Camelia, the Perl 6 bug

IRC log for #padre, 2010-01-26

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

All times shown according to UTC.

Time Nick Message
00:00 mj41_ joined #padre
00:50 patspam joined #padre
01:42 asphere joined #padre
01:47 submersible joined #padre
01:56 cognominal joined #padre
03:11 jazzanova joined #padre
03:44 Alias_ joined #padre
03:52 cognominal joined #padre
03:52 waxhead_ joined #padre
03:56 waxhead__ joined #padre
06:28 Sewi joined #padre
06:28 Sewi gm
06:49 cognominal joined #padre
07:11 kaare joined #padre
07:32 danlucraft joined #padre
07:34 azawawi joined #padre
07:34 azawawi good morning
07:37 Sewi Hi azawawi
07:38 azawawi hi there :)
07:39 azawawi closing regex editor a second time (using ESC key) on win32 seems to crash Padre...
07:51 asphere joined #padre
07:52 Alias_ joined #padre
07:56 * Alias_ decides to migate ORLite to Template::Tiny
07:56 Alias_ (and hope for the best)
08:41 pece joined #padre
08:58 pece2 joined #padre
09:04 Hyppolit svn: r10286 | azawawi++ | http://padre.perlide.org/trac/changeset/10286
09:04 Hyppolit Bye bye Acme-CPANAuthors-Padre (Removed for maintenance issues, adamk++)
09:04 Hyppolit trunk/
09:04 azawawi joined #padre
09:20 marcela joined #padre
10:11 submersible joined #padre
10:30 Hyppolit svn: r10287 | azawawi++ | http://padre.perlide.org/trac/changeset/10287
10:30 Hyppolit Fixed Padre no-document crash with Find Next/Find Previous functionality.
10:30 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
10:30 Hyppolit svn: r10288 | azawawi++ | http://padre.perlide.org/trac/changeset/10288
10:30 Hyppolit Updated Changes
10:30 Hyppolit trunk/Padre/
10:40 asphere joined #padre
10:48 Hyppolit svn: r10289 | azawawi++ | http://padre.perlide.org/trac/changeset/10289
10:48 Hyppolit Added "Use Splash Screen?" to Padre Preferences dialog (adamk++)
10:48 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
10:59 Hyppolit svn: r10290 | azawawi++ | http://padre.perlide.org/trac/changeset/10290
10:59 Hyppolit Fixed uninitialized warnings when Padre starts up
10:59 Hyppolit trunk/Padre/lib/Padre/
11:46 cognominal joined #padre
11:52 daxim joined #padre
12:55 azawawi joined #padre
12:58 Hyppolit svn: r10291 | azawawi++ | http://padre.perlide.org/trac/changeset/10291
12:58 Hyppolit In Open resource, PGDN (or DOWN) transfers focus to the list
12:58 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
12:59 Hyppolit svn: r10292 | azawawi++ | http://padre.perlide.org/trac/changeset/10292
12:59 Hyppolit Perl tidy
12:59 Hyppolit trunk/Padre/lib/Padre/
13:03 azawawi Alias_: ping
13:03 azawawi ping *
13:03 Alias_ pong
13:04 azawawi Any reason why we have Padre::Wx::RightClick ?
13:04 azawawi It is not being used...
13:05 Alias_ Did I write it?
13:05 Alias_ (It wouldn't be out of the question)
13:06 * azawawi takes a look at the log
13:06 azawawi r384 szabgab
13:06 Hyppolit Changeset #384 http://padre.perlide.org/trac/changeset/384
13:07 azawawi seems to be it was started to refactor right click and then it was left there empty...
13:08 * azawawi removing...
13:08 Alias_ Wouldn't be unusual
13:08 Alias_ I've got a few stub classes in there
13:10 Hyppolit svn: r10293 | azawawi++ | http://padre.perlide.org/trac/changeset/10293
13:10 Hyppolit Removed unused Padre::Wx::RightClick
13:10 Hyppolit trunk/Padre/lib/ trunk/Padre/lib/Padre/Wx/
13:31 * Alias_ is doing some rewriting of ORLite to make it faster
13:31 Alias_ I think I can halve the cost of ->select
13:41 Hyppolit svn: r10294 | azawawi++ | http://padre.perlide.org/trac/changeset/10294
13:41 Hyppolit Make sure that windows context key shows the refactor menu items in the right-click pop-up menu
13:41 Hyppolit trunk/Padre/ trunk/Padre/lib/Padre/Document/
13:42 azawawi Alias_: go for it :)
13:52 Hyppolit svn: r10295 | azawawi++ | http://padre.perlide.org/trac/changeset/10295
13:52 Hyppolit Prevented another no-document Padre crash in Preferences "Guess from current document" button
13:52 Hyppolit trunk/Padre/lib/Padre/Wx/Dialog/
13:53 kthakore joined #padre
13:55 kthakore hi
13:55 azawawi hi kthakore
13:55 kthakore what prevented split screen from working?
13:55 kthakore hi azawawi
13:55 * azawawi takes a look at the svn logs
13:57 azawawi donnu :)
13:57 Hyppolit svn: r10296 | kthakore++ | http://padre.perlide.org/trac/changeset/10296
13:57 Hyppolit Only watch if filename is provided
13:57 Hyppolit trunk/Padre/lib/Padre/Wx/
13:58 Alias_ Hrm
13:58 Alias_ There's a big bug in the Syntax Checker
13:58 * Alias_ plays around with the run command stuff
13:58 kthakore azawawi: ok I will try to do it later
13:58 kthakore who was working on the Main.pm doc watcher?
13:59 kthakore I had to add http://padre.perlide.org/trac/changeset/10296
13:59 kthakore to prevent undefined warning on $fn
13:59 azawawi yeah it was annoying...
13:59 azawawi i think it was either adamk or sewi
13:59 kthakore azawawi: yup
13:59 kthakore ok
14:00 kthakore Alias_ or Sewi I messed with you code http://padre.perlide.org/trac/changeset/10296
14:00 kthakore Alias_ or Sewi your warnings are annoying when $fn is undef
14:02 Alias_ I haven't been involved in coding the file watcher
14:04 Hyppolit svn: r10297 | azawawi++ | http://padre.perlide.org/trac/changeset/10297
14:04 Hyppolit Fixed a bug introduced by my r10290 fix. Sorry :)
14:04 Hyppolit trunk/Padre/lib/Padre/
14:05 azawawi kthakore: correction, that would be Corion r10268
14:05 Hyppolit Changeset #10268 http://padre.perlide.org/trac/changeset/10268
14:14 kthakore azawawi: ok
14:15 Alias_ Whoa
14:15 Alias_ What's he mucking with in the single instance server
14:15 Alias_ That's delicate code
14:16 Alias_ He made some other somewhat iffy changes
14:20 azawawi who is Corion btw? :)
14:21 Alias_ http://search.cpan.org/~corion/
14:22 * azawawi slaps himself
14:23 Alias_ Damn, he's made the single instance server need to support long term connections
14:23 Alias_ I'm fairly sure I built it in a way that doesn't support that
14:24 Alias_ It only supports one connection at a time
14:24 Alias_ Which was supposed to be fine, because it would only be connected for tenths of a second
14:44 mj41__ joined #padre
14:56 mj41_ joined #padre
15:29 dapatrick joined #padre
15:49 Hyppolit svn: r10298 | azawawi++ | http://padre.perlide.org/trac/changeset/10298
15:49 Hyppolit Used Module::CoreList::is_deprecated to display deprecated CORE packages in help search title
15:49 Hyppolit trunk/Padre/lib/Padre/Document/Perl/
15:50 Hyppolit svn: r10299 | azawawi++ | http://padre.perlide.org/trac/changeset/10299
15:50 Hyppolit Updated Changes
15:50 Hyppolit trunk/Padre/
15:51 Hyppolit svn: r10300 | azawawi++ | http://padre.perlide.org/trac/changeset/10300
15:51 Hyppolit el-Perlo Tidy-o :)
15:51 Hyppolit trunk/Padre/lib/Padre/ trunk/Padre/lib/Padre/Wx/
15:52 azawawi lol, i got the r10300 :)
15:52 Hyppolit Changeset #10300 http://padre.perlide.org/trac/changeset/10300
15:52 tsee joined #padre
15:57 cognominal joined #padre
15:58 kaare joined #padre
17:45 danlucraft joined #padre
18:32 basti1985 joined #padre
18:33 basti1985 left #padre
18:41 tsee joined #padre
19:11 SvenDowideit_ joined #padre
19:12 SvenDowideit__ joined #padre
19:13 danlucraft joined #padre
19:40 reneeb joined #padre
19:51 SvenDowideit__ joined #padre
20:20 Alias joined #padre
20:32 LeK joined #padre
20:41 LeK would you be interested in a bugreport like "hitting step over button" sometimes open new untitledX tab" ?
20:41 Alias Sure
20:41 Alias Can you elaborate on "sometimes" ? :)
20:43 LeK at first hit, or something every four or five lines.. i haven't tried to check if it was on something like "step over a method call located in another file", but i can try to clarify "sometimes" :)
20:43 tsee Has anyone had a chance to try the Padre-slave-driver branch?
20:45 Alias Not yet
20:45 LeK it looks like both these lines generates the tab opening:   binmode STDERR, ":utf8";      and    my $verbosity=0;
20:46 tsee Doesn't seem like they have anything in common.
20:51 tsee Alias, all: That'll require testing by people other than myself. That's why I'm being so pushy about it.
20:51 tsee submersible, particularly from you wrt. Services.
20:51 Alias ok, will test
20:51 tsee Thanks.
20:52 tsee Is it a joke the the Debugging menu entry is spelled Debuggin?
20:53 Alias There is no Debugging menu
20:53 Alias It's just Debug
20:53 Alias It is in German?
20:53 tsee Ah. Doh.
20:54 tsee Indeed.
20:54 tsee Use of uninitialized value $fn in hash element at /home/tsee/perl/padre/branches/Padre-sl​ave-driver-thread/lib/Padre/Wx/Main.pm line 4147.
20:54 tsee Is that a known warning or is that new with my branch?
20:54 Alias Someone blamed Corion for that and fixed it on trunk
20:58 Hyppolit svn: r10301 | tsee++ | http://padre.perlide.org/trac/changeset/10301
20:58 Hyppolit merge trunk 10285:10300 to branch
20:58 Hyppolit branches/Padre-slave-driver-thread/ branches/Padre-slave-driver-thread/lib/ branches/Padre-slave-driver-thread/lib/Padre/ branches/Padre-slave-driver​-thread/lib/Padre/Document/ branches/Padre-slave-driver-th​read/lib/Padre/Document/Perl/ branches/Padre-slave-driver-thread/lib/Padre/Wx/ branches/Padre-slave-driver-​thread/lib/Padre/Wx/Dialog/
20:59 tsee svn is such a shitty VCS.
21:09 reneeb Hi *, is there a Padre "marketing" text? I prepare a brochure for this years' CeBIT. A "very first" draft is available at http://renee-baecker.de/sonstiges/CeBIT.pdf . If no such text exists, can the Padre community write such a text?
21:16 Hyppolit svn: r10302 | tsee++ | http://padre.perlide.org/trac/changeset/10302
21:16 Hyppolit Remove TODO item that shouldn't be done
21:16 Hyppolit trunk/Padre/lib/Padre/Wx/
21:16 tsee (in cleanup) Can't call method "update_decrement" on an undefined value at /home/tsee/perl/padre/trun​k/Padre/lib/Padre/Lock.pm line 56 during global destruction.
21:18 Alias Interesting
21:20 tsee On an aside, Alias, shouldn't you be asleep?
21:20 LeK I looked a little to find elements about sometimes... and it's strange, if I change the variable name, invert lines, etc... there's no logic except that  it happens at given lines in a reproductible ways... what do you think I could try to get more information? padre_log env? fulltrace? something else?
21:20 tsee You were awake all day (CET) today.
21:20 Alias tsee: I just woke up after 3 hours sleep, early inspection of smoke alarms, ugh
21:21 tsee Ah, I remember you mentioning that.
21:21 tsee We don't even have such a thing.
21:21 Alias You don't have smoke alarms?
21:21 tsee Nope.
21:21 Alias :)
21:21 tsee It's a house entirely made of stone, remember?
21:22 Alias It's possibly a cultural thing
21:22 Alias Australia has somewhat of a different relationship with fire perhaps
21:23 tsee Probably.
21:23 Alias The idea that any of our old houses would be made of stone is funny
21:26 Alias That said, the smoke detector rules are more recent
21:26 Alias New apartment buildings are required to have them, and the body corporate is required to keep them certified
21:26 Alias Which means a 2 minute test once a year
21:26 Alias I'm sure there's excemptions for older buildings
21:29 Alias I'm not entirely sure why you are seeing that destroy/cleanup error
21:30 Alias I can add a workaround though
21:30 tsee I have no idea.
21:30 tsee I've been editing Main.pm
21:30 Alias I know WHY you are seeing it
21:30 Alias Just not how you managed to get the locking system into that state
21:30 tsee With all bells and whistles enables (syntax check, outline, etc)
21:30 Alias Nah
21:30 tsee ... todo list
21:30 tsee Just saying.
21:30 Alias It's coming from something with essentually no movable parts
21:31 Alias essentially
21:31 Alias Maybe ... hrm
21:31 Alias I see no legitimate way for it to happen
21:32 Alias I'll pull in a special case workaround regardless
21:32 tsee Sorry for the hassle.
21:32 Alias n/p
21:33 Hyppolit svn: r10303 | adamk++ | http://padre.perlide.org/trac/changeset/10303
21:33 Hyppolit Special case workaround just in case the locker is destroyed somehow before the last lock refering to it is cleaned up. Shortcut the detruction process in general.
21:33 Hyppolit trunk/Padre/lib/Padre/
21:33 Alias That should fix it
21:33 tsee Thanks.
21:34 Alias It's a reasonable just-in-case to support regardless
21:34 Alias The cleanup process is still a bit magical to me
21:35 * Alias takes the opportunity to fix the stupid syntax checker problem
21:38 Alias Interesting
21:38 Alias It would appear the cwd setting is being ignored on syntax check
21:38 tsee Hmm. Does each thread maintain its own cwd?
21:38 Alias Dunno
21:38 tsee Also, maybe that's system dependent?
21:39 tsee Me neither.
21:39 tsee Easy enough to check.
21:39 Alias I know that we have to be really careful with cwds
21:39 Alias Especially on Windows
21:39 Alias We're supposed to pushd the cwd at the last possible moment during anything that does system stuff
21:39 Alias Because stale cwd cursors prevent directory deletion on windows
21:41 Alias my $pushd = File::pushd::pushd( $self->{cwd} );
21:41 Alias Padre::Util::Win32::ExecuteProcessAndWait( 'cmd.exe', "/C $cmd", 0 );
21:42 Alias My guess would be that "ExecuteProcessAndWait" picks the Cwd to use independantly
21:42 tsee Hmm.
21:42 tsee My threads test says they share the cwd.
21:42 Alias On my machine, cwd during syntax check is my Desktop
21:42 tsee But this is linux.
21:42 Alias Which is the shell default Cwd
21:42 Alias Are you seeing cwd issues
21:42 tsee I don't think so.
21:42 Alias At the top of a .pm file inside a project, add this
21:42 tsee Syntax check was working a couple of minutes ago.
21:43 Alias BEGIN { require Cwd; die "cwd = Cwd::cwd"; }
21:43 tsee BEGIN { chdir("/tmp")}?
21:43 Alias You should see a die message in the syntax checking output that shows the project root directory
21:43 Alias On windows, you see C:\Users\Adam\Desktop or something
21:44 Alias I'm guessing that's the true problem
21:44 tsee It's correct.
21:44 Alias A few other Padre functions fail similarly
21:44 Alias Which probably all stem from the same issue
21:44 Alias I need to replicate system() more closely on Win32
21:47 SvenDowideit joined #padre
21:47 Hyppolit svn: r10304 | adamk++ | http://padre.perlide.org/trac/changeset/10304
21:47 Hyppolit Convert silent acceptance to a TRACE warning when we "inefficiently load Padre::Util::Win32 on non-Win32 system"
21:47 Hyppolit trunk/Padre/lib/Padre/Util/
21:48 Alias Interesting Win32::API code here...
21:48 Alias Mixed case variable names... must be Sewi :)
21:48 Alias I almost want to write a badcode test for mixed-case variables :)
21:49 tsee Nah. I do that all the time in C++ :)
21:49 tsee And "our" vars.
21:49 Alias Unfortunately, when writeing Win32::API bindings, matching variables names to the Win32 API makes sense
21:49 Alias ick
21:49 * Alias always names globals that are manipulared from outside in $ALLCAPS
21:50 Alias Just to make clear they are specialk
21:51 Alias OK, it looks like there's a lpDirectory param to ShellExecuteEx that we just aren't using
21:51 Alias Which means I can fix this reasonably easily I hope
21:54 Alias It's not bad code tbh
21:54 Alias I'm amazed some of the things I'm learning as I work on Padre
21:54 Alias Every now and then you turn a corner in the code and run into a high-density learning experience
21:54 tsee Same here.
21:55 tsee It's the collaboration.
21:55 Alias Ya
21:55 tsee Different people do different things well.
21:55 Alias And the ease of dropping random shit into the app
21:55 Alias Without being judged
21:55 Alias If the 3-4 original guys were code-reviewing everything, half this stuff wouldn't happen
21:56 Alias Because people wouldn't have the patience to explain why just they cut and pasted 100 lines of half-working crap from somewhere else
21:56 tsee That and they simply wouldn't get to learn at their pace.
21:56 Alias ya
21:56 tsee And I think neither of us has the patience to explain the same things over and over again.
21:56 Alias Have you noticed how almost everyone has a rough first few dozen commits?
21:57 Alias Sewi made some large blunders, even Corion is doing some dangerous stuff at first
21:57 tsee (not counting Gabor who's made it his job and who has he friendly temper, too)
21:57 tsee Yep. While Padre's reasonably accessible, it's become a large application.
21:57 Alias People seem to need those first half dozen mistakes to get used to working properly
21:57 Sewi Alias: You converted me to lowercase  a long time ago :-)
21:57 tsee And not exactly well documented.
21:57 Alias Sewi: I know, but I seem to find your "legacy" in a number of places :)
21:58 Alias Sewi: Assuming that nobody else is doing mixed case as well :)
21:58 Alias So I'd chalk up "Helps the adaptation process by letting people make mistakes" as another anyone-commits benefit
21:58 Sewi English only uses uppercase letters for the beginning of a sentence and few other cases.
21:58 Alias So long as you are healthy enough as a community to catch and fix those mistakes
21:59 Sewi German forces you to write many more Words in uppercase Letters, this might by a Reason... :-)
21:59 Alias Interesting
21:59 Alias English forces it for proper nouns
21:59 Alias But even then, the Perl style is to avoid them because casing it used for other reasons
21:59 Alias i.e. $german, not $German
22:00 Alias or $deutch vs $Deutch
22:00 Alias sp?
22:00 Sewi deutsch :-)
22:00 Alias gah
22:00 Alias Sometimes I hate being in a English country
22:01 Alias I look at the nordic countries for what Australia would be like if we were non-English
22:01 Alias (Since our education systems compare somewhat)
22:01 Sewi You write a test for $Uppercase_Vars and I'll fix them, ok? :-)
22:01 Alias Sewi: There's too many, and it's too narcy
22:01 Alias I don't want to start badcode'ing things that aren't actually important functionally
22:01 Alias I'd rather write an automatic refactoring tool :)
22:02 Alias I've been meaning to work on that for a while
22:02 Alias A sort of perlstyle equivalent to perltidy
22:02 Haaarg joined #padre
22:02 tsee You got lexical replace!
22:02 Alias Something that would replace ->foo() with ->foo (or vice versa)
22:02 tsee Just make it work on projects.
22:02 Alias tsee: Right, but that means one at a time
22:02 Alias tsee: I'd then need bulk automation of the detection and replacement
22:02 tsee Yes, of course.
22:03 tsee But you can skip the PPI parsing part.
22:03 tsee Which should make it plenty fast enough.
22:03 Alias I'm not sure I'd want to
22:03 Alias There's $lexicals embedded in strings as ${lexical}
22:03 Alias There's others embedded in regexes
22:03 Alias etc
22:03 Alias Maybe once I get my test-integrity stuff running
22:04 Alias I had this idea for doing risky|PPI refactoring
22:04 Alias You take a project, save-all, then run make test and if it's ok, you apply the refactor and run make test again
22:04 Alias If the second make test fails, you back out the refactor
22:05 Alias That way, you can run 10 or 20 or 30 incremental automated changes
22:05 Alias Checking after each step that it works
22:05 Alias It's a bit on the slow side, but it does create the potential of doing really sweeping changes automatically
22:05 Alias Like, for example...
22:06 Alias "Find every statement in the code we can delete and still have the tests pass"
22:06 Alias Which would be neat for finding untested code :)
22:06 tsee :P
22:07 Alias That kind of thing will be when Padre really starts to come into it's own as an IDE
22:11 Getty and the computer starts writing his own code ;)
22:11 Getty skynet!
22:14 Alias :)
22:14 Alias It's stuff that other people have already
22:14 Alias Just haven't been done for Perl before really
22:14 Alias Not _properly)_
22:39 LeK i've narrowed the search for my bug... seems that getopt is bad for padre's health.... see you in the trac :)
22:40 Alias heh
22:40 LeK heh?
22:41 Alias chuckle-out-loud
22:41 LeK nice chuckling :)
22:44 Alias Excellent, syntax check works properly on Windows now
22:45 LeK great
22:47 Hyppolit svn: r10305 | adamk++ | http://padre.perlide.org/trac/changeset/10305
22:47 Hyppolit - Padre::Util::Win32::ExecuteProcessAndWait doesn't automatically inherit
22:47 Hyppolit the same Cwd as the parent process. Added support for explicit cwd
22:47 Hyppolit parameter and make the syntax checker pass the cwd to it. Syntax checking
22:47 Hyppolit of test scripts and such should now work as intended (ADAMK)
22:47 Hyppolit - Audit uses of Padre::Util::Win32 to only load it via require. Added a
22:47 Hyppolit TRACE warning to verify it never gets loaded on non-Win32 (ADAMK)
22:48 Hyppolit trunk/Padre/ trunk/Padre/lib/Padre/Task/ trunk/Padre/lib/Padre/Task/SyntaxChecker/ trunk/Padre/lib/Padre/Util/
22:48 Alias I wonder if we have a similar problem with script execution
22:48 Alias I wouldn't be entirely surprised
22:49 Alias hrm, nope that's a different problem
22:49 Alias Padre, when running .t scripts, doesn't know to set the Cwd to the project root
22:49 Alias It uses the t directory as the cwd
22:50 Alias Although "Run this Test" works, which is silly
22:50 Alias Shouldn't really be any difference between "running" and "running as test"
22:51 Hyppolit #832: GetOptions( spawns unsavedX tab and crash Padre in StepOver Debugging (new defect) [ http://padre.perlide.org/trac/ticket/832 ]
23:01 Hyppolit #833: Syntax error crash padre when trying to debug (new defect) [ http://padre.perlide.org/trac/ticket/833 ]
23:12 patspam joined #padre
23:27 Sewi Alias (backlog): Run as test should use prove
23:43 LeK bye
23:43 LeK left #padre
23:52 Hyppolit svn: r10306 | adamk++ | http://padre.perlide.org/trac/changeset/10306
23:52 Hyppolit - Tuned the locking for ->close_where, which should make a variety of
23:52 Hyppolit functions like "Close This Project" and "Close Other Projects"
23:52 Hyppolit noticably faster (ADAMK)
23:52 Hyppolit trunk/Padre/ trunk/Padre/lib/Padre/Wx/
23:56 Hyppolit svn: r10307 | adamk++ | http://padre.perlide.org/trac/changeset/10307
23:56 Hyppolit Improved locking code further to also lock the database, and to avoid locking at all if the close_where results in nothing to close.
23:56 Hyppolit trunk/Padre/lib/Padre/Wx/
23:57 Alias joined #padre

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