Camelia, the Perl 6 bug

IRC log for #padre, 2011-02-22

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

All times shown according to UTC.

Time Nick Message
00:39 itcharlie joined #padre
00:47 static left #padre
00:48 jazzanova joined #padre
00:49 Hyppolit svn: r13818 | adamk++ | http://padre.perlide.org/trac/changeset/13818
00:49 Hyppolit Renamed Padre::Task::FileRemove to Padre::Task::File to as a more generic filesystem task
00:49 Hyppolit trunk/Padre/ trunk/Padre/lib/Padre/Task/
00:49 asarch joined #padre
00:52 Alias_ I think I'm going to rewrite the recent files menu
00:52 Alias_ It's rather foreground'y and I open a lot of files on network shares
00:53 Alias_ Move it into the background or something
01:11 Maltus left #padre
01:22 user_6347 joined #padre
02:37 jnap joined #padre
02:49 jnap left #padre
02:53 jnap joined #padre
03:03 jnap left #padre
03:05 itcharlie left #padre
03:19 jnap joined #padre
03:58 jnap left #padre
03:59 jnap joined #padre
04:13 Hyppolit svn: r13819 | adamk++ | http://padre.perlide.org/trac/changeset/13819
04:13 Hyppolit A few nudges to fix xt tests
04:13 Hyppolit trunk/Padre/lib/Padre/
04:29 Alias_ 15_locale.t is still dieing for me
04:29 jnap left #padre
04:44 asarch left #padre
04:54 asarch joined #padre
05:13 asarch left #padre
05:22 Hyppolit svn: r13820 | adamk++ | http://padre.perlide.org/trac/changeset/13820
05:22 Hyppolit The recent files list is now generated in the background
05:22 Hyppolit trunk/Padre/ trunk/Padre/lib/Padre/Task/ trunk/Padre/lib/Padre/Wx/ trunk/Padre/lib/Padre/Wx/Menu/
05:26 kanishka joined #padre
05:27 Alias_ AND it supports UNC paths :)
05:27 Alias_ Which is pretty important, since my work home directory is a UNC path
06:11 szabgab Alias_: have you svn uped?
06:11 szabgab since I changed the t/15 to be TODO
06:12 Alias_ Parse errors: Bad plan.  You planned 11 tests but ran 7.
06:12 Alias_ I appear to be current
06:12 szabgab hmm
06:12 szabgab let me check
06:14 szabgab maybe it cannot open the messages.pot ? on line 50 ?
06:14 Alias_ Died at line 50
06:15 szabgab I'll check on windows
06:19 szabgab wtf when I run the test on windows cwd() returns c:/Users/gabor
06:20 szabgab and not the directory where the test lives
06:20 Alias_ Define "run the tesT"
06:20 szabgab perl t/15..
06:21 szabgab it is probably done by  "use Padre:
06:21 szabgab "
06:21 Alias_ Correct
06:21 szabgab crap
06:21 Alias_ Padre moves the Cwd around to prevent locks
06:21 szabgab so I have to fetch cwd before that
06:21 Alias_ Yup
06:21 Alias_ Well, maybe
06:22 Alias_ It doesn't do it in use Padre
06:22 Alias_ I think it does it in Padre->new
06:22 Alias_ Correct
06:22 Alias_ The constructor records the startup cwd in $main->{cwd} or something, and moves the Cwd to $HOME
06:24 Hyppolit svn: r13821 | szabgab++ | http://padre.perlide.org/trac/changeset/13821
06:24 Hyppolit fix the locale test on Windows as well by fetching the cwd before Padre.pm has a chance to move the cwd around
06:24 Hyppolit trunk/Padre/t/
06:25 szabgab pls try now
06:25 Alias_ szabgab: What I usually do is something like this early
06:25 Alias_ my $t = File::Spec->rel2abs('t');
06:25 Alias_ And then File::Spec->catdir/file from there
06:26 Alias_ Certainly before you construct a Padre instance anyway
06:26 Alias_ But I assume you're doing something similar
06:26 szabgab I use cwd() for that
06:26 Alias_ rel2abs does it for you
06:27 Alias_ But whatever, doesn't matter :)
06:27 Alias_ 15_locale.t passed
06:27 Alias_ Thanks
06:27 Alias_ hrm, I think the startup is looking faster
06:27 Alias_ Now that recent files is done in the background
06:28 Hyppolit svn: r13822 | adamk++ | http://padre.perlide.org/trac/changeset/13822
06:28 Hyppolit Typo
06:28 Hyppolit trunk/Padre/lib/Padre/Wx/Menu/
06:29 szabgab I just had a short e-mail exchange with yet another company that has phased Perl out replacing it by Java
06:30 szabgab the 2 main reasons I was told were    1) no steady supply of Perl developers   2) Packaging applications (including dependencies) is much easier in Java
06:30 szabgab and as I understand the CEO is actually a Perl enthusiast
06:30 Alias_ I pretty much agree with that
06:30 szabgab or at least was
06:31 Alias_ And the latter is more important I think
06:31 Alias_ Easy of use drives users, drives developer availability
06:31 szabgab the latter is a technical issue that seems to be easier to solve
06:31 Alias_ It's a technical issue, but it's still hard
06:32 Alias_ We don't define packaging standards that are any good
06:32 szabgab I should check out the citrusPerl and see if it can help us
06:32 Alias_ PAR and the @INC hooks are hacks
06:33 Alias_ Driven, I guess, but lack of thousands of hours of paid development time
06:34 szabgab btw do you have blog now or since use.perl.org was frozen you don't have one?
06:35 Alias_ I have one, I just don't write in it much
06:35 Alias_ blogs.perl.org
06:35 Alias_ It's not only about use.perl.org
06:35 Alias_ It's also about work/life balance :)
06:35 szabgab ali.as links to use.perl.org
06:35 Alias_ More derby, less time to idly rant on the interwebs
06:36 szabgab skating?
06:36 Alias_ Yup
06:36 szabgab what derby btw?
06:37 Alias_ http://www.sydneyrollerderby.com/
06:37 Alias_ http://en.wikipedia.org/wiki/Roller_derby
06:37 Alias_ Of the remaining private Perl time, and leachable work time, I spend most of it actually hacking
06:38 Alias_ Padre and important CPAN modules, and the odd minor new toy
06:38 Alias_ Less time for 1-2 hour long blog post authoring
06:38 Alias_ Maybe I'll pick up the tempo once the new blog is syndicated and has built up a bit of a post history
06:38 szabgab yeah, it takes a lot of time to write a blog post
06:39 Alias_ I always treated it more as a journal
06:39 Alias_ Lets me write important things down once, so I don't have to repeat it over and over again in IRC
06:39 Alias_ But I'm becoming dissalusioned with text for a few things
06:39 Sewi joined #padre
06:39 Alias_ Trying to learn some basic video stuff
06:40 Alias_ Even though that can mean even more time
06:40 szabgab I have to go and test my car
06:40 Alias_ later
06:40 szabgab prove t/car.t
06:40 szabgab &
06:48 kaare joined #padre
06:52 Alias_ hrm
06:52 Alias_ I found a major bug in the release
06:55 Hyppolit svn: r13823 | adamk++ | http://padre.perlide.org/trac/changeset/13823
06:55 Hyppolit !!!MAJOR BUG FIX!!!
06:55 Hyppolit
06:55 Hyppolit When the TaskManager attempts to spawn a new worker, while also inside of an active "DB" lock (via Padre::Locker) the database needs to be disconnected at the moment of the thread spawn.
06:55 Hyppolit
06:55 Hyppolit The code that reconnects to the database uses the non-existant method Padre::DB->begin_work instead of Padre::DB->begin, resulting in corrupt TaskManager state and a failed unlock.
06:55 Hyppolit trunk/Padre/lib/Padre/
06:56 Sewi I saw this error but didn't see any problems which seem to be related to it.
06:56 Sewi alias++ for fixing
07:00 Alias_ It's clogging up the task manager with failed tasks
07:00 Alias_ It's either the source of the clogged task manager, or more likely obscuring the real problem and sending me down false side-paths
07:08 Alias That needs to go on the release branch
07:10 Sewi Micht this problem result in freezing Padre?
07:11 Alias It crashes the unlock
07:11 Alias But I put in an eval wrapper a while back so it shouldn't completely freeze padre
07:11 Alias Still, it means parts of Padre might die
07:12 Alias directory tree, function list, recent files
07:14 waxhead will this go into the release branch?
07:14 waxhead or, should I merge the branch back to trunk, and then re-branch for a release?
07:14 Alias no no
07:14 Alias Change one line
07:15 Alias Padre::DB->begin_work;    changes to Padre::DB->begin;
07:15 Alias Do it directly on the branch
07:15 Alias It's only line, but it's at pretty much the worse most complicated moment
07:24 pece joined #padre
08:04 marcela joined #padre
08:12 fenderson hi
08:13 Alias hi
08:15 danlucraft joined #padre
08:15 waxhead back
08:15 waxhead ok.. I'll change it
08:17 Hyppolit svn: r13824 | waxhead++ | http://padre.perlide.org/trac/changeset/13824
08:17 Hyppolit Commiting the change from trunk to the branch.
08:17 Hyppolit branches/release-0.82/lib/Padre/
08:18 waxhead alias, I'm still seeing this..
08:18 waxhead Use of uninitialized value in numeric eq (==) at /home/pete/Programming/Perl/Padre/t​runk/Padre/lib/Padre/TaskManager.pm line 346.
08:18 waxhead have you fixed this?
08:18 Alias 346?
08:18 * Alias looks
08:18 Alias I'm not seeing it
08:18 Alias What are the lines around it
08:20 shadowpaste "waxhead" at 217.168.150.38 pasted "error for alias" (19 lines) at http://paste.scsys.co.uk/89435
08:20 waxhead alias ^^
08:21 Alias uh, which line is the one we care about :)
08:22 waxhead next unless $worker->{wid} == $handle->{worker};
08:23 waxhead hmm..  I'm not seeing the error, but I wonder if it's something to do with the SVN plugin... but I can't see why, it doesn't use the TaskManager
08:24 Alias It might be to do with the task manager overloading and restart
08:24 Hyppolit svn: r13825 | waxhead++ | http://padre.perlide.org/trac/changeset/13825
08:24 Hyppolit See if this does anything to cause the warning.
08:24 Hyppolit branches/Padre-Plugin-SVN-take2/lib/Padre/Plugin/
08:25 Hyppolit svn: r13826 | waxhead++ | http://padre.perlide.org/trac/changeset/13826
08:25 Hyppolit Nope.. the error doesn't show up...
08:25 Hyppolit
08:25 Hyppolit I wonder if it was due to the bug fix just applied to the taskmanager??
08:25 Hyppolit branches/Padre-Plugin-SVN-take2/lib/Padre/Plugin/
08:25 waxhead alias maybe...
08:25 waxhead I did what I did before to trigger the error and it didn't
08:25 waxhead so maybe that was it.
08:26 Alias I can't imagine $worker->{wid} would be undef
08:26 * waxhead goes off to make a new Padre tarall
08:26 Alias I think it might have been that error
08:26 Alias It occured after the thread spawned, but before the task handle was assigned to it
08:26 waxhead can't you just test it's defined though?
08:26 Alias It's one of those things where it's NEVER supposed to be undef
08:27 Alias Testing for definedness would be essentially saying that something that should never happen is allowed
08:27 Alias It's turning off a warning by suppressing the warning instead of finding the bug
08:27 waxhead yeah, I sort of thought that after I hit enter...
08:28 waxhead you know.. "maybe you want to get the error since it hsould be defined.."
08:28 Alias I'm moderately certain the bug fix will remove the warning
08:28 Alias If not, it should only happen if the task manager freaks out and does a hard restart
08:50 Sno|Laptop joined #padre
08:53 [Sno] left #padre
09:01 waxhead szabgab, http://padre.perlide.org/trac/wiki/Release
09:01 waxhead just updated it..
09:02 waxhead I need to document the merge process, but it's there as a start
09:12 mrg joined #padre
09:12 mrg hi
09:12 mrg is now known as zenog_
09:13 zenog_ waxhead: fyi, I have added a version called 0.82pre to Trac, so that we can manage tickets related to the release branch.
09:13 zenog_ I hope this okay.
09:17 szabgab zenog_: it is ok but why not call it 0.82 ?
09:19 static joined #padre
09:19 daxim joined #padre
09:22 zenog_ szabgab: Initially I wanted to call it 0.82, but one also needs to enter a release date and this cannot be modified afterwards ...
09:23 szabgab ah, I think we could use the date of the branching
09:23 zenog_ szabgab: Then we could differentiate between the things that happen in the release branch and the things that are reported after the release.
09:23 szabgab that's true
09:23 zenog_ szabgab: Sure, that would also be an option.
09:24 zenog_ basically, I do not really have a strong opinion on how to call it ;-)
09:24 Sewi what about a version "release" which covers everything between branching and releasing (and should be empty once the release is done)
09:24 szabgab then I'd prefer - for the next release - to call it as the version number
09:25 szabgab I am not sure we will want to wait with the relesae till all the bugs are fixed
09:25 szabgab we can find bugs that we say are not that critical
09:25 szabgab and we still found them in 0.82
09:25 szabgab though we can refile them to 0.83 as they are tere too I guess
09:26 Sewi We could move everything from "release" to either "trunk" or "0.82" once the release is done
09:28 szabgab let's see how that works this time
09:29 szabgab once waxhead releases this version we should all give our opinions on the new release process
09:29 szabgab to see how it could be improved
09:41 zenog_ yeah
09:42 zenog_ One thing we could do: Before the release, all tickets for 0.82(pre-or-whatever) must be closed - either fixed or wontfix, and we can re-open and re-assign the wontfix ones after the release (if it has not been fixed in trunk in the meantime)
09:46 waxhead I'm just updating my VM to install the Padre tar ball
09:46 waxhead if anyone else wants to try it out, http://peshel.dnsalias.net/pa​dre-release/Padre-0.82.tar.gz
09:47 waxhead I'm thinking I should rename that to RC or PreRelease or something...
09:47 waxhead anyway.. this is not the final Padre, it's only there if you can try it on a VM or don't care what it does to your production machine.. :)
09:53 daxim -TRIAL.tar.gz
09:54 daxim we have a perfectly cromulent infrastructure for that
09:59 zenog_ waxhead: What about #1138 - does anyone else besides me have those problems with the release branch?
09:59 Hyppolit #1138: Warnings at Padre startup; Padre server not working (new defect) [ http://padre.perlide.org/trac/ticket/1138 ]
10:49 jazzanova left #padre
10:50 szabgab zenog_: I think it is there but was already fixed on trunk, but I am not 100% sure in this
10:50 szabgab waxhead: please don't create two "releases" with the same file anme
10:51 szabgab we will never know which one is which
11:15 waxhead I haven't created any releases.. just the tar ball..
11:15 waxhead zenog_ have you svn update on the brnach?
11:15 waxhead I just started Padre on the VM and it's fine
11:15 waxhead might need to get some files on there to test more though
11:15 waxhead but no errors on start up
11:16 waxhead zenog_, I'm getting it now... had to turn on the single instance setting
11:19 waxhead that's interesting.. I'm getting a segfault when turning on the outline
11:26 waxhead OK.. I'm running the build of padre from the branch, the one that is on the web server currently..
11:27 waxhead with "view/outline" turned on and Padre::Wx::Main in the editor, I get a seg fault.
11:27 waxhead this is on a VM though...
11:27 waxhead ohhhh... let me check memory settings are high enough...
11:30 waxhead nope, even with 1 gig of ram, it still segfaults
11:55 szabgab waxhead: I think we should try to get fenderson to write up such use-cases that will show us various error cases
11:56 szabgab feel free to give him ideas what to do and how to describe
11:59 fenderson yea i'll be happy to do it
12:01 Hyppolit svn: r13827 | mmaslano++ | http://padre.perlide.org/trac/changeset/13827
12:01 Hyppolit Updated Czech translation a little.
12:01 Hyppolit branches/release-0.82/share/locale/
12:06 aiglenoir left #padre
12:06 aiglenoir joined #padre
12:11 aiglenoir left #padre
12:11 aiglenoir joined #padre
12:29 pece left #padre
12:32 pece joined #padre
12:47 aiglenoir left #padre
12:48 aiglenoir joined #padre
13:09 asarch joined #padre
13:20 user_6347 left #padre
13:28 kanishka left #padre
13:46 zenog_ The Linux standalone seems to be missing Wx support, could that be? Or some development bindings? Please have a look at the mailing list ...
14:03 aiglenoir left #padre
14:03 aiglenoir joined #padre
14:05 aiglenoir left #padre
14:08 aiglenoir joined #padre
14:09 aiglenoir left #padre
14:13 aiglenoir joined #padre
14:14 jnap joined #padre
14:14 aiglenoir left #padre
14:18 aiglenoir joined #padre
14:18 asarch left #padre
14:19 aiglenoir left #padre
14:23 aiglenoir joined #padre
14:24 zenog_ left #padre
14:43 zenog re
14:45 asarch joined #padre
15:09 kaare left #padre
15:28 alnewkirk left #padre
15:28 alnewkirk joined #padre
15:31 pece left #padre
15:33 static_ joined #padre
15:35 static left #padre
15:36 static_ is now known as static
15:49 dapatrick joined #padre
15:57 dapatrick left #padre
16:06 Alias__ joined #padre
16:07 user_7609 joined #padre
16:08 user_7609 left #padre
16:25 danlucraft left #padre
16:25 danlucraft joined #padre
16:35 static left #padre
16:45 dapatrick joined #padre
16:45 Sewi left #padre
17:12 pece joined #padre
17:15 asarch left #padre
17:17 asarch joined #padre
17:21 marcela left #padre
17:28 danlucraft1 joined #padre
17:28 danlucraft left #padre
17:34 danlucraft1 left #padre
17:48 Jagdwurst joined #padre
18:01 dodathome joined #padre
18:18 toi joined #padre
18:23 ispy_ left #padre
18:23 ispy_ joined #padre
18:39 daxim left #padre
18:54 pece left #padre
19:01 kaare joined #padre
19:05 fenderson hi
19:17 kanishka joined #padre
19:35 asarch left #padre
20:14 asarch joined #padre
20:14 asarch left #padre
20:29 waxhead regarding the crashing last night, I think it could be the version of perl, I'm running the previous version of ubuntu on it.. I'll do up a perlbrew tonight to double check
20:38 dapatrick left #padre
20:39 dapatrick joined #padre
20:52 Sewi joined #padre
21:04 Jagdwurst left #padre
21:17 Steffen joined #padre
21:21 Sewi left #padre
21:22 Sewi joined #padre
21:30 dodathome left #padre
21:36 Hyppolit svn: r13828 | gyu++ | http://padre.perlide.org/trac/changeset/13828
21:36 Hyppolit not complete, but close: 86%
21:36 Hyppolit branches/release-0.82/share/locale/
21:46 Sewi left #padre
22:01 toi left #padre
22:02 Hyppolit svn: r13829 | zipf++ | http://padre.perlide.org/trac/changeset/13829
22:02 Hyppolit Updated Spanish translation
22:02 Hyppolit branches/release-0.82/share/locale/
22:12 chorny joined #padre
22:37 kanishka left #padre
22:53 kaare left #padre
22:55 dapatrick left #padre
22:56 Steffen left #padre
23:13 Di-ima left #padre
23:22 jnap left #padre
23:24 Di-ima joined #padre
23:44 submersible left #padre
23:48 Alias_ Is it just me or does double click on the function list not work any more

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