Camelia, the Perl 6 bug

IRC log for #padre, 2012-12-12

| Channels | #padre index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
01:34 perlite_ joined #padre
01:36 botbot Welcome perlite
03:43 Foxcool joined #padre
05:24 ssotka joined #padre
05:34 asarch joined #padre
05:40 Foxcool joined #padre
06:33 GlitchMr joined #padre
07:47 sugar joined #padre
08:24 dolmen joined #padre
08:39 dod joined #padre
08:41 Foxcool joined #padre
08:56 dod joined #padre
09:13 dod joined #padre
09:55 sivoais joined #padre
10:18 sewi joined #padre
10:21 SvenDowideit joined #padre
11:19 sivoais joined #padre
11:48 Foxcool joined #padre
11:49 Foxcool unexpected end of string while parsing JSON string, at character offset 145 (before ","url":"http://modul...") at /usr/share/perl5/vendor_pe​rl/CPAN/Meta/Converter.pm line 24.
11:49 Foxcool at /usr/share/perl5/ExtUtils/MM_Any.pm line 824.
11:51 Foxcool Padre 0,96 could not installed
11:51 Foxcool Fedora 18 beta (:
11:54 bowtie joined #padre
12:16 bowtie SvenDowideit, hi
12:17 SvenDowideit heya :) just before i goto bed :)
12:18 SvenDowideit hope that commit isn't to terrible >:}
12:18 bowtie just, a couple of question, as I am a silly billy
12:20 bowtie looked at db::history with cookbook-recipe04, do you think you should run_script_args == full name rather , think of files in a script dir
12:21 SvenDowideit yes, i'm wondering that too
12:22 SvenDowideit and today i was even hit by it :)
12:23 bowtie wx-panel-debugger line 307 -> chdir $dir;
12:24 SvenDowideit y?
12:24 SvenDowideit i copied that one from the 'run tests' code
12:25 bowtie ok, test name please
12:25 SvenDowideit which is why i suspect the chdir to and back should go into the run_command() call itself eventually
12:25 SvenDowideit the UI
12:26 SvenDowideit Wx::Main line 2843 ish
12:26 bowtie the debugger dose not use it, it ask the question where are we, remember dot '.'
12:26 SvenDowideit huh?
12:26 SvenDowideit i added the
12:26 SvenDowideit chdir $arg_ref->{run_directory} if ( exists( $arg_ref->{run_directory} ) );
12:26 SvenDowideit
12:27 SvenDowideit line to Debugger.pm
12:27 bowtie the location, is one of your variables you pass to startup
12:27 SvenDowideit so that the DebugOptions dialog can set a different dir
12:27 SvenDowideit and then afterwards, i'm setting the dir back
12:27 SvenDowideit in case it matters
12:27 SvenDowideit (mmm, ok, is that your q?)
12:27 bowtie ok, got it, silly me
12:27 SvenDowideit sweet :)
12:28 SvenDowideit you give me a nice idea
12:28 SvenDowideit tie to $pwd
12:28 SvenDowideit should chdir for you :)
12:28 SvenDowideit making it possible to local $pwd = '/etc';
12:28 bowtie it's a reset
12:29 SvenDowideit and then to magically be back where you where when you leave the scope
12:29 SvenDowideit what a fun idea ;)
12:30 * SvenDowideit goes to see what cpan thinks :)
12:31 bowtie ok, so now we need to indicate to user that the file is being debugged, with manual override options
12:31 SvenDowideit ah yes
12:31 SvenDowideit i was wondering about things like that
12:31 SvenDowideit and also adding more debug things to the toolbar
12:32 bowtie so when they repeat in future, and get confused as to whats going on
12:32 SvenDowideit oh another fun thing
12:32 SvenDowideit it currently really _is_ the options
12:32 SvenDowideit in that after you run once with those options
12:32 SvenDowideit they are used for non-options debug run too
12:33 SvenDowideit as there's only one history foreach file
12:33 SvenDowideit mmm, i wonder if ingy has already done the tied $pwd in IO::All
12:35 bowtie db::history has a tendency to grow quite large
12:35 SvenDowideit mmm
12:36 bowtie use pp-cookbook to see for your self
12:36 SvenDowideit is there a 'keep the last X entries' of type?
12:36 bowtie lol, of cause not
12:36 SvenDowideit excellent :)
12:37 SvenDowideit i'll have to toss that on my list
12:37 bowtie but you can clean the crap out of it with pp-cookbook
12:37 SvenDowideit as i'm clearly going to make 'large' worse
12:37 SvenDowideit i can, if i'm an advanced user
12:37 SvenDowideit but i'm not :)
12:38 bowtie I normaly manage to keep mind below 5k tuples but some others have been known to be 100k tuples
12:41 bowtie as part of Debug2, i deliberately reduced the number of tool-bar icons, but we could add indication to the debugger panel
12:42 bowtie can we clear a files debug-launch-options ?
12:42 SvenDowideit i keep coming back to the idea that users that come to padre will have expectations
12:43 SvenDowideit and those expectations will be that its similar to other ide debuggers
12:43 bowtie we all have those in our dreams, it's called youth :)
12:44 SvenDowideit i guess i could just make a pp-sven-debugger plugin to make the UI that i prefer to use
12:45 bowtie which ide are you referring to
12:45 SvenDowideit no - you're right, i have not provided a way to discard the options that i set
12:46 SvenDowideit well, i've only used um, turbo(pascal|prolog|c|c++), visual(basic|c|c#), eclipse, komodo, python, er, ok, sorry, thats just the small set i can quicly recal
12:46 bowtie which means I can't test for to display message, you have launch-options set
12:46 SvenDowideit and they all have a very similar UI
12:46 bowtie um?
12:47 SvenDowideit sorry, 2 converstations at once :)
12:47 bowtie lol
12:47 SvenDowideit to back up a little
12:47 SvenDowideit i wrote the options dialog with the presumtion that it was setting the debug options for that file/script combo
12:48 bowtie ok
12:48 SvenDowideit i did not think of it as a seperate debug launch
12:48 SvenDowideit but you are right, without a way to delete those settings, it is a bit presumptious
12:49 bowtie the debugger panel or output panel will need to display that we are using launch-options
12:49 SvenDowideit might have to think a little tho - i've made things a bit magical
12:50 SvenDowideit you're always using debug-options
12:50 bowtie no
12:50 SvenDowideit yes, you are (under the hood)
12:50 SvenDowideit this dialog is only to let you change those options
12:50 bowtie I have not had this shining new hack to provide this,
12:51 SvenDowideit the 'launch' button is a little misleading - but very convenient
12:51 bowtie hence 'debug-launch-options'
12:51 SvenDowideit basically, before my code change
12:52 SvenDowideit there were hard-coded debug launch options
12:52 SvenDowideit which set scriptname==filename
12:52 SvenDowideit now the user can change the options
12:53 bowtie I think we should be able to show 'debug-launch-options' in the top of the Debugger-Output Panel as that has some unused real-estate
12:53 SvenDowideit i agree that in either case, it would be useful for the debug output to tell the user what was launched
12:55 SvenDowideit on top - next to 'Status'?
12:55 SvenDowideit tbh, i've not looked at the debug-output panel yet
12:55 bowtie I will have a look at that, when you are happy that your parameter name are static
12:56 SvenDowideit sorry - can you re-phrase 'your parameter name are static'
12:56 SvenDowideit i'm not sure what that means
12:57 bowtie $arg_ref
12:57 bowtie $arg_ref->{script_args}
12:58 SvenDowideit ok, yup, makes sense
12:58 bowtie we have lots of width, but no hight in that panel at top
12:58 SvenDowideit personally i have neither much
12:59 SvenDowideit running the debugger in my 1024x768 notebook is terrible
12:59 bowtie ok, what is your screen size, cga :)
12:59 SvenDowideit i'd prefer to be able to control the visibility of the 3 debug panels (ie, only manually show the breakpoints panel)
13:00 SvenDowideit i've a very nice 12 inch x61 thinkpad
13:01 SvenDowideit and at the moment, i only work on padre on it - as its 'relaxation'
13:01 SvenDowideit if i goto my desktop, i have to work on sap/perl/foswiki integration
13:01 SvenDowideit and deal with a stack of painful things, so i'm partitioning myself >:}
13:02 SvenDowideit mmm, time to goto bed for me
13:02 SvenDowideit i'll have to add a few things to the trac task to remind me
13:02 SvenDowideit laters :)
13:03 bowtie see other first
14:36 kaare joined #padre
15:20 GlitchMr joined #padre
16:10 sugar joined #padre
16:34 vikdes joined #padre
16:39 kaare joined #padre
17:39 dolmen joined #padre
17:55 dod Weird: when running padre/trunk with debugperl (perl with debugger symbols), I get this error: panic: realloc from wrong pool at lib/Padre.pm line 191.
18:05 dod see http://www.nntp.perl.org/group/perl.​perl5.porters/2009/03/msg144752.html
18:05 * dod thinks that threads are involved there
18:07 * dod guesses wildly that storing data in %ENV should be done before creating threads
18:26 bowtie dod, hi
18:27 dod hello
18:27 dod I found some weird stuff ^^
18:28 dod these are wild guesse, I know next to nothing about threads
18:28 bowtie yes, were you running 'perl dev -d' or 'perl dev -t'
18:30 dod Hmm, that was debian's /usr/bin/debugperl provided by perl-debug package  ( http://packages.debian.org/sid/perl-debug )
18:31 dod I used debugperl to analyse a padre core dump
18:32 bowtie dod, you would have to ask alias about this I think
18:32 dod hmm, okay
18:33 bowtie dod,  if recall you debug Padre with 'perl dev -d'
18:33 bowtie in trunk
18:34 dod okay, let's see if I can trigger another core dump
18:35 dod right... I was using perl -Ilib script/padre :o)
18:35 bowtie yes, just checked, it's a feature that alias built in, but only for trunk
18:36 bowtie dod, on my poor old mc it's a bit slow
18:37 dod hmm, working fine now. And the icon in About is the butterfly instead of Debian's image. I guess that I was halfway between padre 0.96 and 0.97 with my previous setup :-/
18:38 dod sorry for the noise
18:38 bowtie dod, alias is the only user I have asked about debugging padre, it's hard work
18:38 bowtie dod, what noise ?
18:39 dod I think that the issues I had were because I invoke padre in trunk with perl -Ilib
18:39 bowtie good friendly banta
18:39 bowtie lol
18:39 dod okay :-D
18:39 bowtie quite an experience :)
18:41 El_Che lo
18:41 bowtie El_Che, hi
18:42 El_Che damnn, while building a 64-bit perl on SOlaris I end with a 32_bit lib somewhere breking embedded perl
18:42 bowtie what fun
18:44 El_Che compiling by hand, perfect
18:44 El_Che doing it by the gar build system: 1 failing test
18:46 El_Che probably -Dvendorarch doing weird stuff again
18:51 bowtie El_Che, is that -Dvendorarch misbehaving or just SOlaris
18:55 El_Che there is something with the build system I think
18:55 El_Che when I supply it on my manual build it's ok
18:56 Di-ima joined #padre
18:56 El_Che when using the build system it picks an installed perl on that vendorarch path magically
18:58 stkowski joined #padre
19:04 El_Che the installed perl is an old 5.10 32-bit
19:04 El_Che while I am working on 5.16.*, 64-bit
19:07 bowtie El_Che, are you saying you need to be running a 64bit to build a 64bit
19:07 El_Che no
19:08 El_Che I am saying that there is something weird in the build system going on
19:08 El_Che doing it manually or though own scripts is ok
19:08 bowtie ok
19:09 El_Che but being part of a soft repository, better to do things as the rest
19:09 El_Che now that blastwave is dead, sunfreeware and activestate behind a paywall, I hope to help make OpenCSW the place to go for upt-to-date Solaris (sparc and x86) perl binaries
19:10 El_Che the most recent pre-compiled stuff you can get on the internet is 5.10
19:10 El_Che and that's already out of upstream support
19:22 bowtie El_Che, sounds like you are paddling like crazy just trying to keep up
19:23 bowtie what about 5.12, 5.14 and 5.16
19:25 El_Che they will not see the sunlight :)
19:25 El_Che I had good support from p5p to get some solaris bugs fixed or worked around
19:32 bowtie thanks for info
19:32 El_Che just venting :)
20:44 sugar joined #padre
21:10 sugar joined #padre
22:14 mib_neild joined #padre
22:14 botbot Welcome mib_neild
23:20 user_2134 joined #padre
23:21 botbot Welcome user_2134
23:39 bowtie SvenDowideit, take a look at code, as I think there is a side affect, you can not now run a program from within padre :(

| Channels | #padre index | Today | | Search | Google Search | Plain-Text | summary