Camelia, the Perl 6 bug

IRC log for #padre, 2010-02-05

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

All times shown according to UTC.

Time Nick Message
03:41 Hyppolit svn: r10506 | submersible_toaster++ | http://padre.perlide.org/trac/changeset/10506
03:41 Hyppolit add to the growing pile of junk in junkyard/
03:41 Hyppolit trunk/Padre-Plugin-Swarm/junkyard/
04:24 submersible1 joined #padre
06:42 Sewi joined #padre
06:42 Sewi gm
06:45 Alias \o
06:52 pece joined #padre
06:56 awnstudio joined #padre
06:59 marcela joined #padre
07:01 pece joined #padre
08:25 danlucraft joined #padre
08:34 * submersible grins
08:35 submersible when you can 'Run this document in someone else editor' ....
08:35 submersible the possibilities of
08:35 submersible ( `hostname` , `ifconfig` , `ipconfig /all`, `who` );
08:35 submersible or similar - just frigging scare me
08:43 kaare joined #padre
08:59 Alias `rm -rf ~`
08:59 daxim joined #padre
09:02 Hyppolit svn: r10507 | submersible_toaster++ | http://padre.perlide.org/trac/changeset/10507
09:02 Hyppolit move out of the Padre::Wx namespace
09:02 Hyppolit move version
09:02 Hyppolit some pod
09:02 Hyppolit trunk/Padre-Plugin-Swarm/ trunk/Padre-Plugin-Swarm/junkyard/ trunk/Padre-Plugin-Swarm/lib/Padre/Plugin/ trunk/Padre-Plugin-Swarm/l​ib/Padre/Plugin/Swarm/Wx/ trunk/Padre-Plugin-Swarm/lib/Pa​dre/Plugin/Swarm/Wx/Resources/ trunk/Padre-Plugin-Swarm/lib/Padre/Service/ trunk/Padre-Plugin-Swarm/lib/Padre/Swarm/ trunk/Padre-Plugin-Swarm/lib/Padre/Swarm/Message/ trunk/Padre-Plugin-Swarm/lib/Padre/Wx/Swarm/
09:02 tsee joined #padre
09:03 submersible h! tsee
09:03 tsee hey
09:04 submersible i have had some more thoughts re slavedriver
09:04 submersible and the task api
09:04 tsee I'll be back in a couple of minutes, sorry.
09:04 tsee I really need to make coffee.
09:05 tsee Just got up and fired up the computer.
09:05 tsee But feel free to write. I'll read it in a minute.
09:06 submersible . for starters , catching events on something other than the padre main window seems to have merit.
09:08 Haarg joined #padre
09:08 Alias (Padre::Wx::App for example)
09:10 Alias (assuming we modified it so you could make a standalone app object early, without making main at the same time)
09:15 tsee Yes, I agree.
09:15 tsee I used the main window simply because I knew no better.
09:18 tsee I got my coffee, by the way.
09:20 kgish joined #padre
09:25 Alias The other great thing about putting something on top of App is that it's super-fast to get to in Padre::Current
09:25 Alias Because we can use Wx::wxTheApp or whatever
09:27 tsee You still need to fetch the main window in the post-run hook of the Task.
09:27 tsee So you can do the gui updates.
09:27 tsee I think violating that expectation could yield some carnage.
09:28 tsee But I'd be fine with that, too.
09:28 Alias The current model is completely 1.0, there's going to need to be carnage no matter what
09:28 tsee I agree. Though I'm quite happy that it survived as long.
09:28 tsee Means it wasn't a totally whacko 1.0.
09:28 submersible I'm just thinking in terms of the object being available for events to fire on
09:28 Alias That's because I never found the time to learn how it worked :)
09:28 tsee submersible, yes, certainly.
09:29 tsee Alias, ha!
09:29 tsee submersible is right. Using the mw for events is silly.
09:29 tsee As I said, it was born out of cluelessness.
09:29 Alias Hey, it worked
09:29 submersible exactly.
09:29 Alias And you were only the third person to ever touch the code...
09:29 tsee The one thing that worries me is that Mattia started this branch where he separated out the threading into a separately loadable module,
09:30 tsee If we know switc to using Wx::App all over, we might have to change it again to use the Wx::ThreadEventSomething.
09:30 Alias I can live with that
09:30 tsee IF we want the savings that Mattia's branch promises.
09:30 Alias Wait till Wx 3.0 arrives
09:30 tsee But then again, I don't want to push him too much. He's been very patient with my nagging requests.
09:31 tsee Oh!
09:31 tsee He'll be at YAPC::EU in Pisa.
09:31 tsee Everyone! Buy him beer!
09:36 Alias For example, what if we reliably passed a Padre::Current object to the callback
09:36 Alias That way, if we change the event marshalling location later, we can still ensure we pass in the $current
09:37 tsee You mean pass a Padre::Current to the worker loop and thus Task::run?
09:37 jq Alias: what about doing a file::homedir release?
09:38 tsee The main difficulty is always that you have to keep in mind which information (and objects&classes) are even available in the worker threads at all.
09:39 Alias tsee: I meant more that the location where the callback event fires should be irrelevant
09:40 Alias tsee: Something closer to what we have with Process
09:40 tsee Also, I all fairness, I have to note I'm not up for much hacking today. I'll need to go back to bed soon.
09:40 Alias Perhaps the callback should be a string which identifies a fully resolved named function
09:41 Alias So that you can do very simple tasks/callbacks without needing to know about Wx
09:41 Alias Or something...
09:41 azawawi joined #padre
09:41 azawawi good morning
09:42 Alias Heya
09:42 submersible hi
09:42 tsee Hmm. You still want access to the mw in the post-run hook. Which, of course, will be totally fine via some Padre::Current or whatever proxy.
09:42 Alias azawawi: I tried your new shiny regex thingy
09:42 tsee Because that mustn't ever be in a worker thread anyway.
09:42 Alias azawawi: Have you realised that, after you make the regex, there's no button to actually add the s/foo/bar/ to the current document?
09:42 azawawi Alias: yeah
09:43 Alias You need a kind of "Insert" button
09:43 azawawi Alias: i was actually trying PPIx::Regexp
09:43 Alias ah
09:43 * azawawi adds an insert button
09:45 tsee azawawi, how'd PPIx::Regexp work out?
09:45 tsee It seemed useful enough when I checked it our.
09:45 tsee *out
09:46 tsee All the other regexp parsers were too outdated to handle 5.10 regexps.
09:47 azawawi tsee: seems cool
09:48 Hyppolit svn: r10508 | azawawi++ | http://padre.perlide.org/trac/changeset/10508
09:48 Hyppolit /g modifier is useless (was generating warnings while matching)
09:48 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
09:48 tsee uh oh
09:48 tsee The doc viewer in the plugin manager is broken
09:48 tsee It shows garbled POD
09:48 * azawawi begins a tidal wave of commits :)
09:48 tsee azawawi, where's your regexp stuff?
09:48 tsee Plugi?
09:48 tsee *plugin?
09:49 tsee Oh.
09:49 tsee Another issue: Enabling plugins in the plugin manager doesn't show their menus in the plugins menu.
09:49 tsee I think that only happens on restart.
09:49 tsee Alias, might that latter issue be related to some of your code wrangling?
09:52 tsee Actually. The doc viewer in general is screwed.
09:52 tsee Is that something particular to my system or can anyone reproduce it?
09:52 Alias tsee: Possibly, although more likely it's because of whoever did the refactoring of delaying the refresh of each menu until they are shown
09:52 tsee I've seen it for the first time today. And today's the first time I ran Padre from my new harddrive/linux.
09:53 tsee (related note: SSDs *do* kick ass)
09:53 Alias I'm THIS close to converting
09:53 Alias Next laptop for sure
09:53 tsee I caved in and bought an intel x25-M g2 160gb.
09:53 tsee Large enough to hold a fair chunk of my $work data.
09:54 Hyppolit svn: r10509 | submersible_toaster++ | http://padre.perlide.org/trac/changeset/10509
09:54 Hyppolit safely with $document on editor_enable
09:54 Hyppolit trunk/Padre-Plugin-Swarm/l​ib/Padre/Plugin/Swarm/Wx/
09:54 tsee Computer fires up instantly, even considering it's running all sorts of databases and services.
09:54 tsee Compilation time went down a lot. Linking time even more.
09:54 tsee PADRE STARTUP!
09:54 submersible tsee: I see the same plugins-menu problem
09:55 tsee submersible, thanks.
09:55 tsee "perl dev.pl -a" takes < 2 seconds to show the Padre window. It never manages to display the splash screen at all.
09:55 tsee And the -a used to be a real drag.
09:56 tsee Sorry for the rant. I just had to justify the expense to myself.
09:56 Hyppolit svn: r10510 | azawawi++ | http://padre.perlide.org/trac/changeset/10510
09:56 Hyppolit Added insert regex button (Alias++)
09:56 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
09:56 azawawi tsee: Edit/regex editor
09:56 azawawi tsee: i still didnt add my PPIx::Regexp code yet..
09:57 azawawi Alias__: is it okay to make Padre depend on it?
09:57 Alias Personally, I'm fine with it
09:57 tsee Me, too.
09:57 Alias It's built on the same underpinnings as PPI, so it should have almost entirely the same dependencies
09:57 Alias So it should be, I think, a net dependency increase of only 1
09:57 tsee Edit menu has become WAY too large.
09:57 azawawi i think it adds Readonly when installing...
09:57 tsee Yuck.
09:57 Alias It uses Readonly?
09:58 Alias We can refactor that away if need be
09:58 azawawi i dont remember, it installed one package...
09:58 tsee Depends on the stance of the PPIx::Regexp author.
09:58 Alias I was going to offer him a more "official" PPI::Regexp at some point
09:58 Alias So I can make conditions :)
09:58 tsee Error while autoloading 'Wx::wxRE_READONLY' at /home/tsee/perl/padre/trunk/Padre/​lib/Padre/Wx/Dialog/RegexEditor.pm line 156
09:58 * azawawi double checks
09:58 azawawi my internet is very slow today
09:59 tsee azawawi, also, if you can load existing regexps into the editor, you should add a right-click context-sensitive hook.
09:59 Alias tsee: How much was the SSD?
09:59 tsee Or have somebody else do that for you.
09:59 Hyppolit svn: r10511 | submersible_toaster++ | http://padre.perlide.org/trac/changeset/10511
09:59 Hyppolit switch over to Wx::App instead of main
09:59 Hyppolit trunk/Padre-Plugin-Swarm/lib/Padre/Plugin/
09:59 tsee Alias: 380EUR.
09:59 azawawi http://deps.cpantesters.org/?m​odule=PPIx::Regexp;perl=latest
09:59 azawawi # Readonly...
09:59 tsee A shard of my heart as a PhD student, but I expect it to measurably increase my productivity.
10:00 tsee azawawi, so when you add the dependency, also add a dependency on Readonly::XS because that'll automatically be used if available.
10:00 tsee It does away with some of the pain that Readonly incurs.
10:01 tsee Alias: But then again, I do a lot of "edit C++" && make -j3 install && ./testprogram && repeat cycles.
10:01 Alias Readonly is essentially pointless
10:01 Hyppolit svn: r10512 | azawawi++ | http://padre.perlide.org/trac/changeset/10512
10:01 Hyppolit Depends on PPIx::Regexp to do some magic stuff on the Regex Editor...
10:01 Hyppolit trunk/Padre/
10:01 Alias use constant is better in every possible way than Readonly, other than it doesn't interpolate
10:01 azawawi tsee: will do....
10:01 tsee And in some contexts, it needs disambiguation.
10:01 tsee azawawi, thanks!
10:02 tsee I.e. 2*FOO => 2*FOO()
10:03 Hyppolit svn: r10513 | submersible_toaster++ | http://padre.perlide.org/trac/changeset/10513
10:03 Hyppolit I think this might resolve the issue with catching events from main
10:03 Hyppolit branches/Padre-slave-driver-thread/lib/Padre/
10:03 Hyppolit svn: r10514 | azawawi++ | http://padre.perlide.org/trac/changeset/10514
10:03 Hyppolit Added Readonly::XS to dependency list
10:03 Hyppolit trunk/Padre/
10:03 submersible tsee - try that if you get a moment pls
10:04 Alias Do Readonly variables const-away at compile time?
10:04 Hyppolit svn: r10515 | azawawi++ | http://padre.perlide.org/trac/changeset/10515
10:04 Hyppolit Missing quote from previous commit. azawawi--
10:04 Hyppolit trunk/Padre/
10:05 Alias ... when you use the ::XS version
10:05 tsee submersible, on it!
10:05 tsee Alias: Does it? I didn't know that.
10:05 azawawi tsee: any eta on the branch merge?
10:05 tsee I was just trying to say that use-constant needs disambiguation at times.
10:06 tsee azawawi, no. Well, depends on how submersible's fixes work out.
10:06 tsee It's all pretty experimental still.
10:06 tsee And I have but a half brain.
10:07 Hyppolit svn: r10516 | tsee++ | http://padre.perlide.org/trac/changeset/10516
10:07 Hyppolit Support for loading plugins from dev.pl even from branches
10:07 Hyppolit trunk/Padre/
10:08 chorny joined #padre
10:08 tsee submersible, out of curiosity: What does ::swarm use Graph.pm for?
10:08 submersible :)
10:08 submersible maintaining state
10:08 tsee Usage: $thr->kill('SIG...') at /home/tsee/perl/padre/branches/Padre-slav​e-driver-thread/lib/Padre/TaskManager.pm line 434.
10:08 tsee Not saying this is your fault...
10:08 submersible and hopefuly some really glamorous plots later
10:08 submersible mm
10:09 submersible gah - 86-service is borked too
10:09 submersible oops
10:14 tsee azawawi, any ETA on when I can run the regexp editor?
10:14 Hyppolit svn: r10517 | submersible_toaster++ | http://padre.perlide.org/trac/changeset/10517
10:14 Hyppolit crush some warnigns and errors
10:14 Hyppolit branches/Padre-slave-driver-thread/lib/Padre/ branches/Padre-slave-driver-thread/lib/Padre/Wx/
10:15 submersible warnigns eh?
10:15 chorny Alias: Readonly.pm variables are normal variables with readonly flag (for scalars in XS version) or tied (for arrays and hashes).
10:15 tsee submersible, you probably want to add a Swarm dependency on JSON::XS.
10:16 submersible how did i miss that
10:16 tsee Let me take a wager: You had it installed already?
10:16 submersible clearly - but I should also know better
10:17 Hyppolit svn: r10518 | azawawi++ | http://padre.perlide.org/trac/changeset/10518
10:17 Hyppolit Added regex description UI
10:17 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
10:17 azawawi tsee: using PPIx::Regexp?
10:17 tsee azawawi, either way.
10:18 azawawi tsee: it is working now
10:18 tsee Not for me.
10:18 azawawi tsee: but not with PPIx::Regexp
10:18 tsee azawawi, Error while autoloading 'Wx::wxRE_READONLY' at /home/tsee/perl/padre/trunk/Padre/​lib/Padre/Wx/Dialog/RegexEditor.pm line 156
10:18 tsee Wx 0.96
10:18 Hyppolit svn: r10519 | azawawi++ | http://padre.perlide.org/trac/changeset/10519
10:18 Hyppolit Perl tidy
10:18 Hyppolit trunk/Padre/ trunk/Padre/lib/Padre/Wx/Dialog/
10:19 tsee submersible, t/86-service.t ............. 1/? Padre->new has not been called yet at /home/tsee/perl/padre/branches/Padre-sla​ve-driver-thread/blib/lib/Padre/Task.pm line 535
10:19 tsee # Tests were run but no plan was declared and done_testing() was not seen. t/86-service.t ............. Dubious, test returned 255 (wstat 65280, 0xff00) All 9 subtests passed
10:19 azawawi tsee: i see
10:19 tsee And of course, the now classic t/85-task-manager.t ........ 7/18 Use of uninitialized value in null operation during global destruction.
10:19 tsee submersible, other than that, it seems to be working.
10:19 tsee I could even load ::Swarm
10:19 azawawi tsee: im using Wx 0.94 and it is working fine... but that's on vista
10:19 submersible what did u see?
10:19 azawawi tsee: let me upgrade Wx
10:20 submersible srry - afk for some minutes
10:20 tsee submersible, from swarm? The chat window.
10:20 tsee submersible, does ::Swarm allow remote editing yet? Not in the fine-grained sense but in the "give me control over your document NOW" sense?
10:21 tsee oops, well. See you in a couple of minutes.
10:21 azawawi tsee: maybe Wx::Richtextctrl is not supported under linux?
10:21 tsee That would be strange.
10:21 tsee If in doubt, we can bring it up on the wx users listl.
10:21 tsee I'm installing Wx 0.94 for testing.
10:22 azawawi try removing Wx::wxRE_READONLY
10:22 tsee Let me try the older Wx first.
10:22 azawawi tsee: okay
10:24 Alias chorny: Do variables curry away when they have the readonly flag on?
10:24 Alias That's all I really care about
10:25 * Alias is going to remove Test::ClassAPI from PPI to compensate for the addition of PPIx::Regexp
10:26 azawawi tsee: worked with Wx 0.96
10:26 azawawi tsee: Maybe richtextctrl doesnt compile for Linux?
10:26 * azawawi uninstalls Readonly
10:26 tsee Alias: I don't think so.
10:27 tsee azawawi, okay, similarly 0.94 didn't help.
10:27 tsee I'll lose the READONLY.
10:27 Hyppolit svn: r10520 | azawawi++ | http://padre.perlide.org/trac/changeset/10520
10:27 Hyppolit More comments
10:27 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
10:28 tsee Error while autoloading 'Wx::wxRE_MULTILINE' at /home/tsee/perl/padre/trunk/Padre/​lib/Padre/Wx/Dialog/RegexEditor.pm line 156
10:29 chorny Alias: no
10:29 azawawi tsee: drop it also
10:29 tsee Can't locate object method "new" via package "Wx::RichTextCtrl" (perhaps you forgot to load "Wx::RichTextCtrl"?) at /home/tsee/perl/padre/trunk/Padre/​lib/Padre/Wx/Dialog/RegexEditor.pm line 156.
10:29 azawawi tsee: aha
10:29 tsee So richtext might be dead.
10:29 tsee I'll check out the wx build.
10:29 azawawi http://docs.wxwidgets.org/stable​/wx_wxrichtextctrloverview.html
10:31 tsee azawawi, it was built: -rw-r--r-- 1 tsee tsee 3390600 2010-02-05 11:30 RichText.o
10:31 tsee in ext/richtext
10:32 tsee Also passes tests.
10:32 azawawi tsee: about Readonly::XS, i uninstalled Readonly to make sure we can live without it
10:32 azawawi tsee: it breaks PPIx::Regexp
10:35 azawawi tsee: ah i see, i thought Readonly::XS was a replacement for Readonly... ignore my ignorance :)
10:36 chorny Internal::readonly (or somethig like that) is a good replacement for Readonly.
10:36 azawawi chorny: hi there :)
10:36 tsee chorny, InternalS, I think
10:36 chorny azawawi: hi
10:42 tsee azawawi, strange thing it's working for you at all.
10:42 tsee Ah, nevermind. Found something I'd been missing
10:42 azawawi tsee: what is it?
10:43 tsee azawawi, nevermind. I was looking at the XS++ code for richtextctrl
10:43 tsee Can anyone on linux test the regexp editor, please?
10:43 tsee I may have a problem because I have my hacked ExtUtils::XSpp installed which does exception handling.
10:43 tsee Seems unlikely, but maybe that's breaking things.
10:43 tsee Man, I hope not.
10:44 * azawawi powers up his old ubuntu vm
10:46 tsee Aha, I might have found the culprit.
10:46 azawawi tsee: this will take time on my vm... since i havent updated it for a while
10:46 tsee azawawi, I'm looking into the XS while you're testing.
10:46 tsee I appreciate that.
10:46 tsee I was just about to cut an EU::XSpp release.
10:46 tsee Would have sucked to let a bad one escape.
10:49 submersible tsee: not really remote editing yet.
10:49 submersible tsee: if you don't see any output in the chat window. it's not working.
10:49 tsee I do
10:50 tsee "blablah created chat session" or something
10:50 submersible data is just being spewed out on a multicast address. so if u were running padre on the same 'network' other editors will see that chatter
10:50 submersible beware
10:51 tsee I don't think multicast crosses my router.
10:51 submersible the ttl is 1 so they should not
10:51 tsee MAN, Shlomi. Get some decency.
10:52 submersible but the chat window is an event drone. so if you see output when you type in there.. it;'s coming via the network
10:55 submersible tsee: I suspect  that the null oper in global destruction is a db handle
10:56 tsee submersible, ah. Interesting thought.
10:56 tsee That may be dangerous, really.
10:57 submersible if there is a db conn open during thread spawning maybe.. I'm not certain
11:02 Hyppolit svn: r10521 | submersible_toaster++ | http://padre.perlide.org/trac/changeset/10521
11:02 Hyppolit update depends JSON::Xs
11:02 Hyppolit trunk/Padre-Plugin-Swarm/
11:04 tsee azawawi, I found a bug in the new EU::XSpp for sure. I'm testing now whether the bug fix also fixes RichTextCtrl.
11:04 tsee Code generation is such a fragile best!
11:04 tsee *beast, damnit
11:08 azawawi tsee: nearly finished testing, had to pull like 20 packages
11:08 tsee azawawi, okay, my "fixed" xspp still fails on the Richtext.
11:08 tsee I'm pretty sure it's that and not a platform thing.
11:09 tsee But whittling out those bugs isn't going to be fun.
11:09 tsee OTOH it would be sad to see the improvements to EU::XSpp go to waste.
11:09 tsee You can now throw C++ exceptions and have them auto-converted to perl-croaks.
11:10 azawawi tsee: bug confirmed on ubuntu...
11:10 tsee Aha!
11:10 tsee Surprising and unsettling.
11:11 azawawi tsee: same as your issue
11:14 azawawi if it does not work, we can always switchback to HtmlWindow
11:14 azawawi since that's what im using it for... readonly/multiline/multi-style...
11:17 tsee Doh. Wx trunk might have to do with some of the other issues I'm seeing.
11:18 tsee SSD++ # paying of as I build the tenth Wx.pm
11:23 tsee Okay, somehow, I sincerely nuked my setup.
11:23 tsee Does this ring a bell to anyone? unable to resolve overloaded method for Wx::ListCtrl::SetItem at /home/tsee/perl/padre/branches/Padre-slave​-driver-thread/lib/Padre/Action/Plugins.pm line 41
11:23 tsee I guess not.
11:25 submersible joined #padre
11:25 azawawi tsee: what do u think? keep using richtext control or simply use htmlwindow?
11:26 tsee azawawi, I don't know yet. I'll have to repair my ... whole perl environment before I can comment.
11:28 tsee Okay. rm-rf .padre helped a bit
11:28 tsee submersible, there's a Swarm bug.
11:28 tsee If you load the plugin without any previous bottom pane open, it corrupts the GUI
11:29 tsee azawawi, I can now get at the regexp editor
11:29 tsee Using Wx 0.96 (not trunk) and EU::XSpp 0.07 (not trunk)
11:30 azawawi so EU::XSpp is the culprit?
11:30 azawawi tsee: PPIx::Regexp fails gracefully a bit on some captures
11:30 tsee azawawi, nah. strike that.
11:31 tsee I was on the slave driver branch
11:31 tsee Which is behind trunk.
11:31 tsee Padre trunk is still crashing with the usual error.
11:31 tsee It seems EU::XSpp and the richtext problem aren't related after all.
11:33 submersible tsee: there's many bugs in there !
11:33 tsee submersible, just saying.
11:33 tsee Might not even be a problem specific to ::Swarm.
11:33 submersible :) thanks - I have been going on a round of ->Hide
11:35 submersible aha - yep awesome bug
11:38 Hyppolit svn: r10522 | submersible_toaster++ | http://padre.perlide.org/trac/changeset/10522
11:38 Hyppolit hide Resources panel AFTER sizing.
11:38 Hyppolit trunk/Padre-Plugin-Swarm/l​ib/Padre/Plugin/Swarm/Wx/
11:39 * azawawi converts stuff to use Padre::Wx::HtmlWindow
11:40 tsee azawawi, I currently see no other way of fixing the issue.
11:40 tsee BUT
11:40 tsee I think you should bring this up on wxperl-users.
11:41 tsee I'd do it, but I really have no idea what you're doing with it.
11:41 tsee Mattia's been pretty active recently. You're likely to get a swift response.
11:41 azawawi yeah i noticed... the RT emails just keep coming
11:44 Hyppolit svn: r10523 | azawawi++ | http://padre.perlide.org/trac/changeset/10523
11:44 Hyppolit Converting 1st richtext control code to use Padre::Wx::HtmlWindow instead since it is not working on Linux
11:44 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
11:48 Hyppolit svn: r10524 | azawawi++ | http://padre.perlide.org/trac/changeset/10524
11:48 Hyppolit Removing all traces of Wx::RichtextCtrl ...
11:48 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
11:48 * submersible plays evil with EVT_IDLE  MUahahahahaAHAH
11:48 tsee EVT_IDLE?
11:49 azawawi tsee: please it try it now... it seems to work on my VM now :)
11:50 tsee Oh fuc.
11:50 tsee git-svn--
11:50 tsee It just made me commit to a fucking tag.
11:51 tsee azawawi, doesn't crash!
11:51 tsee But you might want to go easy with the vertical screen estate.
11:51 tsee It fills a good 800 pixels.
11:52 tsee Which is the total height of my display.
11:52 tsee Whereas I understand 768 or some such is quite common for the 16:9 laptops.
11:54 Hyppolit svn: r10525 | azawawi++ | http://padre.perlide.org/trac/changeset/10525
11:54 Hyppolit Perl tidy, Set the focus on the regex text field, and More accelerators
11:54 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
11:56 azawawi tsee: so wider instead of taller right for 16:9 screens?
11:56 tsee azawawi, yep.
11:57 azawawi tsee: it was good until i added the non-functioning description field (which is supposed to say in english what the regular expression means)
11:57 azawawi tsee: that's where PPIx::Regexp  is going to be used
11:57 tsee Maybe make that a pop up or something?
11:57 tsee A button?
12:05 Hyppolit svn: r10526 | azawawi++ | http://padre.perlide.org/trac/changeset/10526
12:05 Hyppolit Nuking the use CreateStdDialogButtonSizer in "Quick Menu Access" dialog since it is not multi-lingual
12:05 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
12:05 azawawi tsee: could be... the UI is not final yet... im experimenting :)
12:05 tsee azawawi, just an idea.
12:06 azawawi tsee: thx
12:06 tsee azawawi, is it okay if I experiment with the right-click context hook for the editor?
12:06 azawawi tsee: no problem
12:06 azawawi tsee: go ahead :)
12:06 tsee "looks like regexp" will be much more difficult than "looks like variable" heuristics.
12:07 tsee Maybe it won't work out well.
12:11 Hyppolit svn: r10527 | azawawi++ | http://padre.perlide.org/trac/changeset/10527
12:11 Hyppolit Removed all uses of CreateStdDialogButtonSizer in Padre dialogs since it is not multi-lingual
12:11 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
12:14 Hyppolit svn: r10528 | azawawi++ | http://padre.perlide.org/trac/changeset/10528
12:14 Hyppolit Not a Perl tidy :)
12:14 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
12:16 * azawawi lunch &
12:28 tsee Alias, as the great PPI author, do you have any ideas how one could write a quick-and-dirty piece of code that guesses whether the cursor is inside a regular expression?
12:28 tsee Without PPI, of course.
12:29 Alias As it so happens, I promised szabgab I'd write that exact function
12:30 Alias An I have about 20-50 lines of code in my head that should be able to do it, half formed
12:30 Alias And...
12:31 Alias Or at least, I have a solution that holds for patterns that exist on a single line
12:31 Alias Like finding the class name the cursor is over
12:31 tsee Well. Single line ones would be reasonably feasible.
12:32 tsee Though usually, you can generalize by matching the whole document and using /s
12:32 Alias And there's another alternate method that should work pretty well for regexp specifically
12:32 tsee Can it do "qr fooof"?
12:32 tsee Where the f's are the delims?
12:32 tsee Matching quotelikes in perl is the worst thing ever.
12:32 Alias The regexp one would introspec the lexer
12:32 Alias introspect
12:33 tsee "the lexer"? I'm talking about a non-PPI (i.e. fast) solution.
12:33 tsee For the right-click context menu.
12:33 Alias I'm talking about the scintilla lexer
12:33 tsee Aha!
12:33 Alias You should be able to find the lexical context at the cursor
12:33 tsee Can we call that even if the user chose to use a different lexer?
12:33 Alias No
12:34 tsee Damn.-
12:34 Alias But then scintilla is the only lexer that doesn't suck
12:34 tsee Still smells like a problem waiting to happen.
12:34 Alias It might be able to work based on the colour type, not the processing engine that produces them
12:35 Alias Examine the logical colour type in existance at the point in state, not ask the lexer to process anything
12:35 Alias Which would make it instant, if it works
12:36 Alias That would locate you the really exotic stuff, but not the simple stuff like class names that fall into "word" and so on
12:36 tsee What if you reuse the same color?
12:37 tsee Or is it more abstract than that?
12:37 Alias lexical colour, not literal colour
12:37 Alias And for values that are common/indeterminate "is this a class name" that should hopefully fit into the sub-line thing
12:37 Alias This is all still half-formed of course, and hasn't survived contact with actual code
12:43 tsee Alias: Okay, since you have this laid out in your head, I'll punt on the context menu for the regexps for now.
12:43 Alias I'm more likely to get the "Is the cursor inside a class name" stuff first
12:44 tsee Because just doing a half-decent job would be hard and I'd hate to do a 1.0 implementation that takes quite some effort only to be replaced by something better soon.
12:44 tsee Well, that's moderately easy for the simple cases. At least compared to regexes.
12:44 Alias Well, different methods
12:44 Alias And compared to introspecting the lexer, just getting the current line string and parsing it is relatively easy
12:45 tsee Well, parsing it wrt. regexes and quotelikes is all but easy.
12:48 Alias PPI is pretty fast over very short distances
12:48 Alias Under, say, 10 lines of code
12:48 Alias So there's options there too
12:49 tsee How well does it handle starting the parse at arbitrary points in the code?
12:49 Alias badly
12:49 tsee Using a regex-based parser to determine statement boundaries is difficult, too.
12:49 tsee That's what I though.
12:49 Alias Well, that's not entirely true
12:49 tsee *thought
12:49 Alias It CAN parse line noise
12:50 tsee yeah, well, it spits out "wrong" results, then.
12:50 Alias So it would certainly produce SOME document
12:50 Alias right
12:50 tsee Not helpful.
12:50 Alias So you'd really need to find a kind of "keyframe"
12:50 Alias A known safe point
12:50 tsee That would be a great feature in general.
12:50 Alias It would only need to be safe with regards to quotes and quotelikes
12:50 submersible what is going on w/ padre::wx::dialog
12:50 tsee "find the next keyframe before the start of my currently visible piece of code"
12:51 Alias Because it can tokenenize starting at an arbitrary nested depth
12:51 tsee "find the the next keyframe after" => PPI highlighting
12:51 Alias You might be able to use the lexer for that
12:52 chorny Is there any todo list with near future plans for Padre?
12:53 Alias Officially, there is no roadmap
12:53 Alias We prefer itch scratching
12:53 Alias Find something about Padre you personally wish sucked less or worked better
12:54 Alias Most of the things that "need doing" and haven't been done yet are enormous or vague
12:55 Alias chorny: Anything in particular you'd like to see?
12:56 Alias To give you an example, I hate that syntax checking of unsaved changes to modules doesn't work
12:56 Alias Because it uses a temp file for A.pm, which loads B.pm, which loads A.pm, which results in duplicate subroutine errors
12:57 Alias So I'm working on a new way to marshall temp files and tweak PERL5LIB so that problem goes away
12:59 chorny Alias: I want to implement task management system ( todo list :)) plugin for Padre, but I found out that working with WxPerl is very hard. So I'm trying to work with existing code first.
12:59 Alias GUI work is vicious
13:00 Alias And MOST of Padre's gui code kinda sucks
13:02 azawawi Alias: hey Padre's gui code rocks :)
13:02 Alias *cough*Preferences*cough*
13:02 azawawi hey hey
13:02 Alias I've got some on-again, off-again experiments to try to auto-generate sane dialog code
13:03 azawawi i could transform it to an eclipse-like thing if we want to dump that dependency thing
13:03 Alias You'd what a what now?
13:03 azawawi searchable preferences tree like in eclipse
13:04 azawawi like in directory browser but for preferences
13:05 Alias Ah
13:05 Alias Still prefer a more traditional approach myself
13:05 Alias More of a Firefox approach
13:33 tlb joined #padre
13:43 Hyppolit #844: add wysiwyg-ish pod editing capabilities (new enhancement) [ http://padre.perlide.org/trac/ticket/844 ]
13:54 Hyppolit svn: r10529 | azawawi++ | http://padre.perlide.org/trac/changeset/10529
13:54 Hyppolit First implementation of the regular expression description text field (using PPIx::Regexp). Nothing fancy yet.
13:54 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
13:54 cognominal joined #padre
13:55 Hyppolit svn: r10530 | azawawi++ | http://padre.perlide.org/trac/changeset/10530
13:55 Hyppolit Tidy the untidied Perl code
13:55 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
14:00 basti1985 joined #padre
14:00 basti1985 left #padre
14:03 basti19851 joined #padre
14:03 basti19851 left #padre
14:09 Hyppolit svn: r10531 | azawawi++ | http://padre.perlide.org/trac/changeset/10531
14:09 Hyppolit Regex help comments were not showing up on Linux (since it seems to validate label\tdescription, win32 doesnt)
14:09 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
14:13 aukjan_ joined #padre
14:14 asphere joined #padre
14:19 asphere joined #padre
14:28 azawawi tsee: ping
14:29 tsee pong
14:29 azawawi tsee: i thought about regex highlighting while typing
14:29 tsee ouch
14:29 azawawi tsee: if we do not have a functional richtextctrl...
14:30 tsee You should really send a mail to the list about that.
14:30 azawawi yeah
14:30 azawawi but im sleepy a bit now. Could you do it for me? :)
14:31 tsee No. Because I have NO idea whatsoever about that class.
14:31 tsee Sorry.
14:31 azawawi tsee: np
14:31 tsee My mail would be like "Umm, not worky, okay?"
14:32 azawawi np then i added it my todo-pile :)
14:32 azawawi could you try the latest stuff to check if it is working or not?
14:33 tsee phone.
14:36 tsee azawawi, the description echos the pattern itself
14:36 tsee Aside from the ppi dump
14:37 tsee Ah, sorry. I think it does it only if it contains newlines or so.
14:38 tsee Otherwise, it seems to work fine.
14:41 azawawi yeah it is a first implementation, next step would be the english translation...
14:42 azawawi tsee: thanks for testing it
14:45 Hyppolit svn: r10532 | azawawi++ | http://padre.perlide.org/trac/changeset/10532
14:45 Hyppolit Showing match errors and warnings for regex matches
14:45 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
14:48 azawawi tsee: any progress on the "Edit with Regex Editor" right-click option?
14:48 tsee azawawi, scrapped for now.
14:48 tsee Blocking on Adam implementing "Is this a regexp" function.
14:48 tsee Then, it's trivial.
14:48 tsee Just adding a few simple, lines to Padre::Document::Perl.
14:49 tsee I can't type today. That comma was suppose to go before "simple"
14:50 azawawi *: please send if you have any ideas that we can squeeze into the regex editor...
14:52 Coke is this channel logged?
14:52 Hyppolit svn: r10533 | azawawi++ | http://padre.perlide.org/trac/changeset/10533
14:52 Hyppolit Updated Changes
14:52 Hyppolit trunk/Padre/
14:52 azawawi Coke: yeah ofcourse
14:52 azawawi Coke: irclog.perlgeek.de/padre/today
14:54 Coke bah, didn't even think to check moritz's site. danke.
14:54 * azawawi sleep... have a nice weekend &
14:54 azawawi Coke: np
14:54 Coke azawawi: wanted to point the Readonly author at the chat earlier. =-)
14:55 azawawi Coke: yes?
14:58 Coke nothing in particular.
14:59 tsee Coke, well. I somehow doubt Readonly can be made to do what we expect from a constant: have the optimizer replace it during compile time.
15:00 tsee I doubt Eric could do anything about that.
15:00 Coke eh. he's clever and shifty.
15:00 Coke mostly shifty, mind you.
15:01 Coke besides, there's a kibo-like satisfaction in knowing people are talking about things you've done.
15:01 Hyppolit svn: r10534 | submersible_toaster++ | http://padre.perlide.org/trac/changeset/10534
15:01 Hyppolit add preferences via wxglade codegen, minor tweaks
15:01 Hyppolit trunk/Padre-Plugin-Swarm/ trunk/Padre-Plugin-Swarm/junkyard/ trunk/Padre-Plugin-Swarm/lib/Padre/Plugin/ trunk/Padre-Plugin-Swarm/l​ib/Padre/Plugin/Swarm/Wx/
15:13 dapatrick joined #padre
15:33 rindolf joined #padre
15:33 rindolf Hi all.
15:33 rindolf jq: can we talk?
15:33 jq rindolf: sure
15:34 rindolf jq: OK.
15:34 rindolf jq: I was insulted by your last E-mail.
15:34 jq i was insulted by yours
15:34 rindolf jq: I'm not getting paid to publicise Padre on Freshmeat.net either.
15:34 jq i know
15:34 rindolf And you should mention the changes on the releases.
15:35 rindolf I was giving you good advice in the first E-mail.
15:35 jq from the announce: "What can be said about Padre 0.56 is that it is faster than it has been in a long time. "
15:35 rindolf jq: that's not good enough.
15:36 rindolf jq: and it's announcement, not announce.
15:36 jq rindolf: the announce had a link to alias blog
15:36 jq http://use.perl.org/~Alias/journal/40133
15:36 jq if you want to improve the release announce, fine
15:36 jq but the way you *demanded* things was not ok
15:36 jq you can suggest
15:37 jq pete is not paid to do the release either
15:37 jq and if i have to choose between having a release or having it publicized on freshmeat... my choice is done very quickly
15:37 rindolf That's not exactly a good "What's new"
15:37 jq because if there is no release, there's no point announcing it on freshmeat
15:37 rindolf jq: this is a false dichotomy.
15:37 jq well, i can totally imagine pete being offended by you
15:38 jq and quitting in his release manager job
15:38 jq so the points *are* related
15:38 rindolf http://en.wikipedia.org/wiki/False_dilemma
15:38 jq now, i know that you did not want to be rude on purpose
15:38 jq but the fact is, you were
15:38 rindolf jq: I wrote one E-mail which was perfectly polite and constructive and did not get a reply.
15:39 rindolf And someone once told me that when people are doing gratis work on an open-source project, he expects them to do a good job regardless of the fact that they are getting paid.
15:39 rindolf And someone once told me that when people are doing gratis work on an open-source project, he expects them to do a good job regardless of the fact that they are not getting paid.
15:39 jq sure, we take pride in what we do
15:39 Alias jq: It doesn't matter if he quits as release manager
15:39 Alias (just to chime in)
15:39 jq Alias: i think it does
15:40 dapatrick joined #padre
15:40 jq because it means that we loose a contributor
15:40 Alias jq: The release manager position is intended to move around
15:40 rindolf jq: s/loose/lose
15:40 jq since pete is only doing releases currently
15:40 Alias But can we please not bicker about how successful we are?
15:40 Alias I mean, I LIKE that we are having these problems, but still
15:42 * tsee smirks and takes an aspirin.
15:51 Getty dudes... chill! :)
16:08 awnstudio joined #padre
16:37 [1]awnstudio joined #padre
16:38 [2]awnstudio joined #padre
16:43 CromeDome joined #padre
16:58 awnstudio joined #padre
17:12 isec joined #padre
17:27 isec_ joined #padre
17:30 azawawi joined #padre
17:33 isec joined #padre
17:33 tsee Grmpf. I can't paste into Padre with middle-mouse click. :(
17:39 isec_ joined #padre
17:43 isec joined #padre
17:52 cognominal joined #padre
18:01 Sewi tsee: There is a preference option which might help you
18:02 tsee Sewi, cool, thanks.
18:35 awnstudio joined #padre
20:02 kgish joined #padre
20:18 basti1985 joined #padre
20:18 basti1985 left #padre
20:28 rindolf joined #padre
21:51 basti1985 joined #padre
21:54 basti1985 left #padre
22:22 basti19851 joined #padre
22:39 basti19851 left #padre
22:40 waxhead joined #padre
22:40 waxhead morning
23:08 Hyppolit svn: r10535 | chorny++ | http://padre.perlide.org/trac/changeset/10535
23:08 Hyppolit Add more 'Extended Patterns' returned by sub _regex_groups
23:08 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
23:26 basti1985 joined #padre
23:27 basti1985 left #padre
23:48 basti1985 joined #padre
23:49 kthakore joined #padre
23:49 kthakore Alias: or Alias__ around?
23:49 basti1985 left #padre
23:53 patspam joined #padre

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