarding hashing randomness. 15:00
15:04 dlocaus joined
dlocaus Hello! 15:04
lambdabot dlocaus: You have 1 new message. '/msg lambdabot @messages' to read it.
pugs_svn r18773 | agentz++ | [PCR] bumped the version number to 0.28 15:06
dlocaus @help
lambdabot help <command>. Ask for help for <command>. Try 'list' for all commands
dlocaus @help list
lambdabot list [module|command]
show all commands or command for [module]. www.cse.unsw.edu.au/~dons/lambdabot/COMMANDS
agentzh @list 15:07
lambdabot www.cse.unsw.edu.au/~dons/lambdabot/COMMANDS
dlocaus @clear-messages
lambdabot Messages cleared.
pugs_svn r18774 | dlo++ | [kp6] added link for lambdabot 15:10
r18775 | agentz++ | [PCR] some releng work for 0.28 15:11
pmurias dlocaus: hi 15:15
dlocaus: got my message? 15:16
15:19 justatheory joined 15:25 drrho left 15:29 justatheory left
obra $TIMEOFDAY 15:31
dlocaus pmurias: yes, 15:39
15:39 rindolf left
dlocaus pmurias: I'm currently trying to debug t/kp6/15-capture-from-rule-block.t 15:39
pmurias: the problem is I'm caught up in all of Pair.pm's INIT blocks, and it has a LOT of them. 15:40
15:40 amnesiac joined
dlocaus does anyone know how to force the debugger to skip the INIT { } blocks? 15:40
putting $DB::single and then using "c" doesn't work, the perl debugger still stops on the first line of INIT{ .. }
pmurias dlocaus: did fixing what i mentioned help? 15:43
dlocaus well, yes, it appeared to fix the majority of bugs, but 15-capture...t seams to go into an infinite loop
I think, I'm trying to figure out how to get the debugger to debug it.
pmurias you could temporarly disable INIT blocks in the emitter
dlocaus the INIT blocks are being invoked in Pair.pm 15:44
compiled/perl5-kp6-mp6/lib/KindaPerl6/Runtime/Perl6/Pair.pm
ha ah! 15:45
15:45 Psyche^ joined
pmurias dlocaus: fixed it? 15:45
dlocaus the solution is to create BEGIN { sub daniel_test { print 1; } }; then put main::daniel_test(); where you want the break point, and then use b main::daniel_test to get there. 15:46
pmurias: no, just finally got a break point :)
now, I'm going to try and find out where it infinitely loops (I'm assuming that 4 minutes of no response is an infinite loop.. rather than actuall slowness) 15:47
omg... 15:56
$::MyGrammar dumps to 3k lines of variable content.
15:58 justatheory joined 16:00 justatheory left 16:02 Psyche^ is now known as Patterner 16:07 barney joined 16:16 chris2 joined 16:23 TJCRI left
dlocaus pmurias: ping? 16:24
With the fix you suggested, gammar tests 15 & 17 do not respond, but other tests do. Should I commit the fix, and then skip those tests? 16:25
by "do not respond" I mean they just continue running and there does not seam to be any resolution on the test.
pmurias dlocaus: pong 16:27
dlocaus do you have a suggestion?
pmurias dlocaus: commit the fix
dlocaus what about the test cases? 16:28
pmurias leave them as failures
16:28 justatheory joined
dlocaus they don't fail 16:28
they just keep on going...
as if they were in an infiniate loop
If I commit this, then some smoke server may be waiting a long time... 16:29
for results.
pmurias there isn't a smoke server for kp6 as far as i know
dlocaus should I deliberately break the 15 & 17 tests?, so that they "do" finish, with a failure at least?
pmurias hard question 16:30
dlocaus I know...
which, is why I'm asking :)
pmurias leave them as they are
the person debugging will not have to unbreak them 16:31
dlocaus how will people do the test cases to test their own changes, do the test cases?
everyone working on something else will be stuck.
pmurias break them then 16:32
dlocaus ok
pmurias i remembered that C-c stops the whole test process
have you used KP6_TOKEN_DEBUGGER=1? 16:34
dlocaus humm no 16:35
16:35 ting joined
dlocaus I'll give that a shot, right after I commit 16:35
16:35 justatheory left
pugs_svn r18776 | dlo++ | [kp6] 16:42
r18776 | dlo++ | I have commited a fix on pmurias's suggestion, which appears to fix a has_default.bit() issue, where
r18776 | dlo++ | has_default was not being assigned a "bit" class. It was being left "undef", which caused a but in
r18776 | dlo++ | Emit/Perl5.pm which is depending on the has_default.bit class.
r18776 | dlo++ |
r18776 | dlo++ | I have deliberately broken grammar tests 17 and 15. With this bug fix, the test cases do not appear to complete.
r18776 | dlo++ | They either "hang" or go into an infinite loop. Test case grammar/15 does not complete after 4 minutes.
r18776 | dlo++ |
r18776 | dlo++ | Both test cases, have compiliation errors that will cause them to failure. Comments are included as to why they are
r18776 | dlo++ | forced to fail.
obra dlo++ # long commit messages 16:45
Tene obra++ # rewarding long commit messages 16:47
dlocaus :) 16:50
wait a minute 16:51
@karma dlocaus
lambdabot You have a karma of 5
dlocaus I'm dlocaus
Tene @karma dlo
lambdabot dlo has a karma of 74
pmurias dlocaus++ # working on kp6
dlocaus dlo is my CPAN id, and my svn.pugscode.com user id.
dlocaus is my irc name, "dlo" is already taken on freenode. 16:52
16:52 dlocaus is now known as dlo, dlo is now known as dlocaus, dlocaus left 16:53 dlocaus joined
dlocaus my karma is being given away! wahhhhh! 16:53
Tene @dlocaus-- # no karma for you 16:54
lambdabot Unknown command, try @list
Tene ;)
dlocaus cries
16:56 weinig joined 17:03 drrho joined, ruoso left 17:06 araujo joined 17:08 justatheory joined 17:10 jedai joined 17:12 rindolf joined 17:17 Aankhen`` left 17:28 |Jedai| left 17:32 drrho left 17:35 ebassi left
moritz_ i've got a new hobby: taking a random shell and write `!! | wc -l' ;-) 17:40
rgs you like danger 17:41
Tene Interesting.
17:44 amnesiac left 17:55 lambdabot joined
dlocaus lambabot left for fishies? 17:55
moritz_ @tell lambdabot boo 17:56
lambdabot Nice try ;)
17:57 justatheory left 17:58 rindolf left
dlocaus @tell lambabot lambdabot did a boo boo 18:00
lambdabot Consider it noted.
18:00 TJCRI joined, cmarcelo joined
pmurias dlocaus: die might be a better way of making a test file fail than a undeclared subroutine 18:01
dlocaus pmurias: we have "die"?
moritz_ or `say "not ok $number"', and comment the failing lines 18:02
dlocaus humm, didn't think of that. 18:03
I'll make thos modifications in a bit.
I'm right in the middle of futzing with MOP.pm
18:08 Psyche^ joined 18:11 rindolf joined
pmurias walk& 18:15
moritz_ table tennis & 18:16
18:21 jovs joined 18:26 Patterner left, Psyche^ is now known as Patterner
pugs_svn r18777 | dlo++ | Put in mortiz_'s suggestion on how to fail better 18:36
18:38 Khisanth joined 18:49 TreyHarris joined, TreyHarr1s left
dlocaus hummus & 18:50
19:06 rindolf left 19:08 fridim_ joined 19:13 jisom joined 19:15 devogon left 19:16 devogon joined
pmurias ls 19:19
wrong screen window
dlocaus: have you seen agentzh's tracer?
19:22 fridim_ left 19:24 barney left 19:26 polettix joined 19:36 chris2 left 19:41 jisom left
dlocaus pmurias: sorry, I was eating lunch 19:56
pmurias: no, what is it/
19:56 Southen joined
pasteling "dlocaus" at 63.192.210.11 pasted "Code to catch infinite ::DISPATCH loops." (34 lines, 1.2K) at sial.org/pbot/28405 19:59
dlocaus pmurias: I'm thinking about committing that code that is in the last message, do you have a comment?
it goes in MOP.pm sub ::DISPATCH 20:00
pmurias looks good 20:03
DB::single invokes the debuger?
dlocaus well, if you start it in debugging mode. 20:04
perl -d scripts/kp6 ...
pmurias pugs.blogs.com/pugs/2007/10/a-graphical-tra.html 20:07
lambdabot Title: Pugs: A graphical tracer for Perl 6 regexes based on PCR
dlocaus pmurias: ok, I'll check that out 20:08
pugs_svn r18778 | dlo++ | [kp6] 20:13
r18778 | dlo++ | I have been really hard time chasing down a regex bug. Mostly due to the problem of tracing the stack. I have
r18778 | dlo++ | written code into Runtime/Perl5/MOP.pm ::DISPATCH to detect infinite loops. The premise is simple. If you see the
r18778 | dlo++ | same ::DISPATCH(ARGS) call more than once, you've entered a circular loop.
r18778 | dlo++ |
r18778 | dlo++ | This code is enabled by setting enviromental variable DEBUG = 1 (or 2 for more info)
r18778 | dlo++ | setenv DEBUG 1
r18778 | dlo++ |
r18778 | dlo++ | [email@hidden.address] #perl6
dlocaus pmurias: I think MOP.pm needs to document the hash structures it builds. 20:18
pmurias: do you have any thoughts right off hand?
pmurias looking
dlocaus: it could document the internal object format 20:19
what fields besides _dispatch the object can have 20:20
dlocaus well, I was thinking that DISPATCH($::PROTO,'add_method','func', args ) could include $self->{_document} = [ 20:21
"func belongs to $self->{ name }", .. next func ];
putting _document in the list, would give the human debugger an idea of exactly where he is, when he's busy touring the list. 20:22
pmurias i don't understand
the objects need to keep extra info to make the dumps more readable? 20:24
dlocaus well, I was attempting to understand $::MyGrammar, and x \$::MyGrammar drops 3k lines of "variable settings". When I use the DEBUG code in MOP.pm (I just added it), it produces a trail 4k lines long.
20:24 idiotmax joined
dlocaus When we are going though the ::DISPATCH / $::MyGrammar, we need some way of "knowing" or having an "idea" of where we are. 20:24
pmurias grabbing a better chair& 20:25
dlocaus pmurias: ::DISPATCH( $::MyGrammar, 'tok2', );
that code generated 4k of enter/exit documentation
pmurias you shouldn't dump objects with Data::Dumper 20:26
dlocaus that wasn't Data::Dumper, that was perl's debugger x command 20:27
pmurias i'm not familiar with the perl debugger but i guess it's equivalent 20:28
20:28 cookys joined
dlocaus well, my point is I'm wandering in the forest of $::MyGrammar without a road map. 20:29
pmurias dlocaus: i'll check what you get
dlocaus I'm thinking that MOP.pm should be documenting that forest for me (or doing something so that I know where I am when I'm trying to debug it)
pugs_svn r18779 | dlo++ | cleaned up the output a bit 20:30
pmurias dlocaus: my grammar should have a .perl metho 20:31
d
20:31 cookys left
pmurias which would dump only the intresting stuff 20:31
or are you intrested in the "low level" stuff as the _dispatcher etc.? 20:32
20:32 cookys joined 20:33 cookys left
dlocaus well, I'm trying to figure out how this code got into an infinite loop 20:33
20:33 idiotmax left 20:34 cookys joined
pmurias dlocaus: there shouldn't be ';' and the end of classes/modules 20:37
pugs_svn r18780 | dlo++ | [kp6] Moved edited code from compiled into src MOP.pm
pasteling "dlocaus" at 63.192.210.11 pasted "module Main { $_ = 'a'; MyGram" (14 lines, 167B) at sial.org/pbot/28408 20:38
dlocaus pmurias: that?
pmurias yes
kp6 recognies {} as hash subscripts 20:39
with the ; colons
dlocaus humm. ok 20:40
well, the code still doesn't work :(
20:41 idiotmax joined
pmurias still infites loops 20:41
?
pugs_svn r18781 | dlo++ | [kp6] removed an excessive print (newline) 20:42
dlocaus yes
pmurias: how do I invoke the MyGrammar.perl() ? 20:43
pmurias i don't think there is one 20:44
::DISPATCH($::MyGrammar,'perl') or MyGrammar.perl() if there were
dlocaus is this something I should leave to fglock or someone else to debug? I'm getting a sense I'm starting to thrash. 20:45
pmurias it's not important
dlocaus ok. 20:46
I'm moving on to the next bug. I'll add this to the TODO list?
pmurias perl5rx is the engine which will be used for the bootstap
add it
dlocaus ok
is there a place where I can store code in the repositry in a kind of "my space" that other people can access? 20:47
like a home directory? I have a nice piece of code that I broke down, that would be useful to the next person who has to debug this.
pmurias misc/pX
20:48 HiTech69 joined
pmurias got to buy cola& (not that I would ever drink that horrid stuff) 20:49
pugs_svn r18782 | dlo++ | [kp6] 15-capture-from-rule-block.t goes into an infinite loop. I can't solve this. Someone else will have to try and give it a stab, see misc/pX/dlocaus/kp6-grammar-test-15.pl 20:59
21:03 ludan joined 21:14 dlocaus left 21:18 buubot joined 21:22 idiotmax left 21:27 justatheory joined 21:35 ilogger2_ joined 21:37 thoughtpolice joined 21:41 zostay joined 21:53 cookys joined 21:54 cookys left 21:55 cookys joined 21:59 Limbic_Region joined 22:05 cookys left 22:06 cookys joined 22:09 cognominal_ joined 22:11 fridim_ joined 22:17 the_dormant joined 22:43 the_dormant left 22:46 cognominal_ left 22:54 fridim_ left 23:05 tlb joined 23:20 baalial joined 23:27 luqui joined 23:29 revdiablo joined 23:31 thoughtpolice left 23:50 baalial left