Camelia, the Perl 6 bug

IRC log for #padre, 2011-10-09

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

All times shown according to UTC.

Time Nick Message
00:00 Di-ima joined #padre
00:05 castaway joined #padre
00:10 asarch joined #padre
01:24 ktaragorn joined #padre
01:47 gyles19 joined #padre
02:20 gyles19 Ok, I've managed to create a way to consistently make Padre seg fault and dump core, by opening a short cgi perl script and inserting a 'use' statement.  I have a gdb backtrace of all the threads but I don't know if it's really that useful.  What can I collect that will help the developers diagnose this?
02:26 gyles19 It occurs to me that it may be very useful if that --version command line option printed everything in the About box, not just padre's own version... because y'all always ask for all of that, and I'm trying to automate my life as much as possible.
03:17 submersible_toaster gyles19:  thanks for report .. can you report platform?
03:18 gyles19 Sure, 64-bit Ubuntu 10.04.3
03:19 gyles19 I find that if I nuke the syntax tab before adding the line, the seg fault doesn't occur.
03:19 waxhead gyles19, a friend of mine is reporting a similar problem with Padre..
03:20 waxhead he's been using Dancer, but I suspect it's not the problem of Dancer, but the use statement you are seeing.
03:20 submersible_toaster gyles19:  padre trunk or from cpan or from apt ?
03:21 gyles19 I don't know what 'Dancer' is... I'm wondering if there's a debug mode I can put padre into to track the source of the problem?  From cpan.  Installing from apt simply doesn't work on Ubuntu, their wxperl package is way too old.
03:21 submersible_toaster 0.76.ds1-1 (padre)
03:21 submersible_toaster on 11.04 yuk that is ancient
03:21 submersible_toaster and the icon is just foul.
03:21 submersible_toaster on .90 from cpan ?
03:22 waxhead gyles19, it's a web framework like Mojolicious or Catalyst, but that's not the point, he's seeing the same segfault thing when typing use.
03:22 submersible_toaster probably a calltip then
03:22 submersible_toaster ?
03:23 gyles19 Perl 5.10.1
03:23 gyles19 Wx 0.9901
03:23 gyles19 wxWidgets 2.8.10
03:23 gyles19 Alien::wxWidgets 0.50
03:23 gyles19 Wx::Perl::ProcessStream 0.30
03:23 waxhead yeah, or the syntax checker is firing too quickly and picking up use on it's own and crashing
03:23 gyles19 It locks up as soon as I type the ; to end the line.
03:24 waxhead oh..
03:24 submersible_toaster standby - i will see if i can reproduce
03:24 gyles19 I have a CGI:::Application, and I want to enhance it with CGI::Application::Plugin::DebugScreen, which is installed on my perl... as soon as I finish typing that 'use' statement Padre seg faults.
03:25 gyles19 I have the back trace of all threads from gdb if you would find that helpful.
03:25 waxhead gyles19, sure...
03:25 gyles19 I can also just tarball the project and send you that if you like.  I can make a copy and trim out unnecessary garbage.
03:25 waxhead gyles19, can you register a bug report?
03:26 gyles19 yes, I have a login there.
03:26 * submersible_toaster smells another BEGIN {} problem
03:26 waxhead gyles19, cool... can you register a bug report with the trace info in it..
03:26 waxhead trac!
03:26 gyles19 What I'd really like to do is create an environment for Padre that will automatically capture what information the devs here will find useful.  I've already automated the back trace and a log of padre's error messages.
03:26 waxhead Hyppolit, trac!
03:26 Hyppolit Trac registration opened for 5 minutes. Please visit http://padre.perlide.org/trac/register to register
03:27 waxhead sigh..
03:27 gyles19 I have a registration already, waxhead, but thanks. ;)
03:27 waxhead I was trying to get the URL to register a bug
03:27 waxhead bug!
03:27 gyles19 Ah, I think that's in the welcome message here, isn't it?
03:27 waxhead Hyppolit, bugreport!
03:27 waxhead Hyppolit, report!
03:28 waxhead haha...
03:28 gyles19 Ah, no, it's not.  That may be a helpful addition.
03:28 gyles19 heheeh.  How about a 'help'?
03:29 waxhead Hyppolit, ticket!
03:29 waxhead ticket!
03:29 waxhead uptime!
03:29 waxhead Hyppolit, uptime!
03:29 waxhead Hyppolit, you dope!
03:30 waxhead uptime
03:30 submersible_toaster gyles19: I see a lockup with no segfault here
03:30 waxhead !uptime
03:30 waxhead submersible_toaster, it's not the first report like this I've seen
03:30 gyles19 Give it time.  It's dumping a 200meg core file...  It takes about 10-15 seconds on my box.
03:30 gyles19 And I've got a i7. ;)
03:30 waxhead ouch!
03:31 gyles19 I'm thinking of doing this dump again, but launching an 'strace' on the perl pid before I make it seg fault.
03:31 waxhead gyles19, thanks!
03:31 Hyppolit Trac registration closed
03:32 waxhead Hyppolit, ticket!
03:32 gyles19 (Now you've got me reading Hyppolit's source to see what commands it uses.)
03:32 waxhead gyles19, I would too, but watching the car racing...
03:33 waxhead Hyppolit ticket!
03:33 waxhead ticket!
03:33 waxhead uptime!
03:33 waxhead meh.. :)
03:33 ktaragorn joined #padre
03:34 submersible_toaster brutal
03:34 submersible_toaster the DebugScreen plugin demands that it's caller has add_callback method (during import);
03:35 gyles19 Yes.  Perhaps that's the underlying cause of the seg fault, but still, the text editor shouldn't care what garbage I type. ;)
03:36 submersible_toaster gyles19: another reason why perl syntax check is evil :)
03:37 submersible_toaster gyles19:  consider  BEGIN { `rm -rf ~/` }
03:37 submersible_toaster just syntax checking that sucker and bye bye homedir
03:39 submersible_toaster gyles19:  thankfully it behaves differently on trunk so 0.92 should have this fixed for you :)
03:40 gyles19 hmm.  Now I wish padre had comments of its own I could embed into a file.  like #padre: no_syntax_check_dumbass
03:53 gyles19 I see that hyppolit can return a url to a bug report or a change set if you speak #<number> or r<number> at it.
03:53 waxhead gyles19, yeah, r16001
03:53 Hyppolit Changeset #16001 http://padre.perlide.org/trac/changeset/16001
03:53 waxhead :)
03:54 gyles19 Oh, that's interesting.  It didn't need its nick in the line.
03:56 waxhead nope
03:56 waxhead uptime
03:56 waxhead uptime!
03:56 gyles19 Ah, I just realized I'm not looking at the most recent checkin for that module.
03:56 waxhead I'm sure that used to work
03:57 gyles19 uptime?
03:57 Hyppolit 20:55:25 up 42 days, 17:09,  2 users,  load average: 0.31, 0.40, 0.43
03:57 gyles19 Ha!
03:58 gyles19 It wants the ?
04:00 gyles19 hyppolit?
04:00 gyles19 Hyppolit?
04:00 Hyppolit Hyppolit is a bot currently running version 0.11
04:00 Hyppolit My master is szabgab.
04:00 gyles19 Case sensitive pattern.
04:03 waxhead ticket?
04:03 Hyppolit ticket is http://padre.perlide.org/trac/wiki/Tickets
04:03 waxhead oh..
04:03 waxhead there you go...
04:03 waxhead I guess that makes sense...
04:03 waxhead it's more of a question...
04:03 waxhead where as trac! is an order...
04:05 gyles19 Ah.  It has a config file that defines more of those command words, and you could theoretically add more of them to its memory if you're a trusted nick.
04:05 gyles19 And none of those need the nick in the command.  It doesn't even need to be the only thing on the command.  That could be amusing.
04:06 gyles19 ticket? Or not to ticket? That is the question.
04:06 gyles19 Ah.  yes, it does. nested if.
04:07 gyles19 Yeah, it looks like ! is only needed on trac!, that is a specific regexp test.
04:10 gyles19 explain: $x ||= $y
04:11 gyles19 Ah, well, either that's disabled or it's renamed to something sekrit.
04:23 gyles19 Bingo.  I can reproduce it with a single source file.  Here comes your bug report.
04:39 Hyppolit # 1343 :  Padre seg faults when syntax checking some 'use' statements (new defect) [ http://padre.perlide.org/trac/ticket/1343 ]
04:40 Hyppolit # 1343 :  Padre seg faults when syntax checking some 'use' statements (new defect) new attachment [ http://padre.perlide.org/trac/ticket/1343 ]
04:41 Hyppolit # 1343 :  Padre seg faults when syntax checking some 'use' statements (new defect) new attachment [ http://padre.perlide.org/trac/ticket/1343 ]
04:41 Hyppolit # 1343 :  Padre seg faults when syntax checking some 'use' statements (new defect) new attachment [ http://padre.perlide.org/trac/ticket/1343 ]
04:41 gyles19 There you go.
04:42 gyles19 so #1343 is mine?
04:42 Hyppolit # 1343 :  Padre seg faults when syntax checking some 'use' statements (new defect) [ http://padre.perlide.org/trac/ticket/1343 ]
04:42 gyles19 Cool...
04:51 waxhead gyles19, thanks!!!
04:55 gyles19 I'm making a wrapper that automates collecting crap of interest.  I've just added perl -V and platform info to the log file it generates for me.
05:48 ktaragorn1 joined #padre
05:50 Hyppolit # 1343 :  Padre seg faults when syntax checking some 'use' statements (new defect) new attachment [ http://padre.perlide.org/trac/ticket/1343 ]
06:18 ecocode joined #padre
06:42 toi joined #padre
07:07 Di-ima joined #padre
07:17 ecocode joined #padre
07:20 gyles19 Eh.  I just discovered padre always exits with an error code... glibc is complaining about a malformed doublely-linked list.
07:21 azawawi joined #padre
07:21 azawawi good morning #padre
07:22 gyles19 G'morning.
07:22 azawawi gyles19: hey welcome... just the one that i needed to talk to :)
07:22 azawawi gyles19: gm
07:22 gyles19 Uh oh!
07:22 azawawi gyles19: i was browsing the log and i found your ticket
07:23 gyles19 The one I created a few hours ago?
07:23 azawawi gyles19: a nice suggestion: ## {use|no} padre_syntax_check
07:23 gyles19 Is that something I can put into my source file to disable the syntax check in that particular file?
07:24 azawawi gyles19: it is something that i can add in this release :)
07:24 gyles19 Ah!
07:24 azawawi gyles19: but let us think of it is useful or not
07:25 azawawi gyles19: do you want to stop syntax checking all file or a portion of it
07:25 gyles19 Well, I want padre to stop dumping core while I'm trying to edit a simple 10-line cgi script. ;)
07:25 azawawi gyles19: what 'no padre_syntax_check' will do in the background is to comment the lines when syntax checking them
07:26 azawawi gyles19: hence exported variables/functions will not be there
07:26 gyles19 That would work.  I could disable syntax checking in just that section.
07:27 azawawi on it :)
07:27 azawawi 'use padre_syntax_check' enables the syntax checking behavior
07:28 azawawi 'no padre_syntax_check' disables it
07:28 gyles19 As a ## coment?
07:28 azawawi yup
07:45 Hyppolit svn: r16859 | azawawi++ | http://padre.perlide.org/trac/changeset/16859
07:45 Hyppolit Decrement plan test count to 21 in t/08_style.t
07:45 Hyppolit trunk/Padre/t/
07:46 Hyppolit svn: r16860 | azawawi++ | http://padre.perlide.org/trac/changeset/16860
07:46 Hyppolit Initial comment to parses syntax checking comments blocks
07:46 Hyppolit To disable Padre syntax check, please use:
07:46 Hyppolit ## no padre_syntax_check
07:46 Hyppolit To enable again:
07:46 Hyppolit ## use padre_syntax_check
07:46 Hyppolit trunk/Padre/lib/Padre/Document/Perl/
07:48 azawawi gyles19: i'll write it later today
07:48 azawawi gyles19: but the initial feature is there
07:48 gyles19 Thanks, Azawawi.  That will be very helpful.
07:48 azawawi another feature i was thinking of
07:49 azawawi is counting the ok(...) is(...) statements in a test.t
07:49 azawawi i think it would be useful for the test suite maintainer to quickly identify https://metacpan.org/module/Test::More keywords and test plan count
07:50 ecocode joined #padre
07:52 gyles19 Except that some tests stick a single ok into a for loop...
07:53 gyles19 I cheat; I tell the test it has no plan and then run it and let TAP tell me how many tests there are.  Assumes it runs the way I think it does...
07:54 azawawi true
07:54 azawawi but what if i include the number of oks in a loop - A
07:55 azawawi i mean give you the count of oks outside loops and inside loops as a report
07:55 azawawi this means you can calculate it more easily
07:55 azawawi that manually going through them
07:55 azawawi s/that/than
07:56 gyles19 I don't follow what you mean by "count...outside loops and inside loops as a report"
07:57 gyles19 Do you mean provide another ## comment for it to know how many tests are within the loop?
07:57 azawawi no
07:58 azawawi a normal test file contains Test::More functions inside and outside loops right?
07:58 azawawi if we parse them and show a report telling the user:
07:58 azawawi 5 statements and 2 statements inside a loop
07:59 azawawi the programmer can determine that 2*number of loop + 5 = test count
07:59 gyles19 Ah.  I see, yes.  That would give the maintainer a clue that he needs to count that test file manually.
08:00 gyles19 My "normal test file" is usually a Test::Class thingy, so this wouldn't work for most of the stuff I do.  I'm an old junit hacker so I prefer the class-based unit-test style.
08:02 azawawi at least highlighting test keywords different would help
08:02 azawawi s/different/differently/
08:08 dodathome joined #padre
08:13 gyles19 Well, I use Test::More inside my Test::Class objects, so as long as this counting could occur in a .pm as well as a .t, it would still work for that purpose.
08:17 waxhead you know in my mind it seemed like a good idea, but now I can't figure out how to pull it together... well, not in the way I wanted to... sigh.
08:18 gyles19 Well, I like the idea of highlighting the Test::More functions in some fashion.
08:19 gyles19 Test::Class uses a Test attribute on the methods you create.  Each of those has a test count on it.
08:19 dolmen joined #padre
08:23 gyles19 Do any of the padre tests check the perl exit code after padre terminates?
08:24 pece2 joined #padre
08:27 gyles19 You know, one other thing about the syntax checker...I have to launch padre with a PERL5lIB=t/lib envirornment variable because Padre isn't looking for a t/lib, so it fails horribly to deal with Test::Class tests being under t/lib instead of ./lib or ./blib.
08:32 Hyppolit svn: r16861 | azawawi++ | http://padre.perlide.org/trac/changeset/16861
08:32 Hyppolit Process ## use|no padre_syntax_check comment pragmas in syntax checker. gyles19++
08:32 Hyppolit trunk/Padre/lib/Padre/Document/Perl/
08:33 gyles19 Ooh, karma.
08:35 Hyppolit svn: r16862 | azawawi++ | http://padre.perlide.org/trac/changeset/16862
08:35 Hyppolit Updated Changes
08:35 Hyppolit trunk/Padre/
08:36 Hyppolit # 1343 :  Padre seg faults when syntax checking some 'use' statements (new defect) by azawawi  [ http://padre.perlide.org/t​rac/ticket/1343#comment:1 ]
08:40 pece joined #padre
08:41 azawawi gyles19: any chance you're on trunk to try it out?
08:41 gyles19 My padre was installed from cpan.  I'm not setup as a developer.
08:49 gyles19 It's almost 4am here.  I'm going to bed.  Chat with you in about 9 hours. ;)
08:50 azawawi cool
08:50 azawawi good night! :)
09:01 Hyppolit svn: r16863 | dolmen++ | http://padre.perlide.org/trac/changeset/16863
09:01 Hyppolit Gabor's name is not translatable.
09:01 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
09:04 azawawi interesting...
09:04 azawawi https://github.com/scrooloose/synt​astic/tree/master/syntax_checkers
09:04 azawawi Padre port coming soon lol :)
09:06 * azawawi downloads "Haskell Platform" for win32 :)
09:14 bowtie joined #padre
09:14 bowtie gm
09:22 Hyppolit svn: r16864 | dolmen++ | http://padre.perlide.org/trac/changeset/16864
09:22 Hyppolit "About" dialog refactoring
09:22 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
09:25 bowtie dolmen, ping
09:25 dolmen bowtie: pong
09:26 bowtie dolmen, I am confused, please explain what was wrong, with encapsulation?
09:27 dolmen $offset is a local variable that is only needed to build the system information
09:27 bowtie dolmen,  yes
09:27 dolmen no need to storing for the whole Dialog::About object lifetime
09:28 dolmen s/to storing/store it/
09:28 bowtie dolmen, thanks
09:28 dolmen I'm doing more changes
09:29 bowtie dolmen, good :)
09:34 shadowpaste "bowtie" at 217.168.150.38 pasted "t/08_style.t (Wstat: 65280 Tests: 22 Failed: 1)" (36 lines) at http://paste.scsys.co.uk/148549
09:35 bowtie azawawi, test 08 now fails :)
09:41 Steffen joined #padre
09:42 azawawi joined #padre
09:44 azawawi bowtie: svn st ?
09:44 bowtie azawawi, ?
09:44 azawawi bowtie: maybe you have a left over style in share/(styles|themes) or something
09:45 azawawi bowtie: and gm :)
09:45 bowtie azawawi, O, silly me, yes I have theme called perldoc.txt thanks :)
09:46 bowtie azawawi, version control panel, when we click on a file it opens in editor, but can we set focus to first diff
09:47 ktaragorn joined #padre
09:53 azawawi bowtie: sure, ticket it :)
09:53 azawawi i have downloaded the "Haskell Platform" :)
09:53 azawawi hehe
09:53 bowtie azawawi, will do
09:54 azawawi given that pugs (Perl 6) was written haskell... it is so powerful! :)
09:54 azawawi s/written/written in/
09:55 Hyppolit svn: r16865 | dolmen++ | http://padre.perlide.org/trac/changeset/16865
09:55 Hyppolit More system info changes in "About" dialog
09:55 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
09:55 azawawi bowtie: have you tried ## (no|use) padre_syntax_check ? :)
09:56 * azawawi writes a formal test case for that
09:56 bowtie azawawi, I thought that was still in development :)
09:57 azawawi i already implemented it :)
09:57 azawawi when im excited about an idea, i usually jump in and do it right away
10:01 Hyppolit # 1344 :  version control panel, set focus, on open, to first diff found (new enhancement) [ http://padre.perlide.org/trac/ticket/1344 ]
10:02 bowtie azawawi, ok, so how do we turn on and off, is it per file or section?
10:06 azawawi per file... same ## no critic
10:06 azawawi ## use critic...
10:10 shadowpaste "bowtie" at 217.168.150.38 pasted "## no padre_syntax_check" (18 lines) at http://paste.scsys.co.uk/148551
10:11 bowtie azawawi, have I just confused myself, or?
10:13 azawawi why did you put use CGI::Application::Plugin::DebugScreen;  # die! after ## use padre_syntax_check ?
10:13 azawawi :)
10:16 bowtie azawawi, I thought i would use the test script from backlog, I was testing the switch to see what happens, syntax error is now on empty line 16 which must be wrong, hence I am either confused or syntax checker is confused :) or both :)
10:26 Hyppolit svn: r16866 | azawawi++ | http://padre.perlide.org/trac/changeset/16866
10:26 Hyppolit Added the following tests to t/44_perl_syntax.t:
10:26 Hyppolit - Syntax check off/on pragma block
10:26 Hyppolit - Syntax check off/on pragma block and then error
10:26 Hyppolit - Syntax check off pragma block
10:27 Hyppolit - Syntax check off pragma misspelled
10:27 Hyppolit
10:27 Hyppolit Enjoy :)
10:27 Hyppolit trunk/Padre/t/
10:38 Hyppolit svn: r16867 | bowtie++ | http://padre.perlide.org/trac/changeset/16867
10:38 Hyppolit tweak
10:38 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
10:48 asarch joined #padre
10:54 Hyppolit svn: r16868 | azawawi++ | http://padre.perlide.org/trac/changeset/16868
10:54 Hyppolit When clicking on an annotation, it shows the non-visible syntax check panel without losing focus on the editor
10:54 Hyppolit trunk/Padre/lib/Padre/Wx/
11:12 Hyppolit svn: r16869 | azawawi++ | http://padre.perlide.org/trac/changeset/16869
11:12 Hyppolit Added lang_perl6_auto_detection which enables Padre to detect Perl 6 files       in Perl 5 files. Please note that this is not an accurate method of       detecting Perl 6 files (disabled by default). The previous behavior was to       enable it when the Perl 6 plugin is enabled.
11:12 Hyppolit trunk/Padre/ trunk/Padre/lib/Padre/ trunk/Padre/t/
11:12 Hyppolit svn: r16870 | szabgab++ | http://padre.perlide.org/trac/changeset/16870
11:12 Hyppolit Update the messages.pot file
11:12 Hyppolit trunk/Padre/share/locale/
11:21 Hyppolit svn: r16871 | azawawi++ | http://padre.perlide.org/trac/changeset/16871
11:21 Hyppolit Added "Language Perl 6" tab to preferences.
11:21 Hyppolit trunk/Padre/ trunk/Padre/lib/Padre/Wx/FBP/
11:28 Hyppolit # 1341 :  Wx-Scintilla build fails on wxWidgets 2.8.10 (new defect) by azawawi  [ http://padre.perlide.org/t​rac/ticket/1341#comment:1 ]
11:28 Hyppolit svn: r16872 | azawawi++ | http://padre.perlide.org/trac/changeset/16872
11:28 Hyppolit Fix ticket:1341 'Wx-Scintilla build fails on wxWidgets 2.8.10'. markd++
11:28 Hyppolit trunk/Wx-Scintilla/ trunk/Wx-Scintilla/wx-scintilla/src/scintilla/src/
11:30 Hyppolit svn: r16873 | azawawi++ | http://padre.perlide.org/trac/changeset/16873
11:30 Hyppolit Updated Changes for Wx::Scintilla 0.31_01
11:30 Hyppolit trunk/Wx-Scintilla/
11:31 Hyppolit svn: r16874 | bowtie++ | http://padre.perlide.org/trac/changeset/16874
11:31 Hyppolit another tweak for continuity
11:31 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
11:32 Hyppolit # 1341 :  Wx-Scintilla build fails on wxWidgets 2.8.10 (closed defect) by azawawi  [ http://padre.perlide.org/t​rac/ticket/1341#comment:3 ]
11:35 bowtie azawawi, we have "preferred language for error diagnostics" preferences what dose this do?
11:35 Hyppolit # 1334 :  Syntax checker could be slower (new enhancement) by azawawi  [ http://padre.perlide.org/t​rac/ticket/1334#comment:3 ]
11:35 azawawi bowtie: error diagnostics are localized
11:36 azawawi bowtie: but there is very little translated
11:36 azawawi http://search.cpan.org/~polgab/POD2-FR-0.03/FR.pm
11:37 azawawi https://metacpan.org/search?q=POD2
11:37 bowtie azawawi, we have a preferences option with out any options, that's all :)
11:37 azawawi yup
11:37 azawawi ask szabgab about it :)
11:39 bowtie azawawi, szabgab the font of all knowledge :) will do
11:39 Hyppolit svn: r16875 | azawawi++ | http://padre.perlide.org/trac/changeset/16875
11:39 Hyppolit Updated Changes
11:39 Hyppolit trunk/Padre/
11:40 bowtie azawawi, I have been looking at debugging in Padre, Ouch, why do we use Debug::Client when others use Padwalker ?
11:42 * azawawi prepares Wx::Scintilla 0.31_01
11:48 azawawi cpanm http://cpan.metacpan.org/authors/id/A/​AZ/AZAWAWI/Wx-Scintilla-0.31_01.tar.gz
11:59 Sewi joined #padre
13:00 ktaragorn joined #padre
13:34 kaare joined #padre
14:16 * azawawi starts working on escape sequence Perl lexer highlighting :)
14:44 azawawi submersible_toaster: ping
14:45 bowtie azawawi, submersible_toaster is on gmt +10
14:45 azawawi bowtie: re #1331
14:45 Hyppolit # 1331 :  Specially Handle Escaped characters (new enhancement) by submersible_toaster  [ http://padre.perlide.org/t​rac/ticket/1331#comment:2 ]
14:46 ecocode joined #padre
14:46 azawawi is that escape chars list complete or not in your opinion?
14:47 bowtie azawawi, how do i test for a padre.yml, that that is in root of project
14:48 azawawi Current->project?
14:50 bowtie azawawi, is there any reason why I can not store info inside this file and create if dose not exist?
14:50 azawawi bowtie: i dont follow you
14:50 azawawi bowtie: why? :)
14:50 Steffen joined #padre
14:51 bowtie azawawi, I have been looking at debug, at present it stores breakpoints in object->encapsulation why not in padre.yml :)
14:52 azawawi unfortunately i have to go home now
14:52 azawawi will try to continue discussion from home :)
14:52 azawawi please see #1331
14:52 Hyppolit # 1331 :  Specially Handle Escaped characters (new enhancement) by submersible_toaster  [ http://padre.perlide.org/t​rac/ticket/1331#comment:2 ]
14:52 bowtie ok
14:52 azawawi and try to comment if these list is complete or not
14:53 azawawi thanks
14:53 * azawawi home &
14:53 dolmen joined #padre
14:53 ktaragorn1 joined #padre
15:13 Hyppolit svn: r16876 | szabgab++ | http://padre.perlide.org/trac/changeset/16876
15:13 Hyppolit Update the messages.pot file
15:13 Hyppolit trunk/Padre/share/locale/
15:37 pece2 joined #padre
15:48 bowtie azawawi, this: external_diff_tool => '' is still in Padre::Current->config even though it has been removed from preferences
16:57 Hyppolit # 1343 :  Padre seg faults when syntax checking some 'use' statements (new defect) by gyles19  [ http://padre.perlide.org/t​rac/ticket/1343#comment:2 ]
17:01 submersible_toaster joined #padre
17:13 gyles19 I'm writing a bash script launcher for padre that will gather debugging foo when padre crashes.  (Which for me, seems to be about every 5 minutes.)   This will run on linux, might run okay on cygwin (haven't tried building padre there yet, you don't have a binary for it.)
18:29 chorny joined #padre
18:43 Sewi joined #padre
19:06 Hyppolit # 1328 :  Padre infinite loops when launched without a filename to open (new defect) new attachment [ http://padre.perlide.org/trac/ticket/1328 ]
19:23 Hyppolit # 204 :  Scalars leaked: 1 (assigned defect) by gyles19  [ http://padre.perlide.org/trac/ticket/204#comment:4 ]
19:58 toi joined #padre
19:59 Hyppolit # 1331 :  Specially Handle Escaped characters (new enhancement) by bowtie  [ http://padre.perlide.org/t​rac/ticket/1331#comment:3 ]
20:01 sugar joined #padre
20:31 asarch joined #padre
21:18 smpb joined #padre
21:46 gyles19 Oh, surely not.
21:55 Hyppolit # 1345 :  actionqueue=internal.dump_padre is writing to a closed filehandle. (new defect) [ http://padre.perlide.org/trac/ticket/1345 ]
21:56 asarch joined #padre
22:02 smpb joined #padre
22:24 Hyppolit # 1345 :  actionqueue=internal.dump_padre is writing to a closed filehandle. (new defect) new attachment [ http://padre.perlide.org/trac/ticket/1345 ]
22:26 Hyppolit # 1345 :  actionqueue=internal.dump_padre is writing to a closed filehandle. (new defect) by gyles19  [ http://padre.perlide.org/t​rac/ticket/1345#comment:1 ]
22:50 Alias gyles19: We don't officially support cygwin, and Windows has no issues with crashing
22:50 Alias gyles19: If your thingy only works on Linux, then that's fine
22:50 gyles19 My thingy?
22:54 Alias <gyles19>I'm writing a bash script launcher for padre that will gather debugging foo when padre crashes.
22:56 gyles19 Ah.  This thing I'm writing should be generic enough to run under macOSX as well, but I don't have a machine to test it on.  It would probably need some tweaking there.
23:08 submersible_work Alias: Padre-0.90 on windows hangs for an age on the bug gyles19 reports .
23:09 gyles19 It seems to be going into an infinite loop, if I'm fast enough I can look at 'top' and see padre consuming 100% cpu until it segfaults.
23:09 Alias Where is this?
23:10 submersible_work paste?
23:10 gyles19 I think he's referring to #1343
23:10 Hyppolit # 1343 :  Padre seg faults when syntax checking some 'use' statements (new defect) by gyles19  [ http://padre.perlide.org/t​rac/ticket/1343#comment:2 ]
23:10 submersible_work yup
23:12 Alias Curious it crashes
23:12 Alias Padre absolutely suffers from security problems, the ones I originally described in my first PPI talk at OSCON
23:12 Alias We execute BEGIN blocks arbitrarily
23:12 Alias I had meant to fix that problem before 1.0 :)
23:13 submersible_work Alias: i think it's a funky ::import - not actually a BEGIn
23:13 Alias Same thing
23:13 submersible_work quite
23:14 Alias But a crash in the process execution in a background thread shouldn't cause a crash
23:14 gyles19 This is what prompted azawawi to create the 'use padre_syntax_check' padre pragma comment thing earlier today.
23:14 Alias I saw the commit message
23:14 Alias That's a really bad idea
23:14 Alias Makes Padre more complicated, leaves shit in that does nothing in other editors, and don't actually solve the problem
23:15 Alias gyles19: Can you replicate the crash under perl dev -t ?
23:15 gyles19 It's not unknown.  Vim uses the same sort of thing for things like folding.
23:15 gyles19 perl dev -t?
23:15 Alias I know, it's stupid there too
23:15 Alias :)
23:16 gyles19 I just run padre off cpan.
23:16 Alias Can you replicate on svn version?
23:17 gyles19 I don't have svn installed.  I'm trying to use padre to do real work but I spend more time fighting with it dumping core than actually doing the work really want to be working on. (Stuff for my office.)
23:17 gyles19 I believe azawawi replicated the problem
23:19 gyles19 That die_padre_die.pl file is a simple reproduction of the issue.  I expect if you try that file in your truck it will show you.
23:19 gyles19 s/truck/trunk/
23:19 Hyppolit # 261 :  When Document::XYZ.pm is not available padre crashes on file open (closed defect) by adamk  [ http://padre.perlide.org/trac/ticket/261#comment:1 ]
23:20 submersible_work gyles19: if you shim it by preceeding a sub add_callback {}; does that help
23:20 gyles19 In the die.pl?  I'll try that.
23:22 gyles19 Yes, it does.  And if I then comment that shim, Padre will die again.
23:23 gyles19 Signal SEGV at /usr/local/share/perl/5.10.1/Padre/Wx/Syntax.pm line 352
23:23 gyles19 Padre::Wx::Syntax::task_finish('Pa​dre::Wx::Syntax=HASH(0x72b7010)', 'Padre::Document::Perl::Syntax=HASH(0x7327a98)') called at /usr/local/share/perl/5.10.1/Padre/Task.pm line 495
23:24 Alias ok, doesn't crash for me on Windows
23:24 submersible_work Alias: how long does it hang for?
23:25 gyles19 Doesn't surprise me...windows lets apps scribble all over ram without complaining.  Except 7, I think they finally fixed that there.
23:25 Alias It doesn't
23:25 submersible_work and you have the plugin installed ?
23:25 Alias gyles19: We spawn a process, capture the output, and that's all
23:25 Alias Yes
23:25 Alias Can't locate object method 'add_callback'
23:25 Alias etc
23:26 gyles19 Who are you asking about the plugin, and which plugin?
23:27 Alias me, and CGI::Application::Plugin::DebugScreen
23:28 gyles19 Oh, yes, I do have that plugin installed.  I noticed this issue because I had an earlier version of my padre wrapper in use, and I was trying to add that plugin to my CGI::Application  project... which diverted me completely.
23:28 submersible_work gyles19: fixed on trunk it seems for ubuntu 10.10
23:29 Alias I'm mostly concerned that a crash can propogate across the "external execution" of the program
23:29 gyles19 I'm on 10.04, what changed for 10.10?
23:29 gyles19 Where do I find it?
23:29 Alias It's not like we're compile testing via the dll, we should be just doing a "perl -c blah"
23:29 Alias That would imply running stuff in other non-Perl programs that segfault cascase over as well
23:29 submersible_work gyles19:  no - I mean using trunk and testing on 10.10 I cannot reproduce.   Padre-0.90 does have the problem on same host.
23:30 gyles19 Padre and wx are completely broken on native 10.04, I'm running everything custom-built.
23:30 submersible_work gyles19: yeah - we know :)
23:31 Alias We'll get there :)
23:31 gyles19 (which is why I made the wrapper, to grab the versions of everything you seem to care about, plus the usual platform stuff.  when I was doing TeamB for jbuilder I had a data collection script like this I'd have folks run and post the results.
23:37 Hyppolit # 1327 :  Padre trunk crashes on Ubuntu 10.04 (new defect) by adamk  [ http://padre.perlide.org/t​rac/ticket/1327#comment:12 ]

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