Camelia, the Perl 6 bug

IRC log for #padre, 2010-03-11

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

All times shown according to UTC.

Time Nick Message
00:17 teejay joined #padre
01:01 jazzanova joined #padre
02:11 dorkfish joined #padre
02:44 patspam joined #padre
03:01 patspam joined #padre
03:22 Alias__ joined #padre
03:28 patspam joined #padre
03:34 kyanardag joined #padre
03:47 Alias__ Whoever implemented the survey is doing something too aggressively
03:47 Alias__ It's popping up a shitload of times during the test suite
03:53 szabgab joined #padre
04:20 Hyppolit svn: r11054 | adamk++ | http://padre.perlide.org/trac/changeset/11054
04:20 Hyppolit Moved the survey into the post-init timer, skip it during testing, and load the class for it on the fly.
04:20 Hyppolit trunk/Padre/lib/Padre/Wx/
04:21 Hyppolit svn: r11055 | adamk++ | http://padre.perlide.org/trac/changeset/11055
04:21 Hyppolit Because it's Wx that has the dependency on threads, load threads in the Wx module, rather than in some third module
04:21 Hyppolit trunk/Padre/lib/Padre/
04:21 Hyppolit svn: r11056 | adamk++ | http://padre.perlide.org/trac/changeset/11056
04:21 Hyppolit Add explicit dependencies for just barely enough of Wx to get the slave driver spawned (enough to make a couple of shared event objects)
04:21 Hyppolit trunk/Padre/lib/Padre/
04:21 Hyppolit svn: r11057 | adamk++ | http://padre.perlide.org/trac/changeset/11057
04:21 Hyppolit Add threads.pm to the startup settings
04:21 Hyppolit trunk/Padre/lib/Padre/
04:24 Hyppolit svn: r11058 | adamk++ | http://padre.perlide.org/trac/changeset/11058
04:24 Hyppolit Completed the move of the slave driver into the early part of startup
04:24 Hyppolit trunk/Padre/ trunk/Padre/lib/Padre/
04:25 GabrielVieira joined #padre
04:32 Alias__ Oh crap
04:35 GabrielVieira could anyone help me installing a module that looks for a lib that is installed in another path?! :)
04:35 Hyppolit svn: r11059 | adamk++ | http://padre.perlide.org/trac/changeset/11059
04:35 Hyppolit The SlaveDriver optimisation turns out to be utterly fatal to Padre for the moment, because parts of the task manager expect a cloned copy of the entire Padre application.
04:35 Hyppolit
04:35 Hyppolit Temporarily disabling the optimisation
04:35 Hyppolit trunk/Padre/lib/Padre/
05:52 danlucraft joined #padre
05:56 Sno|Laptop moin moin
05:56 Sno|Laptop Alias__: what on earth did you do :)
05:57 Alias_ Well, the slave driver concept is based on having very little in the way of modules loaded when you spawn the first thread
05:57 Alias_ And then copying THAT to keep all other threads small
05:57 Alias_ While the main thread grows to be large
05:58 Alias_ So Main(big), Master(small), Slave1(small+task-deps), Slave2...
05:58 Alias_ But at the moment, the master is spawned almost as late as the original version
05:58 Sno|Laptop sounds reasonable
05:58 Alias_ So the master is 25-40 meg
05:58 Alias_ The plan was to get it working, then move the master spawn time forwards separately
05:59 Alias_ Gradually recovering memory as we did so
05:59 Alias_ So I tried to put it forwards into the startup
05:59 Alias_ Which saved 20meg of ram
05:59 Alias_ ... and then sent Padre into a death spiral because code running in those background threads didn't have the code loaded they expected
05:59 Alias_ Padre survives for about a minute :)
06:00 Sno|Laptop whereby - in times where people have 4GB+ of RAM, don't waste to much time saving 20mb
06:00 Sno|Laptop hey - Sewi told me yesterday, Padre couldn't exit, so you made progress :)
06:00 Alias_ Sno| bad logic
06:01 Alias_ Sno: 1. Hardcore developers may have 4gig, but not everyone 2. People might want to use that 4GB for something ELSE other than Padre
06:01 Sno|Laptop not really - if it's easy to save 20mb ram, do it
06:01 Alias_ 80meg -> 60meg isn't a small change
06:01 Alias_ Also, these savings are 5, 10 or 15 meg PER THREAD
06:01 Sno|Laptop if you save 20mb per thread, do it
06:01 Alias_ And right now, we don't use many threads
06:02 Alias_ We want to use shitloads
06:02 Alias_ Every meg you save is a meg you can spend on actual features
06:02 Sno|Laptop the point is: ram is there to be used for programs/applications - not for file buffer
06:03 Sno|Laptop and you fail - my workstation has 16GB RAM ;)
06:03 Sno|Laptop only my laptop must "live" with 4 :)
06:04 Sno|Laptop having a look at (I like comparing Padre to) Eclipse+EPIC - Eclipse is wasting more RAM
06:05 Sno|Laptop running eclipse takes around 1-1,5GB, running Padre around 0,5GB (process memory, 64-bit arch)
06:18 Alias_ Running Padre for me takes about 100-150meg
06:18 Alias_ Problem is we're not doing anything yet
06:22 Sno|Laptop on windows/32, right?
06:29 Alias_ Ya, for me it's Win32
06:29 Alias_ Half a gig seems surprising
06:30 Sno|Laptop NetBSD/amd64 :)
06:30 Sno|Laptop but it's the same on Linux/x86_64 - IIRC
06:31 Sno|Laptop whereby I talk about the entire memory, the perl process running Padre consumes
06:31 Alias_ Me too
06:33 Sno|Laptop k, half a gig was to big ...
06:33 Sno|Laptop 24390 trevor    85    0   326M  134M select/0   0:04  0.00%  0.00% perl
06:33 Alias_ We have done various work to try and cull it down
06:34 Alias_ Removed deps, etc
06:34 Alias_ So it fluctuates, obviously
06:35 Sno|Laptop from my point of view it's ok - price/effort looks reasonable
06:35 Alias_ There's a great bit in the online videos of MIT's algorithms course
06:36 Alias_ In the first lecture, the head of the school is explaining why it's important
06:36 Alias_ His argument is that performance is like currency
06:36 Alias_ Even if you don't need it now, it's important to spend less so that your budget for future improvements is larger
06:36 Alias_ And you aren't locked into that increased spending unexpectedly
06:37 Sno|Laptop yes - for sure - but if you reach 80% of max performance, the other 20% will cause cost explaining in reach ...
06:37 Alias_ You only know where those lines lie from the inside though
06:38 Sno|Laptop I'm not english native speaker, so I might not explain correctly what I mean
06:38 Alias_ you do, I think
06:38 Alias_ "Diminishing returns"
06:38 Sno|Laptop have a look at the prices for High-Availability - cold stand by doubles costs, but you may have downtimes up to 1h per takeover
06:39 * Alias_ nods
06:39 Sno|Laptop 99,999% gives you downtimes of 5min/year - but prices are "jenseits von gut und b�se"
06:39 Alias_ Our cost for a downtime over an hour as about $50,000, so we do a lot of redundancy stuff
06:40 Sewi joined #padre
06:40 Sewi gm
06:40 Sno|Laptop that's why I thought (with my 16GB RAM and padre consuming 350MB): wtf he want's to save 20MB?
06:40 Sno|Laptop moin Sewi
06:40 Sno|Laptop Alias had made padre exiting again - so you can release :)
06:40 Alias_ Sno: I have a 4-core desktop, and I'd like to be running 6-8 background jobs to fill them all
06:41 Sno|Laptop so you need more RAM, Alias_ :)
06:41 Alias_ Reverse the logic :)
06:41 Sno|Laptop heh
06:41 Alias_ "You MUST have $this much RAM to use Padre"
06:41 Alias_ Screw you developing world!
06:41 Sno|Laptop I understand :)
06:42 Alias_ Anyways, I want a ton of threads
06:42 szabgab Alias_, that's the killer statement
06:42 Alias_ And at 40meg of ram per thread before any actual work is done, that's pricey
06:42 szabgab and gm everyone, I see you are enjoying yourself :-)
06:42 Alias_ :)
06:42 Sewi Sno|Laptop: I think, I did a workaround for this exit bug, don't know if it's a fix :-)
06:42 Alias_ I must be getting old
06:42 Alias_ I've spent the day wondering if people disagree with me just to be difficult
06:43 Sewi Alias: Is r11050 a fix?
06:43 Hyppolit Changeset #11050 http://padre.perlide.org/trac/changeset/11050
06:43 Sno|Laptop is there any good tool for perl to see which (perl) data structures consumes how much ram?
06:43 Alias_ Look through Devel
06:43 Alias_ There's a few
06:43 Sno|Laptop Devel is big o.O
06:44 Alias_ Devel::Size and related
06:44 Alias_ There's half a dozen different mechanism people have used
06:44 Alias_ Sewi: I don't know threading well enough in practice
06:45 Sno|Laptop so there's no recommendation by perl itself?
06:45 Sewi I'll try to ask Steffen / tsee, I think...
06:46 Alias_ Sno: It depends a lot on what it is you want to measure
06:47 Alias_ But no, not really
06:47 Sno|Laptop Alias_: trivial: want to have interactive flowchart about memory consumption of padre ;)
06:48 Sno|Laptop no, just kidding - I just want to have sth. to analyse ram usage (e.g. some people are whining, S::S consumes to much ram etc.)
06:48 Sewi Sno|Laptop: There is a Devel module used by --fulltrace. If you add a memory check after every line, you could get the delta
06:49 Sno|Laptop must be done in perl-debugger - removing the memory he debugger consumes ...
06:50 Alias_ S:S?
06:51 Sno|Laptop SQL::Statement
06:52 Sno|Laptop Sewi: your patch - I would recommend using tv_interval and threads::yield()
06:53 pece joined #padre
06:53 Sewi Sno|Laptop: Use it :-)
06:53 Sewi I was just happy to remove that endless loop
06:54 Sno|Laptop as I said, I have enough to develop/fix - I just thought you want a review and comments
06:55 waxhead joined #padre
06:56 marcela joined #padre
06:59 waxhead hey everyone...
06:59 Sewi Hi waxhead
06:59 Sewi guys, _any_ reply on the spy function would be nice :-)
07:00 Sewi waxhead: Did you get any feedback about the unstable releases? I 'd like to see the next release stable again (also meaning: Do the next release not before we're stable again)...
07:01 Sno|Laptop but Sewi: if you think sth. could be improved by implementing it in XS, I might be able to help you :)
07:01 waxhead Sewi, I don't ever see any feedback
07:01 kaare joined #padre
07:01 waxhead yes it would be nice to get back to stable, but it can only happen once someone gets it stable...
07:02 Sewi We're working on it :-)
07:02 waxhead I see Alias is back on the case for that... and broke trunk for a bit ..
07:02 Sewi Sno|Laptop: Sorry, "sth."?
07:04 waxhead Sewi, so once trunk is stable we'll release a stable version... :)
07:05 Alias_ We're stable again, I'm fairly sure
07:05 * Sewi running make test
07:05 Alias_ I checked in the SQLite code, and it does flush the driver objects in CLONE
07:05 Alias_ So SQLite should be completely safe now
07:05 Sewi But we still crash very badly on locked DB.
07:05 Sno|Laptop sth. = something (isn't it the official abbreviation?)
07:05 Alias_ Statement Handle
07:05 waxhead wow.. some decent traffic in the dev list
07:05 Alias_ "locked"?
07:06 Alias_ Sewi: This is the multiple instance case?
07:06 Sno|Laptop I thought we're talking english here - about perl, and not perl about english ;)
07:06 Sewi I think so. I saw most cases when a looping trying-to-end-myself Padre process was hanging around
07:06 Alias_ hrm
07:06 Alias_ I think that's a connection option thing
07:07 waxhead Alias_, serious ?  We're stable?
07:07 Sewi Sno|Laptop->send(Wx::gettext('Maybe'));
07:07 Alias_ waxhead: It was only one specific bug
07:07 waxhead so even the threading stuff is sorted?
07:07 Alias_ waxhead: And I knew how to fix it, just had to find a block of time
07:07 Sewi I submitted a workaround but don't know if it's a fix.
07:07 waxhead Alias_, what about this afternoons breakage?  ( I watch the log from work )
07:07 Alias_ It was an optimisation attempt
07:07 Alias_ I've disabled it for now
07:08 Alias_ We should have been stable since this morning
07:08 * Sewi hands over Alias a block of time: http://ecx.images-amazon.co​m/images/I/41yzkRyKVyL.jpg :-)
07:08 waxhead Alias_, so the thread db handle thing is sorted out.. i think that was really the major problems
07:08 waxhead problem...
07:09 waxhead well, I'm happy to roll out a stable release if people think it's right to advertise it as such...
07:10 waxhead Has azawawi finished  the advanced or key bindings stuff yet...
07:10 Alias_ Might be worth waiting for the keybindings to actually be savable somewhere :)
07:10 Alias_ Would be nice if that actually worked
07:10 Sewi We should wait if we get any major bug reports at least over the weekend and I won't feel well if we get the spy function into the release without any comment (and withouot the server being running)
07:10 Alias_ spy function?
07:10 waxhead I have to head out for a while now.. will catch up on things when I'm back..
07:11 Alias_ Sewi: BTW, I don't like the survey
07:11 Alias_ Sewi: I'd prefer not to bug people on first startup
07:11 Alias_ Instead to run the "first time" stuff like the survey after maybe 10 hours of use
07:12 Sewi Sure, also a good idea.
07:12 Alias_ So people that are just looking at it and will never run it again are excluded
07:12 waxhead OK.. no rush on the release.. so we'll go for early next week.. I'd like to see some time spent making sure things are working OK... there's been a few things mentioned in irc lately about Padre not closing etc..
07:12 Sewi make test passes here (without author tests)
07:12 Alias_ waxhead: If that's in 0.56 it would be worth releasing
07:13 waxhead alias_, I thought it was trunk...
07:13 waxhead ok gone
07:17 Alias_ If it's trunk, no biggie
07:17 Alias_ No need for release
07:17 Alias_ We Warned Them
07:17 pece joined #padre
07:18 Hyppolit svn: r11060 | Sewi++ | http://padre.perlide.org/trac/changeset/11060
07:18 Hyppolit Remove badcode test failures
07:18 Hyppolit trunk/Padre/lib/Padre/
07:35 pece joined #padre
07:36 alvar joined #padre
07:57 Hyppolit svn: r11061 | Sewi++ | http://padre.perlide.org/trac/changeset/11061
07:57 Hyppolit Some other test failures
07:57 Hyppolit trunk/Padre/lib/Padre/Document/ trunk/Padre/lib/Padre/Task/
08:09 Sno|Laptop joined #padre
08:13 joerg joined #padre
08:14 joerg gm
08:14 Sewi Hi
08:25 mrg joined #padre
08:25 mrg hi everyone
08:26 zenog When I start the installed version of Padre (not via dev.pl), then it crashes on startup:
08:27 zenog Can't locate Padre/Wx/Dialog/WhereFrom.pm in @INC (@INC contains: /home/mrg/.padre/plugins /home/mrg/software/zeno/perl-modules /etc/perl /usr/local/lib/perl/5.10.0 /usr/local/share/perl/5.10.0 /usr/lib/perl5 /usr/share/perl5 /usr/lib/perl/5.10 /usr/share/perl/5.10 /usr/local/lib/site_perl .) at /usr/local/share/perl/5.10.0/Padre/Wx/Main.pm line 338.
08:27 Sewi Seems you're mixing the trunk and installed version...
08:28 Sewi WhereFrom.pm is used since yesterday night.
08:28 zenog hm.
08:28 zenog I'll delete everything and do a reinstall ...
08:28 Sewi make clean
08:39 daxim joined #padre
08:39 Sewi Hi daxim
08:40 daxim 'lo
08:41 danlucraft joined #padre
08:49 zenog Sewi: works now
08:49 pece joined #padre
08:51 Sno|Laptop Sewi: http://www.hirschbeutel.de/padre/​patch-lib::Padre::Taskmanager.pm
08:53 Alias__ grr
08:53 Alias__ I swear somewhere in this damned startup code is a way to get the thread cost down to something normal
08:55 Sewi Sno|Laptop: The sleep made threads finish themselfs in most cases, this is why I added it. A select undef,undef,undef didn't work.
08:56 Sno|Laptop yield() is no select(undef,undef,undef), isn't it?
08:57 Sewi No, but you removed that sleep...
08:57 Sno|Laptop sure - it's enough
08:58 Sno|Laptop sleep does: while( !waited around the required nanosec ) { yield() }
08:58 Sewi I think sleep is a kernel call
08:58 Sno|Laptop yes
08:58 Sno|Laptop and the kernel does exactly what I wote (a bit more complicated, but in principle that way)
09:00 Sewi patch it in and I'll try it out :-)
09:00 Sewi I doesn't care about replacing my code as long as it's an improvement :-)
09:00 Sno|Laptop I'm here since 16:00 CET
09:00 Sno|Laptop s/since/until/
09:17 jagd joined #padre
09:23 joerg ping Sewi
09:24 Sewi Sno|Laptop: go ahead
09:24 Sno|Laptop ?
09:25 Sno|Laptop I don't remember having commit bit
09:25 Sewi patching
09:27 Sewi Which is your trac username?
09:28 Sno|Laptop rehsack - like my cpan username
09:28 Sno|Laptop but you can commit it
09:29 Sno|Laptop or do you prefer have my name there for svn blame? :)
09:29 Hyppolit svn: r11062 | Sewi++ | http://padre.perlide.org/trac/changeset/11062
09:29 Hyppolit bless $rehsack,Padre::SVN::CommitBit;
09:29 Hyppolit subversion/
09:30 Sno|Laptop now I must guess my password :)
09:30 Sno|Laptop once set, saved in firefox and forgotten
09:31 Sewi You could get a new one, I guess
09:31 Sno|Laptop I'll try a few standard passwords first ...
09:31 Sno|Laptop but first: new coffee
09:35 Sno|Laptop Sewi: is the commit bit for Padre repository only, or for all repositories?
09:35 Sewi svn.perlide.org
09:35 Sno|Laptop Fayland asks me for patching my RT-submits directly
09:35 Sno|Laptop ok
09:55 Alias__ The whole repo is under one permission set
09:55 Alias__ You are encouraged to fix stuff anywhere you like
09:57 Sno|Laptop You'll never get a stable release then :)
10:02 Sewi Commit bits aren't constants :-)
10:15 zenog Does "English (New Britstralian)" mean "English (UK/New Zealand/Australia)"?
10:15 zenog Shouldn't it be "English (New Britstralia)" then? :-)
10:16 zenog btw, the View -> Language option does not allow the locale "C" and "English (US)"
10:18 Sewi zenog: Ask Alias, he does the English translation :-))
10:18 Alias__ Good point!
10:19 Alias__ We use (Country) now, don't we
10:20 zenog Alias__: I guess, at least that is what's done most of the time in software i18n.
10:20 Alias__ Sure thing
10:29 Sewi Do we want|need (up to) three different English versions?
10:31 zenog Sewi: I guess if there are US users who want to have "color" instead of "colour", then they can step up and do their thing.
10:32 zenog Sewi: Same goes for German-speaking Swiss people who may want to have an interface w/o "ß", and Austrians who want "Jänner" instead of "Januar".
10:34 daxim burn it with fire!
10:37 Sewi daxim: You're always so understanding and liberal about people's questions :-)
10:37 Sewi zenog: You're right, but let's add it the day some actually does this translation.
10:37 Sewi Otherwise I'd add "Klingon" as a language, too :-)
10:43 Alias__ hrm
10:44 Alias__ Just tried my second attempt at making SlaveDriver operation work properly
10:44 Alias__ Segfault within a tenth of a second :(
10:47 zenog Sewi: That's what I meant. If someone wants do it, then it's okay.
10:50 Sewi Sure, but don't add these languages just for extending the list without any changes to AU English
10:51 zenog Sewi: no no, don't worry.
10:51 zenog What does "Select to matching brace" do?
10:52 Sewi Go to a opening/closing bracket and use it. It creates a selection up to the matching bracket.
10:55 zenog Sewi: cool, thanks
10:55 Sewi I requested, but didn't do that :-)
11:08 Hyppolit svn: r11063 | rehsack++ | http://padre.perlide.org/trac/changeset/11063
11:08 Hyppolit Update thread-finish code to wait real for 2 minutes and don't guess
11:08 Hyppolit how long a sleep(2) really takes.
11:08 Hyppolit trunk/Padre/lib/Padre/
11:08 Alias__ zenog: It's been a point of honour that we refuse to support American
11:08 Sno|Laptop Sewi: first try pw-guess succeeds :)
11:08 Alias__ zenog: Because every other application in the universe uses American english as the default
11:08 Alias__ zenog: And we have no American developers
11:08 Sewi Sure, we got Ryan
11:09 Alias__ And BTW, our core English is actually British
11:09 Alias__ English English
11:09 Sno|Laptop but it doesn't tell how much pw-misses I had :)
11:09 Alias__ It just so happens there are no words that are different between British, Australian and New Zealand in the words we use in this application
11:10 Alias__ And only in the English translation do we use the "New Britstralian" joke
11:10 Alias__ All the other translations use "British"
11:10 Alias__ Or localised translations of the word "British"
11:11 Alias__ I think in the American locale, you see British too :)
11:11 waxhead re
11:11 Getty Sewi: if you really use klingon, there are books and stuff about it, its really defined, so dont take it to funny ;)
11:13 Sewi I know, a friend of mine spoke it and Google has a translation, too.
11:16 waxhead Sewi, what bad code test fails did you fix?
11:17 waxhead I thnk there's a patch for the badcode test to make it work under some environment variable.. rehbus did it, I tested it out and it seemed to work fine...
11:17 Sewi waxhead: I didn't touch the test, just fixed a DB::single in a Padre mod
11:17 waxhead but when the tests ran they found a lot of "beginner" errors, which looked more like errors with the way the beginner error thing worked
11:20 Haarg joined #padre
11:21 Alias__ OK guys, I need advice
11:21 Alias__ I've got a working implementation of the slave driver stuff as it was meant to run
11:21 Alias__ i.e. Spawning as early as possible
11:22 Alias__ The change in memory base load (with master thread + one worker) is a reduction from 82.7meg to 56.3meg
11:22 Alias__ But it crashes on exit
11:22 Alias__ Should I commit?
11:22 Sewi Fix the crash first :-)
11:22 Alias__ I'm not sure how
11:23 Alias__ Cause I don't know threads
11:23 Sewi me too
11:23 Alias__ I just know the startup sequence really really well :)
11:23 Alias__ Because I wrote most of it
11:23 Sewi commit it, we may revert it as needed.
11:23 Alias__ It can be disabled with a one-line change
11:23 Alias__ I'll put a note at the location
11:23 Sewi ok
11:34 waxhead seems reasonable..
11:34 waxhead maybe ping steffen to see if he can take a look at the threading stuff
11:34 waxhead I know he reads the dev list, why not post something there about it?
11:35 Hyppolit svn: r11064 | adamk++ | http://padre.perlide.org/trac/changeset/11064
11:35 Hyppolit Landing the slave master quick-spawn change, which reduces average memory cost of Padre by 25%
11:35 Hyppolit trunk/Padre/ trunk/Padre/lib/ trunk/Padre/lib/Padre/ trunk/Padre/lib/Padre/Wx/
11:35 Alias__ ok
11:42 waxhead where's the "but breaks padre when closing" at the end of the commit?  :)
11:47 Alias__ :)
11:47 Alias__ shhhh
11:50 Sewi It's there!? First line between the two | chars... If there is either adamk or sewi, it means: "Something is broken now" :-)
11:51 Alias__ heh
11:51 Alias__ Actually
11:51 Alias__ For me, the rule would be s/Landing//
11:51 Alias__ er
11:51 Alias__ m/landing/i
11:52 Alias__ I suspect if we looked back in history, there would be comments like "Landing Config 2.0" :)
11:52 Sewi :-)
11:53 Alias__ OK, my work is done for today
11:53 * Alias__ goes to shoot some Russians
11:53 Sno|Laptop have fun
11:53 waxhead I prefer being a ninja myself. :)
11:54 waxhead ok.. sleep time.
11:54 waxhead nite
11:54 * Sno|Laptop prefers shooting users on cpan :)
11:54 * Sewi hides
12:09 Sewi Is there any possibility to read a text key out of the language file without using gettext?
12:10 cognominal joined #padre
12:15 zenog Sewi: Locale::PO
12:17 Sewi thanks
12:18 zenog "Jump to the matching opening or closing brace: {, }, (, )" - this also works for < > and [ ] - is it okay if I change the message there?
12:19 Sewi sure
12:19 zenog Cool. Maybe I ask too much here, I should just do it ...
12:19 Sewi Just do it as long as it's any kind of improvement. If unsure, ask :-)
12:25 ecocode joined #padre
12:29 Hyppolit svn: r11065 | Sewi++ | http://padre.perlide.org/trac/changeset/11065
12:29 Hyppolit Add untranslatable language codes/names to Wx/Menu/View.pm
12:29 Hyppolit DEVELOPERS: PLEASE ADD YOUR LANGUAGE AS YOU WRITE IT LOCALLY TO THE TOP OF THIS FILE!
12:29 Hyppolit trunk/Padre/ trunk/Padre/lib/Padre/ trunk/Padre/lib/Padre/Wx/Menu/
12:30 Sewi Argh, meant TRANSLATORS
12:34 alvar joined #padre
12:41 daxim ridiculous
12:47 Hyppolit svn: r11066 | zenogantner++ | http://padre.perlide.org/trac/changeset/11066
12:47 Hyppolit German translation updates
12:47 Hyppolit fr_CA is French (Canada), not French (France)
12:47 Hyppolit comments about edit command involving parantheses: also mention < > and [ ]
12:47 Hyppolit gettext-ified reload dialog
12:47 Hyppolit trunk/Padre/lib/Padre/ trunk/Padre/lib/Padre/Action/ trunk/Padre/lib/Padre/Wx/ trunk/Padre/lib/Padre/Wx/Dialog/ trunk/Padre/share/locale/
12:48 zenog So, lunch break is over ... back to work.
12:52 Hyppolit svn: r11067 | daxim++ | http://padre.perlide.org/trac/changeset/11067
12:52 Hyppolit let's do i18n correctly and without duplicating work that's been already done
12:52 Hyppolit trunk/Padre/lib/Padre/Wx/Menu/
13:12 sjn joined #padre
13:12 sjn hey folks
13:12 zenog hi sjn
13:13 sjn quick q: is there a setting in Padre that enables the shell/emacs-type cut/paste keys? (C-k/C-y)
13:17 daxim it's not configurable. bug #411
13:17 Hyppolit #411: configurable key bindings (new enhancement) [ http://padre.perlide.org/trac/ticket/411#comment:3 ]
13:17 sjn ah well
13:18 zenog sjn: That's work in progressSee also bug #120
13:18 Hyppolit #120: Emacs Keybindings (new enhancement) [ http://padre.perlide.org/trac/ticket/120#comment:1 ]
13:20 sjn ok, sweet
13:20 zenog sjn: Sewi offers his help in dealing with the plugin in the bug report. He is quite active here in #padre
13:22 Getty damn, if all of the perl on a system is to "old" you cant really install local::lib
13:23 Getty mh, now i'm stocked...
13:23 Sno|Laptop install new perl to another place on the system, eg to /usr/pkg/bin/perl :)
13:25 Getty yeah actually really good idea DIRECTLY make own perl onto that
13:25 Getty i got a compiler "wow" ;) so this could go
13:26 Getty i got a managed hosting facility for lots of sites and cause i want to "go perl" now with the business sites, too, i just wanted to precheck if that plan works at all
13:26 Getty but yes with own perl anything goes, i forgot....
13:30 Sewi sjn look at the emacs plugin
13:31 Sewi Getty: apt-get update;apt-get upgrade ;-)
13:31 rindolf joined #padre
13:35 Getty if i would have root..........
13:38 Sno|Laptop can't you use any built-in local root exploit? linux must have many of them ...
13:38 Sewi lol
13:39 Sno|Laptop I should reduce reading bofh and gbo ...
13:50 sjn Sewi: which one is the emacs plugin, btw? can't find an obvious one and the ::CommandLine one does something else than what I'm after
13:58 daxim test Padre::Plugin::EmacsMode
13:59 sjn can't find that one on cepan
13:59 sjn cpan*
14:03 sjn anyhoo, I have some minor suggestion, if anyone wants them :)
14:04 sjn - in Prefrences->Appearance, let default right margin be 76 instead of 80 (avoids wrapping when sending text in some mail clients)
14:07 sjn - in Preferences->Behaviour, the "Clean up file content on saving" feature, although useful, should have a default cleanup setup that allowes single empty lines (it's good to have some "space" in the code)
14:08 sjn one or two, even
14:09 sjn right now, they're all removed, which makes the code less readable (it removes any sectioning of the code text one has done using empty lines)
14:09 sjn it also fucks up the POD
14:20 Sewi I think one empty line is good and should be accepted, only multiple empty lines usually don't make sense.
14:21 sjn Sewi: you about that
14:21 Steffen joined #padre
14:21 sjn lots of people who have their own opinion about this
14:22 sjn Sewi: one thing that's easy to agree on at least, is that removing _all_ empty lines is silly
14:23 Sewi Sure, same with spaces at line endings.
14:25 sjn yes
14:32 Steffen Hmm. Spaces at line endings must die.
14:33 Sewi Hi Steffen.
14:33 Sewi Did you read what Alias wrote earlier?
14:33 Steffen Not really.
14:35 Steffen Sewi, well, what did he say, then?
14:35 Sewi He reduced the mem needs of Padre but added another exit-crash.
14:35 Sewi Do you have time to look at it?
14:35 Steffen Ah. Curious I missed that.
14:35 Steffen No, absolutely not.
14:35 Steffen Sorry.
14:35 Steffen Next week is conference week.
14:53 jagd joined #padre
15:11 Hyppolit svn: r11068 | adamk++ | http://padre.perlide.org/trac/changeset/11068
15:11 Hyppolit Now we have a Tools menu, move the regex editor there
15:11 Hyppolit trunk/Padre/lib/Padre/Wx/Menu/
15:21 Hyppolit svn: r11069 | adamk++ | http://padre.perlide.org/trac/changeset/11069
15:21 Hyppolit The next step in the renaming of Plugins to Tools
15:21 Hyppolit trunk/Padre/lib/Padre/ trunk/Padre/lib/Padre/Action/ trunk/Padre/lib/Padre/Wx/ trunk/Padre/lib/Padre/Wx/Menu/ trunk/Padre/xt/
15:23 Alias joined #padre
15:26 zenog Alias: Next step would be to move the preferences from "Tools" to "Edit" ;-)
15:27 Alias Heh, we just moved it FROM Edit :)
15:28 Alias Actually, the next big menu change is to get rid of the "Perl" menu
15:30 zenog hehe
15:30 Alias No really :)
15:46 dapatrick joined #padre
15:52 Sno|Laptop we had a big discussion about that, but Alias as Windows(tm) user followed MS gui style
16:03 Sewi Why did you move the RegEx editor?
16:04 Sewi People where just used to find it.
16:04 kyanardag joined #padre
16:04 Sewi Alias: Please try to do menu changes before the release and not month after something was published :-(
16:13 zenog Sewi: How often are Padre releases?
16:14 Sewi Evers 1 to 30 days :-)
16:14 Sewi Usually every 2-3 weeks
16:16 jagd 2-3 weeks are for testing version or for the stable one?
16:16 Sewi stable - as long as we are stable
16:22 * sjn install 0.58
16:22 sjn installs*
16:22 jagd :) it is amazing,   people, who are hacking Padre only for fun, can create stable releases so oft
16:22 Sewi Blame it on waxhead- we won't have so much releases without him :-)
16:24 ecocode hmm... stable ?
16:25 ecocode depends on what you call stable
16:25 Sewi 0.57 and 0.58 are not stable.
16:25 ecocode yep, seen that ;)
16:27 Sno|Laptop hmm, I didn't encounter big problems with Padre-0.58 on NetBSD/amd64
16:27 zenog Should the test suite work fine? It does not for me. "make test" gives errors.
16:28 Sno|Laptop works for me (tm) :)
17:02 Sewi Anybody using the current trunk (r11064) having crashes on exit?
17:02 Hyppolit Changeset #11064 http://padre.perlide.org/trac/changeset/11064
17:03 jagd Sewi: i
17:04 Sewi Which OS?
17:04 jagd Linux
17:04 jagd i tried again
17:04 Sewi I tried with and without open files and got no crash
17:04 jagd wait moment
17:08 jagd panic: free from wrong pool at /home/wu/projects/padre/padre/Padre/lib/Padre.pm line 242.
17:09 jagd Revision 11069.
17:12 * Sewi updating again...
17:14 pece joined #padre
17:16 Sewi No crash on --die but 1 unjoined and 1 running thread without
17:18 jagd genau.   sorry
17:18 Sewi np
17:25 Hyppolit svn: r11070 | Sewi++ | http://padre.perlide.org/trac/changeset/11070
17:25 Hyppolit Maybe fix for the exit-crash?
17:25 Hyppolit trunk/Padre/lib/Padre/
17:25 Sewi could you svn up and test this, please?
17:26 GabrielVieira_ joined #padre
17:39 CromeDome joined #padre
17:39 jagd Sewi:   with pleasure
17:39 Sewi thanks
17:39 jagd Sewi:  keine Meldung mehr
17:40 Sewi cool
17:41 cognominal joined #padre
17:44 jagd aber falls ich etwas anders machen,  wie z.B. Debuging,  mit RegexEditor umspielen.  und dann Padre zu schlie�en,  es gibt noch Fehlermeldung aus:
17:44 jagd panic: free from wrong pool at /home/wu/projects/padre/padre/​Padre/lib/Padre/SlaveDriver.pm line 246
17:45 jagd END failed--call queue aborted at /usr/lib/perl5/site_perl/5.10.0​/i586-linux-thread-multi/Wx.pm line 246.
17:45 jagd panic: free from wrong pool at /home/wu/projects/padre/padre/Padre/lib/Padre.pm line 242.
17:45 Sewi thanks
17:47 Sewi Could you check what exactly causes the errors?
17:47 Sewi Which action?
17:47 Sewi RegEx editor seems to be ok
17:47 jagd entschuldigung,    das Programm ruft Wx.pm unter /usr/lib,  weil ich habe das Wx Beispiel durchgef�hrt
17:48 jagd ich versuche nochmal
17:48 Sewi find, replace, about one of them seems to be the reason
17:48 Sewi Or the error is just random :-)
17:53 jagd die Meldung, die /usr/lib....  enth�lt, ist wahrscheinlich ein Problem in meinem lokalen Wx Modul.
17:55 jagd aber wenn ich debug (step over) das Wx Beispiel,  dann beschwert es immer eine andere Meldung:
17:56 jagd panic: free from wrong pool at /home/wu/projects/padre/padre/Padre/lib/Padre.pm line 242.
17:56 jagd sorry,  die Meldung ist dieselbe ..  :P
18:15 Hyppolit svn: r11071 | jagd++ | http://padre.perlide.org/trac/changeset/11071
18:15 Hyppolit update chinese translation
18:15 Hyppolit trunk/Padre/share/locale/
18:34 alvar joined #padre
18:38 Sewi Which Wx do you have?
19:11 Sewi left #padre
19:24 szabgab ge
19:24 szabgab lots of test failures on my machine now
19:28 jagd Sewi: Wx 0.9701  from CPAN
19:37 jagd Failed 5/38 test programs. 5/1211 subtests failed.
19:39 danlucraft joined #padre
19:45 Sewi joined #padre
19:47 LeK joined #padre
19:47 Sno|Laptop joined #padre
19:50 LeK hi
19:52 Sno|Laptop hi
20:07 Sewi hi
20:08 Sewi Would you two like to try out the current trunk (svn up) if there are any crashes?
20:14 szabgab I have lots of test failures
20:14 szabgab details sent to mailing list
20:16 Sewi szabgab: Don't worry, I might add another one soon
20:18 Hyppolit svn: r11072 | Sewi++ | http://padre.perlide.org/trac/changeset/11072
20:18 Hyppolit Add version check
20:18 Hyppolit trunk/Padre/t/
20:18 * Sewi didn't break it, just noticed that it's broken again
20:21 Sewi But I think, /me broke local.pm
20:24 Hyppolit svn: r11073 | szabgab++ | http://padre.perlide.org/trac/changeset/11073
20:24 Hyppolit move localization code to Padre::Locale
20:24 Hyppolit trunk/Padre/lib/Padre/ trunk/Padre/lib/Padre/Wx/Menu/
20:27 szabgab Sewi,   In the View/Language menu did you want to have what we call "utf8text" in the Padre::Locale file?
20:28 szabgab or in English for every language?
20:29 Sewi I thought about utf8
20:31 Sewi Like "(?"?) ?????, ??? ????? ?????, ???? ????, ??? ?????
20:31 Sewi
20:31 Sewi (?') ????, ?????, ????? ??????, ?? ??????
20:31 Sewi
20:31 Sewi (?"?) ????, ??? ????, ??? ???? " for hebrew
20:31 Sewi (Result of babylon ;-)
20:32 Sewi Fran�ais for french
20:32 Sewi "Deutsch" for German
20:32 Sewi and so on...
20:36 szabgab so that is the utf8text that we have
20:37 szabgab but for some reason it comes out broken for me, still researching
20:39 Sewi that locale error you posted on the ml is fixed now
20:40 Hyppolit svn: r11074 | Sewi++ | http://padre.perlide.org/trac/changeset/11074
20:40 Hyppolit Fixed locale test errors
20:40 Hyppolit trunk/Padre/lib/Padre/
20:41 Sewi svn up and try again - it might be responsible for your broken output, too.
20:41 Sewi What about moving wherefrom.cgi to the document root?
20:41 Hyppolit svn: r11075 | szabgab++ | http://padre.perlide.org/trac/changeset/11075
20:41 Hyppolit use the utf8text to display on the menu - but it shows up broken
20:41 Hyppolit trunk/Padre/lib/Padre/ trunk/Padre/lib/Padre/Wx/Menu/
20:47 Hyppolit svn: r11076 | szabgab++ | http://padre.perlide.org/trac/changeset/11076
20:47 Hyppolit decode utf8 string for display
20:47 Hyppolit trunk/Padre/lib/Padre/
20:47 szabgab Sewi, I'll take a look at it tomorrow, I am just too tired for anything real now
20:47 szabgab but pls check the View/Language menu now
20:51 Sewi Looks good
20:52 Sewi (Except that I can't read most things, but this is not a display problem :-) ).
21:00 dorkfish joined #padre
21:07 Hyppolit svn: r11077 | szabgab++ | http://padre.perlide.org/trac/changeset/11077
21:07 Hyppolit fix version test problem
21:07 Hyppolit trunk/Padre/script/
21:08 Sewi thanks
21:14 TheBlack joined #padre
21:41 danlucraft1 joined #padre
21:42 szabgab hmm who is TheBlack ?
21:43 Getty a dude from cebit
21:43 Getty if i dont mixup the nicks right now
21:44 szabgab Getty, I have not seen a report from you about CeBIT, have I missed it?
21:45 szabgab and btw what is your PAUSE user?
21:45 Getty getty
21:46 Getty and i didnt made one till now, i was busy all week with bringing my work to go, and also trying to setup the blog where i wanna write it
21:46 szabgab that is surprisingly empty there
21:46 Getty yeah, so far
21:46 Getty on github there is something
22:09 Hyppolit svn: r11078 | szabgab++ | http://padre.perlide.org/trac/changeset/11078
22:09 Hyppolit add some prereqs, another blog entry
22:09 Hyppolit trunk/Perl-Staff/ trunk/Perl-Staff/lib/Perl/
22:12 SvenDowideit joined #padre
22:13 Hyppolit svn: r11079 | szabgab++ | http://padre.perlide.org/trac/changeset/11079
22:13 Hyppolit add auto_provides to Makefile.PL
22:13 Hyppolit trunk/Padre/
22:46 cognominal joined #padre
23:13 cognominal joined #padre
23:15 acidtrip joined #padre
23:25 acidtrip left #padre
23:35 mib_2znizy joined #padre

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