Camelia, the Perl 6 bug

IRC log for #padre, 2010-06-09

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

All times shown according to UTC.

Time Nick Message
00:19 Alias jq: Pong
01:02 jazzanova joined #padre
01:24 submersible1 joined #padre
02:20 szabgabx__ joined #padre
02:42 Hyppolit svn: r11569 | adamk++ | http://padre.perlide.org/trac/changeset/11569
02:42 Hyppolit The ErrorList view has been upgraded to the new View API, and (mostly) to the new task API.
02:42 Hyppolit branches/Padre-TaskRewrite/lib/Padre/ branches/Padre-TaskRewrite/lib/Padre/Action/ branches/Padre-TaskRewrite/lib/Padre/Task/ branches/Padre-TaskRewrite/lib/Padre/Task2/ branches/Padre-TaskRewrite/lib/Padre/Wx/ branches/Padre-TaskRewrite/t/
02:44 Hyppolit svn: r11570 | adamk++ | http://padre.perlide.org/trac/changeset/11570
02:44 Hyppolit Noting why bits are incomplete
02:44 Hyppolit branches/Padre-TaskRewrite/lib/Padre/Task2/
03:35 psybermonkey joined #padre
03:51 |Sno| joined #padre
03:52 user_8554 joined #padre
03:52 user_8554 hi folks
03:55 Alias hi
03:57 user_8554 any guys from padre dev team
03:57 Alias Yes
03:57 Alias This is a all in one mixed channel
03:58 Alias Dev channel + support channel + user chat
03:58 user_8554 i was interested in contributing to padre
03:58 [Sno] joined #padre
03:58 user_8554 downloaded the padre source from svn and had a nightmare with wxwidgets but finally got the bild working
03:58 user_8554 build*
03:58 Hyppolit svn: r11571 | adamk++ | http://padre.perlide.org/trac/changeset/11571
03:58 Hyppolit Adding LWP dependency to the core.
03:58 Hyppolit Nobody has a good reason for multi-mechanism driver tactics.
03:58 Hyppolit branches/Padre-TaskRewrite/
03:59 Hyppolit svn: r11572 | adamk++ | http://padre.perlide.org/trac/changeset/11572
03:59 Hyppolit Completed the migration of the background HTTP request tasks to the new Task 2.0 API
03:59 Hyppolit branches/Padre-TaskRewrite/lib/Padre/Plugin/ branches/Padre-TaskRewrite/lib/​Padre/Plugin/PopularityContest/ branches/Padre-TaskRewrite/lib/Padre/Task/ branches/Padre-TaskRewrite/lib/Padre/Task2/ branches/Padre-TaskRewrite/lib/Padre/Wx/ branches/Padre-TaskRewrite/lib/Padre/Wx/Dialog/
04:00 user_8554 any startup guide with architecture would help me to get started
04:00 Alias What kind of stuff are you interested in?
04:01 user_8554 in devloping or writing a fix also write some automated cases for smoke
04:01 Alias Well for the testing side of it, Padre is just a normal CPAN distribution
04:02 Alias So if you look through the family of tests in /t/ you should be able to steal the technique from a test that does similar things to what you want to do
04:02 Alias That's the "t" directory, if that didn't render right for you
04:02 user_8554 ok
04:03 Alias Second, start up the checkout version with "perl dev.pl"
04:03 Alias And turn on the Padre Developer Plugin
04:04 Alias That gives you some toys so you can start looking at the internals of the live running Padre instance
04:04 user_8554 ok alias u meant .../Padre/t directory correct ?
04:04 user_8554 ok
04:04 Alias And screw around with it
04:04 Alias Correct
04:04 Alias The repository trunk contains a set of distributions
04:04 Alias With /Padre/ being the distro for the core editor
04:05 Alias And the others being some support modules we made in the process of writing the editor, various plugins in all kinds of different states (some live, some dead, some never worked at all)
04:06 Alias One of the more interesting tools you might want to do is "Dump Top IDE Object"
04:06 Alias That gives you the entire Perl structure of the running IDE
04:09 user_8554 i guess the developer plugins is enabled by default i we got a build from source ?
04:09 user_8554 ok
04:09 user_8554 great thanks Alias
04:09 Alias It's installed by default, but not enabled
04:10 user_8554 ok
04:10 Alias waxhead: I'm now down to three or things left to migrate to Task 2.0, when do you want to do the 0.64 release so I can merge to trunk?
04:10 * Alias afk
05:12 psybermonkey joined #padre
05:18 Xaero joined #padre
05:59 danlucraft joined #padre
06:02 pece joined #padre
06:08 Xaero joined #padre
06:27 kaare joined #padre
06:30 ecocode joined #padre
07:17 jq Alias: cf the mail i sent you
07:24 * Alias opens mail for the first time todaty
07:25 Alias Interesting
07:25 Alias Config vs data is problematic
07:25 jq why?
07:25 Alias On Windows, the idea that you've EVER edit a config file yourself is rediculous
07:26 Alias You'd
07:26 Alias So in that sense, the application is just storing config, which really makes it treatable as "data"
07:26 Alias And who cares if it's mixed in with other types of data
07:26 jq indeed, and the config can also go in the registry
07:26 Alias Right
07:26 Alias The user should never have to care
07:27 jq so, what to do?
07:27 jq my_config() returns my_data() on windows?
07:27 Alias On Linux, there's value to the separation because you don't want to store a billion cache files and config files in the same place
07:27 Alias I really don't know
07:28 Alias I REALLY REALLY don't want to put anything in File::HomeDir that is our guess at correctness
07:28 Alias Everything so far is 100% to spec
07:28 Alias Which is why the FreeDesktop stuff only landed very recently
07:28 Alias I wasn't game to half-ass it
07:29 Alias The my_dist_data concept as you propose doesn't REALLY violate that too far
07:29 Alias Because "data" -> "vendor" -> "product" holds
07:29 Alias Which seems to be the primary pattern
07:30 jq do you think it should go in another dist?
07:30 Alias my_config is a bigger problem
07:30 Alias I don't see a compelling reason to have it separate
07:31 jq me neither, but i prefer asking the question *now* :)
07:31 Alias There's no extra dependencies, and now we support FreeDesktop it's fairly reasonable for us to make the call that we can just mandate $HOME/.perl by fiat
07:31 Alias (for the plain Unix driver)
07:32 Alias (since it's not only the absense of a standard, but also the absense of any potential future standard)
07:32 Alias absence
07:32 Alias blah
07:33 jq so, to summarize: my_dist_data() is ok as is for you, my_dist_config() needs more thoughts
07:33 Alias Anything to do with config needs more thought
07:33 azawawi joined #padre
07:33 Alias If we're assuming config implies human-editable
07:33 azawawi hello
07:34 Xaero o/
07:34 Alias Probably need to read the freedesktop spec again to see how they differentiate
07:34 azawawi human-editable? :)
07:34 jq btw. in the doc, you wrote "the module tries extensively to make sure that directory exists" (or sthg like that)
07:34 Xaero are you guys discussing about padre?
07:34 jq ==> do you think we should create the directory?
07:34 Alias Xaero: File::HomeDir
07:34 jq Xaero: nope, about file::homedir
07:34 Alias jq: Different meaning
07:34 Alias jq: "If the module returns a path, it will exist and you don't need to test -e yourself"
07:34 Alias jq: Otherwise, it returns undef
07:35 jq right. my_dist_data() doesn't do that
07:35 Alias On some platforms, that's a little non-trivial
07:35 Alias On Windows, the directory won't exist until you make the system call to ask if it exists, and Windows creates it on the fly
07:35 jq ie, it's pretty guaranteed that the dir won't exist when called for the first time
07:35 Alias yeah, that's a new twist
07:36 Alias You also haven't covered the dist vs module problem
07:36 Alias See File::ShareDir
07:36 jq since i'm at it... would you accept me to patch file::homedir to accept parms during loading to return path::class objects?
07:36 Alias Wouldn't that be a global change in behaviour?
07:37 jq eg: "user File::HomeDir returns=>'Path::Class';" (or whatever)
07:37 Alias That's probably a bad idea
07:37 jq no, by default it returns regular string
07:37 Alias File::HomeDir could easily have multiple things calling it
07:37 danlucraft joined #padre
07:37 Alias Your code ----> HomeDir  <---- Some other module
07:37 Alias Action at a distance etc etc
07:38 Alias Plus it would mean more dependencies, and threaten a future migration to the core
07:38 Alias (Unless Path::Class is core...)
07:38 jq ok, i take that as a "not now" answer ;-)
07:38 Alias That part almost sounds better as a different dist
07:39 Alias File::HomeDir::PathClass (for example)
07:39 Alias Which would be a lookalike API that just does a pass-through
07:39 Alias You could even method-scan normal homedir to work out which shims to create
07:39 azawawi Alias: any eta on the task2-rewrite merge?
07:40 jq Alias: ok, so can you take some time to think about my_dist_config()?
07:40 Alias azawawi: 1. I've got 3 things left to migrate (SyntaxChecker, DocBrowser and PPI-based transforms)
07:40 Alias azawawi: 2. waxhead hasn't done his "final" 0.64 release yet
07:40 jq we'll discuss it later when you'll have some time to think about it
07:40 Alias jq: Sure
07:41 Alias azawawi: I'm pretty close now though
07:42 Alias I just detoured to rewrite the way tools and panels are done, which has made it easier
07:42 Alias I figured since we're breaking every plugin already, we should probably sneak in any other minor stuff that would break lots of plugins
07:42 azawawi Alias: thx for the update
07:42 Alias Lots of commits landing though, so making good headway
07:43 Alias And Padre still seems to work just fine
07:43 Alias Maybe some scalar leaking to readdress
07:43 azawawi Alias: I needed that info to migrate to Task2 API for Perl6 plugin
07:43 Alias And I've regressed a few minor features for the moment
07:43 Alias ah
07:43 Alias You use outline right?
07:43 Alias Big changes there
07:44 Alias Outline is well and truly finished, so if you wanted to start migrating the Perl 6 plugin now would be a reasonable time to start
07:44 cognominal joined #padre
07:44 Alias I'm happy enough with the APIs, just slogging through the last of the migration
07:44 Alias Well, almost happy enough
07:44 Alias Maybe a few small class naming tweaks
08:08 waxhead Alias, I'll send out an email tonight asking for translations etc.  Do a release firday night, paving the way for the weekend to do the merge...
08:08 waxhead how does that sound?
08:09 Alias_ joined #padre
08:16 azawawi joined #padre
08:20 waxhead I'm running padre under the branch at hte moment..
08:20 waxhead I'll try and take a look at what's left and have a go at getting it migrated.. cooking dinner at the moment though
08:21 azawawi Alias: I think we need some kind of a padre project file. (e.g. define PERLLIB, pick Perl 5/6 interpreter, etc...)
08:22 azawawi Alias: i know we can figure out some stuff from Makefile.PL and META.yml, ..etc but that's 100% correct
08:22 azawawi Alias: im just thinking. Ignore it if you're busy :)
08:23 danlucraft1 joined #padre
08:26 Alias_ Sorry, been fiddling with projector
08:26 Alias_ "project file"?
08:26 Alias_ We kind of already do
08:26 Xaero ah right! did you get a new cable yet?
08:27 Alias_ Different problem
08:27 Alias_ Horizontal banding, as if part of the LCD lost input
08:27 Alias_ Rather scary
08:27 Alias_ Turned away for 15 minutes to find and read the warranty and when I look back it's gone
08:27 Alias_ Slightly more scarier
08:28 Xaero what happened? the projector not working?
08:29 jq Alias: when you have an idea regarding my_dist_config(), can you reply by mail please?
08:30 Alias_ Temporarily not working, worse
08:41 ingy joined #padre
08:49 Steffen joined #padre
09:12 Steffen left #padre
09:22 sorear joined #padre
09:29 waxhead Alias_, email sent to the dev list... we'll go with 0.64 on friday night, then you're free to do what you like to trunk over the weekend..
09:29 Alias_ ok
09:41 waxhead I honestly expected to see Padre in the heavy 100...
09:49 Alias_ It is
09:49 Alias_ Well, it was
09:56 waxhead I'm poking around the Task2 stuff... I have no idea where to start... :-/
09:56 * waxhead goes back to looking at the code
09:59 Alias_ Padre made it as high as 90th
09:59 Xaero on what?
10:00 Alias_ On the Heavy 100 list
10:00 Xaero what is that?
10:00 Alias_ The 100 modules with the highest number of dependencies
10:00 Alias_ See http://ali.as/top100
10:01 Alias_ waxhead: Padre is now about 20 dependencies outside of the top 100
10:01 Alias_ Although I just re-added LWP as a dep on my branch, so I expect we'll be closing that gap a bit
10:03 Alias_ hrm
10:04 Alias_ Looks like DBIx::Class just added Moose as a dependency
10:04 Alias_ Which means Padre is immune from the dependency bump
10:04 Alias_ So we're now about 40 dependencies outside the top 100
10:05 Xaero i thought it would be good to compare after reads the jobs newsgropup, so i make http://i46.tinypic.com/29e1yc6.png :P
10:09 Alias_ There's a mistake in that graph
10:09 Xaero ?
10:11 Alias_ You've use two different graph types for data with a similar nature
10:11 Alias_ The bar graph is an error, it should be a line
10:11 Alias_ Although admittedly there is higher variation in the bars
10:11 Alias_ Maybe you could smooth it
10:12 Alias_ In fact, ignore me
10:12 Alias_ I see why you did it
10:12 Xaero the line graph for jobs looks as scary as hell, i tried smoothing too, it would need more than 6 of smoothing for it to be smooth
10:17 Alias_ Scary because it's jagged?
10:17 Xaero it looks more like noise
10:21 Alias_ fair enough
10:39 aukjan joined #padre
10:54 Alias_ waxhead: The Task2 stuff will probably make more sense once all the tasks are migrated, and I've had time to write more documentation
10:54 Alias_ Although, tbh, it's designed to be easy to use as a task writer, not so much to understand the deeper internals
10:55 Alias_ They are fairly elegantly written, but rather complicated to understand
11:49 Steffen joined #padre
11:50 Steffen left #padre
13:16 aukjan joined #padre
13:20 psybermonkey joined #padre
13:51 rindolf joined #padre
13:57 rindolf Hi all.
14:00 Xaero o/
14:01 rindolf Hi Xaero
14:01 Sewi Hi rindolf
14:01 Sewi Are you from AU, too?
14:01 rindolf Sewi: no.
14:01 rindolf Sewi: I'm from .il.
14:01 rindolf Sewi: like szabgab
14:04 Alias_ Bloody hell it's cold
14:04 Alias_ Although by Sydney standards, that's 10c
14:12 pece joined #padre
14:12 Xaero 10C? "Bloody hell it's cold" is justified then
14:14 Sewi Alias: We got around 30 here today...
14:18 Alias_ 30 is "nice and warm"
14:18 Alias_ Now 40, 40 is "hot"
14:20 Alias_ Turns out quite a few people in Sydney die from the cold each year
14:20 Alias_ Because none of the houses are insulated against the cold properly, and nobody owns proper cold weather gear
14:21 Alias_ I lost my nice warm English jacket a few years ago :(
14:21 Alias_ Fortunately, I now have an awesome Icelandic jacket :)
14:22 Xaero newspaper works just as well is what I hear
14:23 * sjn notes that Alias_ thinks 40 is hot :)
14:23 Alias_ sjn: Of course, 45 is "bloody hot"
14:24 sjn there we go
14:24 Alias_ Which I note is also the point at which road tar starts to melt :)
14:24 * sjn sends alias the letters M, I, L and F :)
14:24 Alias_ heh
14:26 Xaero Alias, what was the problem with the heavy100 dependencies thingy everyone is talking about?
14:26 Xaero is it fixed now?
14:26 Alias_ Turns out that DBIx::Class just added Moose as a dependency
14:27 Alias_ And the Top 100 is overrepresented with big database-driven apps
14:27 Alias_ Moose is not exactly light-weight when it comes to deps
14:33 Sewi 42 Deps is not really lightweight :-)
14:33 * Xaero moos
14:34 Sewi DBIx::Class 131 - must be more than Padre
14:34 Xaero one thing I always wondered about, is there anyone in here who is not a professional developer?
14:36 Sewi Sure, Zenog is student
14:37 Xaero but he is learning programming in school/college, right? and will be a pro someday?
14:39 Sewi don't know
14:42 patspam joined #padre
15:06 kaare joined #padre
15:12 Sven_Windows__ joined #padre
15:17 [1]awnstudio joined #padre
15:19 Alias joined #padre
15:19 waxhead joined #padre
15:19 Hinrik joined #padre
15:19 mattp joined #padre
15:47 Hyppolit svn: r11573 | adamk++ | http://padre.perlide.org/trac/changeset/11573
15:47 Hyppolit Completed the migration of the syntax checker to the Task 2.0 API
15:47 Hyppolit branches/Padre-TaskRewrite/lib/Padre/Document/ branches/Padre-TaskRewrite​/lib/Padre/Document/Perl/ branches/Padre-TaskRewrite/lib/Padre/Task/ branches/Padre-TaskRewrite/lib/Padre/Task2/ branches/Padre-TaskRewrite/lib/Padre/Wx/ branches/Padre-TaskRewrite/t/
15:47 Alias_ Right, that's one more done
15:47 Alias_ Took a shitload longer than I expected
15:48 Alias_ Down to just a few bits and pieces now
15:48 * Alias_ bed
15:48 Xaero gn
15:49 Sewi gn8 alias
16:15 Steffen joined #padre
16:21 Alias__ joined #padre
16:24 Alias joined #padre
17:35 aukjan joined #padre
19:15 szabgabx joined #padre
19:44 LeK joined #padre
20:38 danlucraft joined #padre
21:55 szabgabx I am a bit busy at LinuxTag, could someone please fix the download page to point to the new Strawberry + Padre?
22:14 Haarg joined #padre

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