Camelia, the Perl 6 bug

IRC log for #padre, 2010-03-13

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

All times shown according to UTC.

Time Nick Message
00:06 dorkfish joined #padre
00:17 teejay joined #padre
02:34 Hyppolit svn: r11088 | adamk++ | http://padre.perlide.org/trac/changeset/11088
02:34 Hyppolit Don't run the auto_provides unless we've previously run make manifest
02:34 Hyppolit trunk/Padre/
03:57 Sno|Laptop joined #padre
05:48 Alias <Steffen> Everybody seems to have preferred the alias. But using my actual first name has the advantage of naturally teaching people how to spell it ;P
05:48 Alias Steffen++
06:14 Alias Did someone break all the "Run command" stuff?
06:14 Alias None of the module tools seem to work any more
06:14 Alias In fact, their code doesn't remotely seem to match the API either
06:14 Alias And I can't get any program to write output to the Output window
06:28 SvenDowideit_ joined #padre
06:53 pece joined #padre
06:55 danlucraft joined #padre
07:14 pece joined #padre
07:22 pece joined #padre
07:40 Alias Goddam, no wonder
07:40 * Alias finally figures out how to get Wx::Display working
07:40 Alias Which means I can finally implement that smarter default size and position logic
07:59 Alias Crap
07:59 Alias So Display logic is possible, but file logic isn't sucked in from Wx
07:59 Alias Meaning no object way to implement incremental search stuff
07:59 Alias ...obvious
08:18 szabgab Alias, I see you enjoy talking to yourself :-)
08:21 Alias Call it stream-of-conciousness blogging :)
08:22 Alias I really really want to have this damned recursive search
08:22 Alias I'm slowing going insane waiting on things to happen
08:22 Alias I'm starting to think that maybe we need to integrate AnyEvent or something
08:22 Alias Since there's no asyncronous file IO
08:25 Alias I think it might be time to play around a bit with the idle event handler
08:25 Alias Starting with the directory search
08:36 Hyppolit svn: r11089 | adamk++ | http://padre.perlide.org/trac/changeset/11089
08:36 Hyppolit Dumping display geometry works
08:36 Hyppolit trunk/Padre/lib/Padre/Plugin/
08:45 Hyppolit svn: r11090 | adamk++ | http://padre.perlide.org/trac/changeset/11090
08:45 Hyppolit Everything that needs threads should make sure they are loaded themselves. This removes the need to load threads during a client call to the single instance server.
08:45 Hyppolit trunk/Padre/script/
08:55 waxhead joined #padre
09:07 Alias Wow, the task system does some pretty evil things
09:08 szabgab You know I am more and more worried that Padre got so unstable lately
09:08 Hyppolit svn: r11091 | waxhead++ | http://padre.perlide.org/trac/changeset/11091
09:08 Hyppolit Rename the new method to a more correct named.  Added comment to explain why it was added.
09:08 Hyppolit
09:08 Hyppolit I'm not sure it belongs here.
09:08 Hyppolit trunk/Padre/lib/Padre/Wx/
09:08 Alias szabgab: I agree. It's not just that we're unstable, it's the SIZE of some of them
09:08 Alias szabgab: Maybe it's time to start doing dev releases
09:09 Alias I'm also a little concerned by stuff beyond just crashing
09:09 szabgab basically that's what happening now
09:09 Alias Entire areas of Padre have stopped working, due to changes in those very areas
09:09 szabgab except we still call them with full version numbers
09:10 Alias True
09:10 Alias That's only really happened in the last couple of releases in a really serious wat
09:10 Alias That's only really happened in the last couple of releases in a really serious way
09:10 Alias Although the current codebase can be made stable with a one line change
09:10 Alias Those exit-time-crashes can be easily stopped
09:11 szabgab exit() as the first line of padre ?
09:11 Alias Look at Padre::Startup line 111
09:12 Alias I only committed something that crashy because I knew wxhead/etc can stop the crashing when they build the release
09:12 szabgab I am actually less worried about that - though its nasty -
09:12 szabgab more about things like F2 not working
09:12 szabgab let me check if that is still the case
09:13 waxhead hey guys
09:13 szabgab oh it is working now on trunk
09:13 waxhead Padre is very crashy
09:13 szabgab hi waxhead
09:13 waxhead open Main.pm for a whlie
09:13 waxhead while even
09:14 waxhead I also have Padre::Wx::Menu open too...
09:14 waxhead I just noticed it's crashed in the back ground...
09:14 Alias I get a window opening
09:14 Alias That then hangs
09:14 Alias oh, for a while
09:14 Alias Then it works
09:14 Alias waxhead: Padre::Startup line 111
09:14 waxhead I get a memory dump...
09:14 Alias waxhead: Change what it says in that comment and then try again
09:14 waxhead ok
09:15 waxhead I was actually going to alt-tab to do that when you said it a bit ago.. only to find Padre had died..
09:15 waxhead oh.. I'm running dev.pl -a
09:15 szabgab (in cleanup) Can't locate object method "refresh_view_document_as" via package "Padre::Wx::Main" at /home/gabor/work/padre/Padre/lib/Padre/Wx/Main.pm line 1133.
09:15 szabgab I just got the above warning on the console
09:16 waxhead ARGH!
09:16 Alias Weren't you the one playing around with the View menu?
09:16 Alias :)
09:16 waxhead ok.. let me fix that...
09:16 waxhead I changed the method name and didn't change it in the refresh...
09:16 Alias You didn't do a Find in Files when changing a method? Naughty
09:16 szabgab and t/15-locale.t is now failing
09:17 waxhead szabgab, I opened a ticket for that one
09:17 waxhead local that is
09:17 szabgab I guess I should invest more energy in Padre, I just have negative time nowdays
09:18 szabgab waxhead, it was passing just a few revisions ago
09:18 Alias It passes for me
09:18 Hyppolit svn: r11092 | waxhead++ | http://padre.perlide.org/trac/changeset/11092
09:18 Hyppolit Fixed the refresh method to call the renamed refresh_view_menu method.
09:18 Hyppolit trunk/Padre/lib/Padre/Wx/
09:19 waxhead szabgab, yes it was...
09:19 waxhead I think it started failing for me post 0.58 release.. so it's only recent
09:19 szabgab DBD::SQLite::db do failed: SQL logic error or missing database cannot VACUUM from within a transaction at (eval 163) line 36.
09:19 szabgab that's a nice one
09:19 waxhead szabgab, svn up now
09:19 szabgab using git :-)
09:20 waxhead Alias, I'm not sure what I have done should be done... I've added in a method in Main.pm to force the View Menu to be updated so the View document as works correctly when tabbing between documents
09:20 waxhead szabgab, are you?
09:20 waxhead szabgab, like it?
09:21 szabgab I keep getting confused and mostly use it for offline commits
09:21 waxhead I see that subversion is now an apache project
09:22 Alias waxhead: That probably shouldn't be necesary
09:23 Alias waxhead: The reason there's some methods there that only refresh part of the menu is because those parts of the menu are horribly expensive
09:23 waxhead Alias, I added it in as the change of tabs wasn't changing the view document as...
09:23 szabgab interesting now the test passes
09:23 Alias ... and curiously, everything ELSE gets updated?
09:23 waxhead it's over the top to do the whole menu, it just needed to update the view document as list
09:24 waxhead does it?
09:24 Alias The disabled'ness of the undo command
09:24 Alias The top level Perl menu appears and dissapears
09:24 Alias Tons of menu elements need to be changed when you tab
09:25 Alias And it's mostly fast
09:25 Alias It's just that TWO parts of the menu (the recent files list and the window list) are very very expensive, and only change in rare situations
09:25 Alias So they have their own sections
09:25 Alias Wait, view document as?
09:26 Alias There's an entire section in the view menu for this
09:26 Alias Why isn't that gettings fired anyway?
09:26 waxhead yeah there is
09:26 Alias It should be
09:26 waxhead I don't know.. hence the horrible hack to force the whole view_menu
09:27 szabgab Alias, waxhead what do you think about moving the "File/New"  inside the "File/New../"  ?  and same with open ?
09:27 Alias szabgab: I don't want to have to navigate two menu levels to the most common file operations of all
09:27 waxhead try taking it out.... you can test it simple enough, open a perl file check what's set, then open a text file, check that... then tab back to the perl file.. if you check the view document as it's text still
09:27 szabgab you'd probably press ctrl-o and cntrl-n anyway?
09:28 waxhead well given the multitude of ways to open things...
09:28 Alias It means instead of click->drag-down->release (which is fast) you have to do click->drag-down->wait-for-submenu->drag-right-b​ut-not-too-far->try-to-understand-which-one-is-j​ust-ordinary-new-file->drag-down-again->release
09:28 waxhead but I'd rather see how other applications do it
09:28 Alias szabgab: I can't reach ctrl-o and ctrl-n with one hand
09:28 Alias I never use them
09:28 szabgab I have two hands :-)
09:28 Alias I have one hand on the mouse half the time
09:29 szabgab oh the mouse
09:29 szabgab where is my mouse?
09:29 szabgab I have not seen it for some time now :-)
09:29 Alias Anyways, I put them there because in ever other application I've had that did it that way, I HATED it
09:29 Alias I like the symmetry too
09:29 szabgab so that's the big difference between people used to Windows and Linux (vi/emacs) I think
09:30 Alias "New" "Complicated New" "Open" "Complicated Open"...
09:30 szabgab ok
09:30 Alias If we get the popcon server working, I'd bet good money that "New" and "Open" are used more than anything other kind of new or open combined
09:31 Alias Oh, also
09:31 Alias It gives us permission to make the submenus are horrendously complicated as we like
09:31 Alias And let plugins hack the shit out of them
09:31 Alias s/are/as/
09:32 Alias waxhead: Rip out your hack, something else is going wrong
09:32 Alias I just spent most of a week untangling all the menu/refresh logic :)
09:32 waxhead remove it clean or comment it out for now?
09:32 Alias Remove it clean
09:33 Alias It's the wrong approach regardless
09:33 Alias If that's not firing, then the per-document word-wrap stuff isn't firing either
09:33 Hyppolit svn: r11093 | waxhead++ | http://padre.perlide.org/trac/changeset/11093
09:33 Hyppolit Removed all traces of the refresh view menu as requested by Alias.
09:33 Hyppolit trunk/Padre/lib/Padre/Wx/
09:34 Alias And the font size
09:34 Alias And anything else in View that's per-document
09:34 Alias Where is the event handle you were looking at?
09:36 waxhead Padre::Wx::Menu::View line 75
09:36 Alias The event handle
09:36 Alias That's the actual update code
09:36 waxhead oh.. is it?
09:37 waxhead I don't know.. I don't think I ever found it...
09:37 waxhead <blush>
09:37 Alias Wait, if you didn't even find it, what on earth was calling that custom refresh function
09:37 waxhead Padre::Wx::Main was...
09:37 Alias (... that you wrote)
09:38 Alias ??? -> Main->refresh_view -> ::View->refresh
09:38 waxhead there's  a refresh method in that that seems to selectively refresh parts of Padre
09:38 Alias You added a new refresh method
09:38 Alias And what did you have called THAT
09:38 waxhead I just removed it..
09:38 waxhead Padre::Wx::Main::refresh
09:39 Alias oh...
09:39 Alias That's a very very big refresh
09:39 waxhead I was just following what was done before...
09:39 waxhead I had no idea where else it should be done..
09:40 waxhead what's worse, I only wanted to fire the view document as only..
09:40 waxhead but couldn't work that out, so did the lot...
09:40 waxhead as in the whole of the View menu...
09:41 Alias Well, ->refresh_menu is the refresh for the whole menu
09:41 Alias And that should be included in the refresh lock on tab changes
09:41 Alias hrm
09:42 waxhead there's not refresh-menu, only refresh_menubar
09:42 Alias oh right
09:42 Alias I'm trying to remember how this works now
09:43 Alias The notebook event you want is in ::Notebook
09:43 waxhead well, I spent a long night trying to work out how it does it thing... and failed!
09:43 waxhead then stumbled on the bits that I needed to fix the view document as bug...
09:43 waxhead then found the refresh problem...
09:44 Alias on_auinotebook_page_changed
09:45 waxhead I'm looking at it, but don't see what I'm looking at.. :)
09:45 Alias hrm
09:45 Alias ->lock
09:45 waxhead is this where we should do the view document as update?
09:46 Alias oh man, there's some junk I didn't clean up I think
09:47 Alias What the hell
09:47 waxhead you know, this is where I'd like to live an easy walk away and drop in and see what you're up to... it would make learning a whole lot simpler.. :)
09:47 Alias :)
09:47 Alias I've been pondering copying the "code walkthrough" idea from Dreamwidth
09:48 Alias Except where there idea of a "walkthrough" is "a description of the things we changed" mine would be "How subsystem $foo works"
09:48 Alias their
09:48 Alias I think I've made a mistake here
09:48 Alias Or some gremlin has been "optimising" things again
09:48 waxhead Alias, I think that's one of the major problems all major projects suffer.. that "how things work" stuff...
09:49 Alias Especially as things get more sophisticated
09:49 Alias Locking and refresh is definitely not really very intuitive any more
09:49 waxhead I really am just on the outside looking in with all of this... I can flounder my way around, but I'm in no way "knowledgeable" with this...
09:49 Alias So basically, someone has made ->refresh_menubar into a null stub
09:50 Alias Or I forgot to revert it when I reverted a dozen other things
09:51 waxhead yeah, I can see that...
09:52 waxhead I was looking in Menu, since the ->menu I assumed it was menu, not menubar....
09:52 Alias Yeah, we could probably name that better
09:52 Alias But that's for another (more pedantic) time
09:52 waxhead I'll rename it... :)
09:53 waxhead will it break much?
09:53 Alias Every now and then I go on a naming-pedant rampage
09:53 Alias Right now it would break stuff
09:53 waxhead I'm thinking I might have a crack a making the ack output a little nicer... it's a request in trac at the moment
09:53 Alias I'll do it next time I do a round of GUI structure cleanups
09:53 waxhead OK...
09:54 waxhead oh.. wait, the ack output or the rename?
09:54 Alias The rename
09:54 Alias We actually need to do an entirely new generation of menu update in light of the action system
09:54 Alias But there's still bigger issues to fix first
09:56 waxhead Alias, there's always going to be bigger issues to fix...
09:56 Alias Oddly, I've been hoping to make Ack irrelevant :)
09:56 Alias By writing our own recursive search
09:56 waxhead oh.. no worries then...
09:56 Alias But unfortunately, that requires that I learn the task system
09:56 Alias So you go ahead
09:58 Alias We'll probably keep the same GUI stuff anyway, just replace ack itself with something else
09:59 waxhead sure...
09:59 waxhead next time I feel like a hair pulling exercise...
10:02 waxhead wow the SDL guys are doing some stuff...
10:03 Alias They are doing a huge rewrite yeah, because the old maintainer who didn't want to do a full rewrite has pretty much walked away from SDL now
10:04 waxhead cool...
10:05 Hyppolit svn: r11094 | adamk++ | http://padre.perlide.org/trac/changeset/11094
10:05 Hyppolit Untangle the situation with refresh -> menu
10:05 Hyppolit trunk/Padre/lib/Padre/Wx/ trunk/Padre/lib/Padre/Wx/Menu/
10:05 Alias Done
10:05 waxhead isn't there some game written in SDL that's quite popular... I remember being surprised it find out it was written in Perl.. can't think what it's called now... has a penguin shooting balls...
10:05 Alias yes
10:06 waxhead that game and Padre really showed me perl isn't that slow...
10:06 Alias Although I think for the newest version they ported away
10:06 Alias Indeed
10:06 Alias Case in point
10:06 Alias I just made the entire menu tree refresh fire on tab change
10:06 Alias Padre is not even REMOTELY slower
10:07 Alias (Because we ripped out the two specific things that were slowing menu refresh down, windowlist and recent files
10:07 waxhead is that all you changed?
10:07 Alias Yup
10:07 Alias Move ->refresh_menubar to ->refresh_menu
10:07 Alias And then deleted the ->refresh_menubar and the Menubar->refresh_top methods
10:10 waxhead yeah, I see...
10:10 Alias Which is what it probably should have been in the first place
10:11 waxhead this is odd...
10:11 waxhead Can't locate package PPI::Structure for @PPI::Structure::Constructor::ISA at /home/pete/Programming/Perl/Padre​/trunk/Padre/lib/Padre/Wx/Ack.pm line 274.
10:11 Alias oh rly
10:11 Alias Curious
10:11 Alias Does it still happen when you disable the slave driver optimisation
10:11 waxhead I just fired up cpan and installed PPI
10:11 waxhead how the heck was it not installed?
10:13 waxhead shouldn't we list PPI as a requirement?
10:13 Alias It is
10:13 waxhead hmm..
10:16 waxhead Alias, nice.. it's working fine...
10:19 waxhead oh.. I meant the menu refresh works fine...
10:20 Alias yup
10:20 waxhead the ack errors are still happening.. I'll try turning off the optimisation thing in Startup
10:20 Alias That's what it should have done all along
10:22 shadowpaste "waxhead" at 217.168.150.38 pasted "ack and optimisations are disabled" (5 lines) at http://paste.scsys.co.uk/40682
10:22 waxhead Alias, check the nopatse
10:23 waxhead paste even..
10:23 waxhead I've got if( 0 ) { at line 111 in start up..
10:23 waxhead no PPI Errors, but get these instead...
10:23 Alias In that case, I don't know what the problem is
10:23 waxhead how the heck do you even start to debug stuff like that?
10:24 Alias You don't
10:24 Alias You leave it to the thread'perts
10:24 Alias Or you learn the threading system
10:24 waxhead Steffen would be about it wouldn't it? :)
10:24 Alias Steffen, submersible, Sewi to a lesser extent, and I'm trying to learn it
10:25 waxhead submersible is across the threading stuff?
10:25 Alias Yup
10:25 Alias He wrote the ::Service version
10:25 waxhead that's neat... I really thought it was only down to Steffen...
10:25 Alias And Swarm plugin
10:26 waxhead yeah I knew about swarm...
10:27 waxhead oooo... outline doesn't like ::Service!
10:31 waxhead hmm outline is seems broken
10:34 waxhead ahh!!
10:34 waxhead Alias, diabling the optimisation in Startup breakes the Outline view
10:34 waxhead disabling even
10:39 Alias What?
10:39 Alias DISABLING it breaks something?
10:39 Alias Now _that_ shouldn't be possible
10:42 waxhead Alias, try it out...
10:42 waxhead at least turning off the disabling brings back the outline
10:42 * Alias has other targets
10:42 waxhead ok
10:43 waxhead I'm off... very tired.. long day... early start on a saturday!
11:10 karl joined #padre
11:27 zenog joined #padre
11:27 zenog hi guys
11:30 Alias hi
11:31 LeK hi
11:35 zenog One question: Is there a reason why two different search methods exist (normal and incremental)?
11:35 Alias Because we haven't reconciled the two, and because of things like Find All
11:36 zenog Ah okay.
11:37 zenog But if the incremental search had all necessary features, we could get rid of the popup-window-based one?
11:37 Alias Nah
11:37 Alias You add MORE features to the dialog one, and then you make Ctrl-F open the incremental search
11:37 Alias And then ANOTHER Ctrl-F move it from incremental to full dialog
11:38 zenog hm
11:39 Alias For example, "Find method, find blah blah"
11:50 Hyppolit svn: r11095 | adamk++ | http://padre.perlide.org/trac/changeset/11095
11:50 Hyppolit One last tweak for the menu/menubar fix
11:50 Hyppolit trunk/Padre/lib/Padre/
11:55 Hyppolit svn: r11096 | adamk++ | http://padre.perlide.org/trac/changeset/11096
11:55 Hyppolit Adding a new library for calculating various useful window geometries
11:55 Hyppolit trunk/Padre/lib/Padre/Wx/ trunk/Padre/t/
12:08 Sewi joined #padre
12:11 Sewi Hi all
12:11 Sewi Did anybody change the default editor fontsize?
12:30 Hyppolit svn: r11097 | zenogantner++ | http://padre.perlide.org/trac/changeset/11097
12:30 Hyppolit one English message fix; more German translation updates
12:30 Hyppolit trunk/Padre/lib/Padre/Action/ trunk/Padre/share/locale/
12:45 Alias Sewi: Not that I know of
13:02 danlucraft joined #padre
13:30 Sewi Ok, must be something local...
13:36 karl_ joined #padre
13:57 Hyppolit svn: r11098 | adamk++ | http://padre.perlide.org/trac/changeset/11098
13:57 Hyppolit - New Padre::Wx::Display screen geometry library for handling multiple
13:57 Hyppolit screens, weird geometry setups and other weird stuff that coders
13:57 Hyppolit have on their development setups. Padre's main window uses this to
13:57 Hyppolit calculate an elegant golden-ratio uniform-margin default position
13:57 Hyppolit and size (ADAMK)
13:57 Alias Kapow!
13:57 Hyppolit trunk/Padre/ trunk/Padre/lib/Padre/Wx/
13:57 Alias And at a single stroke, one thing I hate ENORMOUSLY about Padre is fixed :)
13:58 sjn :)
13:59 sjn does that mean there's a new release soon?
13:59 Alias God knows
13:59 Alias Not my job
13:59 Alias I just try to make Padre suck less, or at least suck faster
14:00 Alias And I think it's time I learned threads, because there's some stuff in there like looks rather like abominations
14:00 Alias And I'd very much like to know if they are necesary
14:00 * sjn throws a cookie to szabgab # Go and make a release!
14:04 Sewi sjn: waxhead does this atm.
14:04 sjn \o/
14:04 Sewi But we need to be a bit stable before :-(
14:04 sjn aw
14:05 sjn "stable" in what manner?
14:05 sjn spurious crashes?
14:05 Sewi hopefully non of them
14:07 Alias Yeah, we've done two bad ones
14:07 Alias Be nice to do a good one
14:07 Sewi Would be great - I promised that 0.58 will be a stable one to the people on CeBit :-)
14:22 submersible braaaains
14:33 Hyppolit svn: r11099 | adamk++ | http://padre.perlide.org/trac/changeset/11099
14:33 Hyppolit When showing a toolbar panel for the first time when it wasn't needed at startup time, make sure to set the interface lock state on or off properly to match the main_lockinterface configuration state.
14:33 Hyppolit trunk/Padre/lib/Padre/Action/ trunk/Padre/lib/Padre/Wx/
14:34 Alias submersible: BTW, the ABC needs your help
14:34 submersible ABC?
14:34 Alias http://www.abctv.net.au/tv/catalyst/su​rvey/fatherhood/index.php/form/page/2
14:35 submersible yeah - the new dads show
14:35 submersible not sure how keen i am
14:35 Alias Figured since we had two...
14:35 submersible :)
14:35 Alias You don't need to be ON the show, it's a damned survey
14:40 Hyppolit svn: r11100 | adamk++ | http://padre.perlide.org/trac/changeset/11100
14:40 Hyppolit Noting the lockinterface thing
14:40 Hyppolit trunk/Padre/
14:42 Hyppolit svn: r11101 | adamk++ | http://padre.perlide.org/trac/changeset/11101
14:42 Hyppolit Piddling tweaks (mostly docs)
14:42 Hyppolit trunk/Padre/lib/Padre/
14:43 Hyppolit svn: r11102 | adamk++ | http://padre.perlide.org/trac/changeset/11102
14:43 Hyppolit Removing an import
14:43 Hyppolit trunk/Padre/lib/Padre/
14:43 Alias submersible: Ping?
14:44 submersible pong
14:44 Alias submersible: I'm reading the threading/task code
14:44 submersible mmm
14:44 Alias And we dirty the hell out of the task objects
14:44 submersible explain 'dirty'
14:44 submersible it is not pretty in there
14:44 Alias We write stuff into it's hash
14:44 Alias Would doing what we do with plugins work?
14:45 Alias Wrapping a ::TaskHandle wrapped around the task object
14:45 Alias Pass the TaskHandle into the thread, and let it unpack, load the code etc in there
14:46 Alias We also suggest that people generate their own shared variables
14:46 submersible presently i think that the return event collection is based on refaddr/stringamafication of the task instance
14:46 submersible but in principal it can be any old thing like an assigned id from a task handle or similar
14:46 Alias But, if we do this slave master thing, you CAN'T make your own shared variables right? Because you have to set up the share before the first spawn?
14:46 Alias Or am I wrong ?
14:46 submersible what is the problem u want to solve
14:47 Alias I want to throw events into Padre with impunity
14:47 submersible runtime :shared seems to work for service queues
14:48 Alias OK, but what if we just registered two shared event objects?
14:48 Alias ::TaskMethod and ::WxMethod
14:48 submersible from a task - throw arbitary events back to Main
14:48 submersible ??
14:48 Alias The first is an event that just passes it's params through to a named method on the task
14:48 pece joined #padre
14:48 Alias The second is just an event that passes it's params through some object it gets from Wx by id
14:48 Hyppolit #886: Crash on locked database (new defect) [ http://padre.perlide.org/trac/ticket/886 ]
14:49 submersible there is no event handler on the slave side remember
14:49 Sewi Sorry, I tried but wasn't able to figure out where to fix this
14:49 Alias submersible: For now, I only mean coming back FROM the task
14:49 Alias So TaskMethod triggers a method in the main-thread task object
14:50 submersible i don't see why not. end of the day - the wx::NewEventType is just an incrementing counter ... so you could try forgetting the :shared guff and just pass down the integer to the slave
14:50 submersible I think
14:52 Alias I notice we've also broken some other stuff
14:52 Alias Because now the task modules aren't actually loaded in the workers
14:53 Alias Or we potentially have anyway
14:53 submersible yeah - the slave should perhaps check that the instance it has been passed is in fact loaded :)
14:55 Alias OK, so in principle, I really should be able to dig in and rewrite a bunch of stuff?
14:55 Alias As long as I'm fairly careful
14:55 submersible I think the tests are pretty poor.
14:55 submersible u could break almost everything before they complained
14:56 Alias And I have
14:56 Alias The configuration rewrite was epic
14:57 Alias I think from memory it broke 100% of plugins
14:57 submersible the config_read config_wriite stuff?
14:57 Alias Oh
14:57 Alias The human/host split, all that
14:57 Alias I renamed about 75% of them in the process
14:58 Alias OK, so another thing
14:58 Alias Can you do a similar thread event pushing thing without using Wx at all?
14:58 Alias Could we gateway it through naked threads.pm doohickies?
14:58 Alias Because if we could, we could spawn the slave master before we've loaded Wx at all, and that would save us BUCKETs of memory
14:59 submersible there is still more juice to squeeze from wx
14:59 Alias OK
14:59 Alias So first step, have zero native Wx objects anywhere in the data at all
14:59 submersible I think it is possible to load JUST the wx PlThread-ness early
14:59 Alias And switch everything to ids
14:59 Alias So if we can go native later, it doesn't hurt
14:59 Alias Sound reasonable?
15:00 submersible to me - RAM scraping is not very reasonable... but that's just me
15:00 submersible afk 5 mins.. sry
15:00 Alias I gotta head home
15:01 Alias At work doing load testing guff, badly
15:01 sjn testing guff isn't fun
15:02 sjn (how does one test guff, btw?)
15:02 sjn :)
15:08 Hyppolit svn: r11103 | Sewi++ | http://padre.perlide.org/trac/changeset/11103
15:08 Hyppolit Add a test which crashes Padre
15:08 Hyppolit trunk/Padre/xt/
15:13 ElwoodJ joined #padre
15:18 submersible ping alias
15:23 Hyppolit svn: r11104 | Sewi++ | http://padre.perlide.org/trac/changeset/11104
15:23 Hyppolit Add two actions
15:23 Hyppolit trunk/Padre/lib/Padre/Action/
15:29 Karl joined #padre
15:33 Alias_ joined #padre
15:33 Alias_ submersible, btw this is the hyper-fast api I was talking about
15:33 Alias_ http://search.cpan.org/~mlehm​ann/AnyEvent-5.251/lib/AE.pm
15:33 Alias_ You were right, I think
15:34 submersible ok - regarding padre and wx and threads
15:36 Karl prove -blv t/15-locale.t
15:36 submersible to launch your slaves early, instead of a threads shared queue input and wxevent output... we replace the whole thing w/ queues in both directions... then let the foreground wx poll or trap new data on the return queue.
15:36 Karl 16:26:33: Error: Cannot set locale to 'de_DE'.
15:36 Karl ok 4 - ->change_locale(de)
15:37 Alias_ submersible, you'd have to poll it bloody quickly though wouldn't you?
15:38 submersible what latency do u neeed from a task
15:38 submersible ?
15:38 Alias_ Depends how small the jobs you want to spew out are
15:39 submersible u can listen to WxIDLE and check queue there too
15:39 Alias_ Absolutely
15:39 Alias_ I have big plans for IDLE
15:39 submersible or some thing whacky w/ signals
15:40 Alias_ Probably wouldn't work, given platform hijinx
15:40 Alias_ So then we stick with a highly stripped down Wx for now
15:40 Alias_ And solve the other problem later
15:40 Alias_ Once we have a speed measurement under ideal conditions
15:41 Alias_ But I imagine something like 50milliseconds?
15:41 Alias_ 20/second
15:41 Alias_ On the plus side, that lag would effectively be ammortised across all the workers
15:42 Alias_ And idle would cut it even more
15:46 Alias_ Anyways, bed
15:46 Alias_ I'll leave you to do whatever it is you are doing at this time
15:46 Alias_ :)
15:47 submersible baby poo
16:32 wikik joined #padre
16:32 wikik hi
16:32 wikik i'm new with padre and perl
16:32 wikik i'm trying to install a cpan module but i don't know how to do it
16:33 wikik i'm doing it on windows...
16:33 wikik anybody can help me?
16:36 submersible wikik: what perl are you using?
16:36 submersible the padre standalone ?
16:37 wikik i've installed the last padre version
16:37 wikik and nothing else
16:37 submersible run padre - look in the tools menu => module tools
16:38 wikik i tried clicking on installing a CPAN module
16:38 wikik the problem is that padre craches
16:39 submersible you might need to talk to a more windows savvy person.
16:39 submersible sry
16:39 wikik jejej
16:39 wikik it doesn't matter
16:39 wikik thank you for your help
16:39 wikik :D
16:39 submersible np
16:39 submersible stick around
16:40 submersible maybe someone else has suggestion
16:44 dolmen joined #padre
16:44 LeK try doing it manually
16:45 LeK you should have a strawberry directory under c:\
16:45 wikik yes
16:45 LeK if perl's on your path, just type 'cpan' and it should work
16:45 LeK module install from menu works for me with .57 r10863
16:45 Hyppolit Changeset #10863 http://padre.perlide.org/trac/changeset/10863
16:46 LeK (funny; i had never used this menu before)
16:48 wikik i used the terminal
16:49 wikik i've done get Net:SSH:Perl
16:49 wikik and then make install Net:SSH:Perl
16:50 Sewi wikik: Type "cpan Net::SSH::Perl" - it should be enough.
16:50 LeK install Net::SSH::Perl should have be enough :)
16:50 Sewi If you're already running the cpan client, true
16:50 LeK yep
16:50 * LeK is away... brb in half an hour
16:52 wikik when i compile the script it returns the same error as before
16:52 wikik Can't locate Net/SSH/Perl.pm in @INC (@INC contains: C:/strawberry/perl/lib C:/strawberry/perl/site/lib C:\strawberry\perl\vendor\lib .) at C:\Users\Vicente\Desktop\regresion.pl line 7.
16:52 wikik ups
16:52 wikik i haven't seen this
16:53 wikik after installing the CPAN module it finishes with an error...
16:53 wikik it fails during the tests
16:58 wikik how can i download the last changeset of padre?
17:04 Sewi List of changes or newest version?
17:04 wikik the newest version to don't have this problem when installing the CPAN modules
17:04 wikik or maybe is better to install an older version i don't know...
17:06 Sewi I don't think that it's a Strawberry version problem.
17:06 Sewi Which error did you get?
17:06 wikik well i have the problem when i tried to install the CPAN module from the padre IDE
17:06 wikik in this case the application craches
17:07 wikik the other problem is installing the SSH module using the terminal
17:07 wikik in that case it appears an error when testing the module
17:07 wikik and after that it doesn't finish....
17:10 Sewi _which_ error appears
17:13 wikik Test Summary Report
17:13 wikik -------------------
17:13 wikik t/03-packet.t    (Wstat: 2304 Tests: 0 Failed: 0)
17:13 wikik Non-zero exit status: 9
17:13 wikik Parse errors: Bad plan.  You planned 10 tests but ran 0.
17:13 wikik //hint// to see the cpan-testers results for installing this module, try:
17:13 wikik reports TURNSTEP/Net-SSH-Perl-1.34.tar.gz
17:13 wikik Running make install
17:13 wikik make test had returned bad status, won't install without force
17:13 wikik Failed during this command:
17:13 wikik TURNSTEP/Net-SSH-Perl-1.34.tar.gz            : make_test NO
17:15 Sewi I'll try myself...
17:15 wikik ok
17:15 wikik thanks
17:16 Sewi This is the error: t/03-packet.t ...... Use of uninitialized value $ENV{"HOME"} in concatenation (.
17:16 Sewi ) or string at C:\strawberry\cpan\build\Net-SSH-Perl-1.34-S798LQ\blib\lib/Net/SS
17:16 Sewi H/Perl.pm line 108.
17:16 Sewi The getpwuid function is unimplemented at C:\strawberry\cpan\build\Net-SSH-Perl-
17:16 Sewi 1.34-S798LQ\blib\lib/Net/SSH/Perl/SSH1.pm line 30.
17:16 Sewi # Looks like your test exited with 9 before it could output anything.
17:17 wikik do you know how can i solve the problem?
17:17 Sewi wikik: It's might be a Strawberry problem, but it's no Padre problem.
17:17 Sewi Please type /join #win32 to go to the Strawberry channel.
17:18 Sewi I already posted the error message there, so someone should be able to help you.
17:18 wikik thanks a lot Sewi
17:18 Sewi I'm off now...
17:40 wikik left #padre
18:01 ElwoodJ Hi there. Is here anybody who uses Padre under Mac OS X 10.6.2 Snow Leopard?
18:02 szabgab ElwoodJ, here and there we hear about people
18:02 szabgab but they need to install it from source
18:03 ElwoodJ from Source? via CPAN or the Sources from svn?
18:10 szabgab I'd say CPAN for a start
18:14 ElwoodJ ok, I still take a running start. It's a great IDE, I use it under Windows at work, but now I want to use it at home under Mac.
18:33 mib_2jrcvr joined #padre
18:36 mohsen joined #padre
18:41 mohsen joined #padre
18:45 szabgab I am happy you like it!
18:47 Hyppolit svn: r11105 | karl.forner++ | http://padre.perlide.org/trac/changeset/11105
18:47 Hyppolit Development related to the braces highlighting and matching.
18:47 Hyppolit
18:47 Hyppolit Added some methods to enforce a uniform definition of what is a brace (so currently it is {}[](). All features using braces share this same definition :
18:47 Hyppolit - highlighting
18:48 Hyppolit - goto matching brace
18:48 Hyppolit - select to matching brace
18:48 Hyppolit
18:48 Hyppolit Moved all related code into Editor.pm.
18:48 Hyppolit
18:48 Hyppolit Modified goto matching brace (CTRL+1) and select to matching brace in order to:
18:48 Hyppolit - use the defined braces
18:48 Hyppolit - respect the inside/outsideness of starting point, meaning that is the cursor is inside the braces; its detination point is also inside
18:48 Hyppolit - the select to matching brace should also now work properly if there is an existing selection to the right of a closing brace
18:48 Hyppolit trunk/Padre/lib/Padre/Action/ trunk/Padre/lib/Padre/Wx/
18:49 mohsen joined #padre
18:53 Hyppolit #484: Select a complete block from one brace to another (closed enhancement) [ http://padre.perlide.org/trac/ticket/484#comment:3 ]
19:20 Sewi ping karl
19:20 Karl yop
19:21 Sewi r11105 is nice. It's no problem in this commit, but if you continue to work on brackets, please remember that there might be cases where other languages behave differently.
19:21 Hyppolit Changeset #11105 http://padre.perlide.org/trac/changeset/11105
19:21 Karl I used what I saw the code in Editor.pm
19:21 Sewi Padre is written in Perl and mostly for Perl, but we try to support other languages, too, so all Perl-specific-stuff should go to the Document/Perl.pm (and maybe below)
19:21 Sewi I'm just mentioning this because you might face it in the future...
19:23 Karl to prevent matching on <>
19:24 Karl and anyway the scintilla does not support other braces
19:24 Karl but anyway I have encapsultated the code, so it could allow us to implement our own brace matching code in the future
19:53 Karl I'd like to implement the xemacs Paren Highlighting=Expression way of highlighting braces
19:53 Karl meaning that all the code between the two braces is highlighted
19:53 Karl what do you think ?
19:58 Karl I'll add a setting to enable it.
19:59 Karl I don't understang a thing about the scintilla styles, so I guess I'll reuse the code from the editor smart highlighting
20:04 szabgab Karl, I guess that's the way to go
20:05 Karl so it wolhd be of the same colour
20:05 Karl will
20:11 szabgab go for something and we'll see
20:23 Karl I can't get rid of this xxxxx survey dialog !!
20:43 Sewi Sure?
20:43 Sewi Please look at the about->info panel where your config dir is.
20:43 Sewi Do you have a config.yml there?
20:43 Sewi Does it have a "FeedbackDone" - Line?
20:46 Karl no "FeedbackDone"
20:47 Karl I think the vote does not work in my virtual box for networking reasons
20:47 Karl but still it is very annoying
21:00 Sewi Did you try to click on "skip" - this should set the feedbackdone without any network access
21:20 Karl ok, sorry. BTW the skip button does not display its text in its entirety
21:20 Sewi again? :-(
21:21 Sewi I don't know how to make it bigger
21:34 Hyppolit svn: r11106 | karl.forner++ | http://padre.perlide.org/trac/changeset/11106
21:34 Hyppolit added the find_matching_brace method and implemented brace expression highlighting like xemacs paren highlighting expression mode.
21:34 Hyppolit You have to enable it via the editor_brace_expression_highlighting setting.
21:34 Hyppolit trunk/Padre/lib/Padre/ trunk/Padre/lib/Padre/Wx/
21:34 Karl ok I got it working if you want to give a try
21:35 Karl and that's enought for today :-)
21:35 Karl btw when will I get pay ? ;-)
21:36 Karl ... enough - paid
21:42 karl_ joined #padre
21:55 ElwoodJ left #padre
22:36 Sewi gn8
22:44 karl_ joined #padre
23:50 Alias_ joined #padre

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