Camelia, the Perl 6 bug

IRC log for #mojo, 2012-11-12

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

All times shown according to UTC.

Time Nick Message
00:21 Molaf__ joined #mojo
01:01 laouji joined #mojo
01:12 mattastrophe joined #mojo
01:16 Mike-PerlRecruiter_ joined #mojo
01:38 d4rkie joined #mojo
01:52 Kulag joined #mojo
02:00 lukep joined #mojo
02:03 dvinciguerra joined #mojo
02:18 jpn joined #mojo
02:44 CountryGeek_away joined #mojo
03:24 Foxcool joined #mojo
03:32 Kulag joined #mojo
03:33 noganex_ joined #mojo
03:44 Foxcool_ joined #mojo
04:00 xaka joined #mojo
04:19 Foxcool joined #mojo
04:29 Foxcool joined #mojo
04:52 kitt_vl joined #mojo
05:00 Kulag joined #mojo
05:04 laouji joined #mojo
05:22 Drossel joined #mojo
05:24 ver joined #mojo
05:41 Kulag joined #mojo
06:00 Kulag joined #mojo
06:20 spleenjack joined #mojo
06:35 Kulag joined #mojo
06:47 Kulag joined #mojo
06:54 yakudza_ joined #mojo
07:10 Molaf_ joined #mojo
07:11 dpetrov_ joined #mojo
07:11 AmeliePoulain joined #mojo
07:12 AmeliePoulain joined #mojo
07:16 Molaf joined #mojo
07:44 mire joined #mojo
07:45 mtths joined #mojo
07:49 dod joined #mojo
07:55 Molaf_ joined #mojo
08:01 mtths joined #mojo
08:10 hlin joined #mojo
08:10 Molaf joined #mojo
08:16 marcus joined #mojo
08:18 dross :3
08:20 Molaf_ joined #mojo
08:22 dross news
08:22 dross oop
08:25 Molaf__ joined #mojo
08:31 Molaf joined #mojo
08:34 dod joined #mojo
08:35 Molaf_ joined #mojo
08:39 dod joined #mojo
08:41 Molaf_ left #mojo
08:44 coff joined #mojo
09:04 mire joined #mojo
09:08 Averna joined #mojo
09:14 batman joined #mojo
09:21 Vandal joined #mojo
09:26 cosmincx joined #mojo
09:33 fhelmber_ joined #mojo
09:45 Britzel joined #mojo
09:47 mire_ joined #mojo
09:58 sr joined #mojo
10:13 sri jberger: re the Devel::Cover thing, i would really like to see a version of Mojo::Template that works with #line comments
10:14 sri in fact, i have an unfinished patch
10:16 sri it has the potential to require a lot less code and be more reliable
10:16 sri but there are a few edge cases i never got around to fixing
10:17 sri http://hastebin.com/raw/xitekudoxo # the half finished patch rebased on master, in case anyone wants to work on it
10:18 sri i think it's pretty much only errors in nested templates causing problems
10:22 sri summary for everyone else, we currently parse perl errors to find the right line number for Mojo::Template error messages with context
10:23 sri that might not be necessary if we can add support for this to Mojo::Template http://perldoc.perl.org/perlsyn.h​tml#Plain-Old-Comments-%28Not!%29
10:28 sri there are only 3 test cases broken
10:28 sri and they are prolly related
10:41 Molaf joined #mojo
10:42 sri http://hastebin.com/raw/hebawejini # and down to one error
10:42 sri browser error messages are losing a little detail though
10:42 sri should be worth it if we can make it work though
10:42 sri leaves one test fail in t/mojo/template.t
10:45 sri hmm, that one might actually be unfixable
10:48 marcus o/
10:48 sri \o
10:57 sri ok, i don't think heredocs and multiline strings can work with line directives
10:57 sri but
10:57 sri i have a fully working patch
10:57 sri http://pastie.org/5365057
10:57 sri jberger, marcus, tempire, crab: review!
10:59 sri https://github.com/kraih/mojo/bl​ob/master/t/mojo/template.t#L716
10:59 sri that test is kind of a big problem
10:59 sri if we decide that we want it to keep working i don't think we can ever use line directives
11:01 sri oh, and i'm wrong :)
11:02 * sri is a moron
11:04 sri you only need one line directive for the rest to work
11:05 sri ok, now for real :)
11:05 sri jberger, marcus, tempire, crab: review http://pastie.org/5365090
11:07 * sri is +1 on the patch
11:07 sri this line detection strategy is a lot more future proof
11:11 good_news_everyone joined #mojo
11:11 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/fz-GHw
11:11 good_news_everyone mojo/master 99b1521 Sebastian Riedel: improved error message accuracy in Mojo::Template by using line directives
11:11 good_news_everyone left #mojo
11:11 sri FUCK IT WE"LL DO IT LIVE
11:19 good_news_everyone joined #mojo
11:19 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/6mpDAw
11:19 good_news_everyone mojo/master 45af12e Sebastian Riedel: small optimizations
11:19 good_news_everyone left #mojo
11:21 sri only downside i can see is the loss of some information in the actual error messages
11:21 sri like https://github.com/kraih/mojo/commit/99b15​21110cd75897c69371f2b6d1806823c67e2#L5L692
11:22 sri that information is still present in log files though
11:25 sri the fact that Mojo::Exception now has full test coverage alone is well worth it imo
11:32 batman does hypnotoad close STDOUT and STDERR when started as a daemon?
11:33 Vandal joined #mojo
11:39 good_news_everyone joined #mojo
11:39 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/NugNZQ
11:39 good_news_everyone mojo/master 303e9bc Sebastian Riedel: more context for Mojolicious template errors
11:39 good_news_everyone left #mojo
11:39 sri and the downsides are gone :)
11:40 sri line directives just don't allow ", everything else seems fine
11:42 marcus sri: does this break any backwards compat?
11:42 sri marcus: don't think so
11:42 marcus it seems like a good change.
11:42 marcus sri: +1 from me then
11:43 sri only if you were checking error messages char for char
11:43 marcus which would be pretty crazy
11:43 sri ye
11:46 ladnaV joined #mojo
11:47 ObseLeTe joined #mojo
11:50 Vandal joined #mojo
12:04 ryozi joined #mojo
12:05 Averna joined #mojo
12:17 ladnaV joined #mojo
12:21 Vandal joined #mojo
12:24 ladnaV joined #mojo
12:27 AmeliePoulain joined #mojo
12:28 Vandal joined #mojo
12:33 ladnaV joined #mojo
12:34 AmeliePoulain joined #mojo
12:40 good_news_everyone joined #mojo
12:40 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/yCl-1w
12:40 good_news_everyone mojo/master 01df8cf Sebastian Riedel: documentation tweaks
12:40 good_news_everyone left #mojo
12:43 Vandal joined #mojo
12:46 motoboi joined #mojo
12:47 ladnaV joined #mojo
12:48 mire_ joined #mojo
13:07 tholen42 joined #mojo
13:18 Mike-PerlRecruiter_ joined #mojo
13:49 good_news_everyone joined #mojo
13:49 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/UHzURA
13:49 good_news_everyone mojo/master 76f320d Sebastian Riedel: better template names for commands
13:49 good_news_everyone left #mojo
13:50 sri interesting that those names now show up correctly on stack traces too
13:51 sri not sure if having "template foo.mt from DATA section" is better than "foo.mt" in a stack trace
13:51 sri hmm
14:08 inokenty joined #mojo
14:09 sh4 joined #mojo
14:25 banana joined #mojo
14:27 mattastrophe joined #mojo
14:28 banana hello .. i want to store my Controllers in a sub folder and name them like "MyApp::Controllers::MyController" .. how can I still use the "/:controller/:action" route ?
14:32 D4RK-PH0ENiX joined #mojo
14:33 Ulti is there a recommended way of forking from within mojolicious so that all file descriptors are closed in the child?
14:37 labrown joined #mojo
14:45 banana found :   $r->get('/:controller/:action')->to(   namespace => 'My::SubFolder',   controller=>$self->stash("controller"),   action=> $self->stash("action")   );
14:45 dod joined #mojo
14:54 d4rkie joined #mojo
14:56 jberger sri: sorry, I was out last night
14:56 jberger exciting that you seem to have been making progress on the line directives
14:56 jberger I'm going to go back and look at it now
14:57 sri jberger: i think it's solved :)
14:57 jberger excellent!
14:58 jberger so if I run cover on galileo's test suite, will I see the templates now?
14:58 sri you tell me!
14:58 * jberger tries
14:58 jberger need to update mojo first
14:58 sri it's 30 lines of code less, so it's a win no matter what
14:59 jberger what was the issue? I'm sure I can glean it from the code, but is there a quick summary?
15:03 sri no issue, just better code
15:03 sri no clue if it helps with Devel::Cover, i just wanted to change that code for a long time
15:06 jberger I'm checking now, I thought I should update Devel::Cover too, just to be sure I'm up on all the new codez
15:10 dabudabu joined #mojo
15:14 jberger I'm still getting warnings about not finding the templates
15:14 jberger http://hastebin.com/raw/secikoheqa
15:15 jberger I do find it interesting that it knows that its a template
15:15 jberger is it just that it can't find the file I wonder?
15:18 sri ah
15:18 sri it knows because we tell it the file is called "template foo.mt"
15:19 sri that's what i was wondering about earlier
15:21 sri i guess Devel::Cover would require the full file system path
15:21 sri without any decorations
15:22 jberger so its trying to open a template file named "template page/show.html.ep"?
15:22 sri don't think we can do that
15:22 sri prolly
15:22 sri https://github.com/kraih/mojo/commit/30​3e9bc2d4cd619c52414166f1737bbcf444c8b1
15:22 jberger maybe this is where Devel::Cover could meet us halfway?
15:23 sri it's pretty arbitrary how we call templates
15:23 jberger yeah, I guess thats true
15:23 sri we optimize for humans
15:24 sri also, we give relative paths in a neutral format, not os dependent
15:25 sri my goal so far was to get as close as possible to the previous state while making the code better
15:25 sri if you want to change error messages that's a different discussion
15:26 jberger no, I don't disagree
15:27 jberger internally mojo must use the #file directive with the name as the filename?
15:27 ObseLeTe joined #mojo
15:27 sri the commit i linked was the one where i made everything Cevel::Cover unfriendly again i suppose
15:28 sri #line 1 "foo/bar.html.ep"
15:28 sri or rather
15:28 sri #line 1 "template foo/bar.html.ep"
15:28 jberger rigt
15:28 sri in some cases even
15:28 sri #line 1 "template foo/bar.html.ep from DATA section"
15:29 jberger even that doesn't help if the files aren't relative to the root, I imagine
15:29 sri they may be in one of many search paths
15:30 jberger in galileo, could specify my template names to work
15:30 jberger so that it doesn't have to be a mojolicious goal, but I could make it one for galileo?
15:30 sri i don't think you can do it portably
15:30 sri windows paths and so on
15:31 jberger test coverage doesn't have to be portable I suppose
15:31 jberger I don't expect that my amount of coverage would be platform dependent
15:31 dod joined #mojo
15:31 jberger and I wouldn't have the user-side install run cover anyway
15:32 sri i think the big question here is if we want to keep error messages user friendly
15:32 sri "Died at template dies.html.ep from DATA section line 2." vs "Died at dies.html.ep line 2."
15:33 * sri pokes marcus, tempire and crab
15:34 marcus sri: I was them user friendly
15:34 marcus I want them...
15:34 marcus Pretty please.
15:34 * sri tends towards user friendly too
15:35 marcus Can't we just hide templates from Test Coverage? :-P
15:35 jberger not that I disagree, but isn't the fact that its .ep imply that its a template
15:35 jberger marcus: I want coverage
15:36 marcus jberger: I want a Pony.
15:36 marcus and designer friendly error messages
15:36 jberger or at least I think its a worth endevour and its not that far off if I understand correctly
15:36 jberger why does Devel::Cover want an MD5 of the file anyway?
15:37 jberger just for repeated use right?
15:37 jberger so I don't think that that is the root problem
15:37 jberger I need to dig deeper into d::C
15:37 sri it surely needs a real path to match the generated code to a file
15:37 knshaum joined #mojo
15:38 sri without making it "Died at c:\lalala\myapp\templates\foo\bar.html.ep line 2." it will never work with Devel::Cover i think
15:40 sri DATA and inline templates are out of the question for Devel::Cover anyway
15:40 sri (as far as i understand)
15:45 jberger sri: for DATA and inline you would probably have to use the #line directive in the same file, but I'm not sure you would want to do that
15:45 sri true, that would work and suck :)
15:45 * jberger nods
15:47 sri finding the actual file could also turn out to be impossible
15:47 sri DATA can be redefined for a class
15:47 jberger I know that PDL has to do lots of magic with #line, but I know that its a lot of hassle for the devs
15:48 * jberger shivers thinking about *DATA manipulation
15:48 sri i can imagine it getting easily out of hand
15:49 jberger btw I just tried installing mojo when checked out at 45af12e (just before you changed the messages) still didn't work, so it seems that it will need full paths
15:50 sri line directives are at least a lot better than parsing error messages :)
15:50 jberger I'm just seeing IF I can get it to work, then we can decided if we want it to
15:50 jberger yeah, your earlier work is good in any case
15:51 * sri had no idea you only need one directive per file
15:56 d4rkie joined #mojo
15:57 cafe joined #mojo
16:00 sh4|2 joined #mojo
16:05 mire_ joined #mojo
16:09 jberger ok, so here's where it stands
16:09 jberger if I name the template $path
16:09 jberger cover notices it and puts it in the database
16:09 jberger but it doesn't seem to parse it
16:10 jberger as I originally thought, I think D::C needs to use the parsed template, not the raw one
16:10 jberger then report the line numbers related to the template
16:11 jberger and I have NO IDEA if thats possible
16:11 jberger until then, making the template name something "designer" prevents the file from being in the cover_db report
16:12 jberger but means that there are warnings issued (which is probably what it should do)
16:12 jberger it reports the "right error" for the wrong reason
16:12 jberger can'
16:12 jberger can't open file ... presumably for coverage testing
16:14 jberger and the counter point is (as always) if you really need test coverage reporting on your templates, then you have too much logic in your views
16:14 jberger so who knows if its worth it
16:15 jberger sri: either way, fine work on the line directives
16:16 sri guess that's true
16:16 sri helpers should be covered after all
16:17 jberger I had mostly wanted to get coverage on some 'if' branches
16:18 jberger but since I can test the resultant HTML output, even that is testable
16:18 jberger I just don't get a coverage report saying I hit all of them
16:18 jberger which is how all testing worked before D::C right :-P
16:22 ruz joined #mojo
16:39 Polarn joined #mojo
16:55 alester joined #mojo
17:36 yakudza joined #mojo
17:37 dpetrov_ joined #mojo
17:40 xxtjaxx_ Where do you put templates when installing a mojolicious application?
17:40 xxtjaxx_ And how do you do that with Module::Build
17:43 sri http://mojolicio.us/perldoc/Mojolicious/Guides​/Cookbook#Making_your_application_installable
17:54 coff joined #mojo
18:03 xxtjaxx_ sri: Thanks
18:09 batman joined #mojo
18:24 xaka joined #mojo
18:41 dod joined #mojo
18:51 marty joined #mojo
19:38 dvinciguerra joined #mojo
19:42 sri my macbook is supposed to arrive tomorrow \o/
19:43 * sri wonders if he will get a samsung or toshiba ssd
19:44 sri apparently toshiba sucks with filevault2 :S
19:50 coff sri: The problem with getting new gadgets are that productivity drops, at least for a time...
19:50 coff sri: 1st world problem. :)
19:51 sri and happiness goes through the roof \o/
20:00 Polarn_ joined #mojo
20:19 sri still wonder what to do about Mojolicious::Routes::Match
20:20 sri has to be some way to make the api more elegant
20:29 mattastrophe joined #mojo
20:32 rem_lex| joined #mojo
20:36 lukep joined #mojo
20:51 rem_lex joined #mojo
21:16 Kulag joined #mojo
21:46 batman left #mojo
22:08 mire_ joined #mojo
22:25 marty I got my macbook pro a few months ago and I'm still not very productive.  Of course, I've also been living in the woods for the last few months.  :)
22:29 SmokeMachine_ joined #mojo
22:38 marcus marty: you could always run linux in parallels
22:40 marty Marcus. That's exactly what I do and it works well.   My lack of productivity is due to my poor self motivation rather than my platform.  ie: I wanna go play in the woods more.  :)
22:41 marcus :D
22:41 marty http://www.flickr.com/photos/87096950@N​04/8180255603/in/set-72157631511738720
22:43 Averna joined #mojo
22:54 jzawodn joined #mojo
23:00 marcus marty: how many square meters is that cabin? It looks a bit small?
23:01 marcus I live fairly near the forrest as well, but not actually *in* it  - https://www.evernote.com/shard/s22/s​h/18175932-87d7-46f2-892b-95e4532599​fa/d448ef83cda4e32844b3a4ea03dea1b9
23:02 marty marcus:  Yes, it's small.  about 74 square meters.  It's just for weekends at this point.  But some day..... full time.   :)
23:02 marcus marty: that's pretty good for a weekend home tho :)
23:03 marcus marty: interesting architecture.
23:03 marty the best part is. I can get a 3g signal there.  Soon, I'll be able to take my laptop with me and work on Mojo while I'm in the woods.   Weeeee
23:04 marty It's a heavy snow area.   the A-Frame design is fairly common for high snow areas here in the states.
23:04 marty Saw a bear and some Elk last weekends.
23:05 marty bear was about 20 feet away.  Caught be by surprise.
23:05 marty marcus:  From you map it looks like you are near a LOT of forest area.
23:06 marcus marty: I'm also 15 minutes from downtown Oslo tho.
23:06 marty Lots of lakes on that map.  I'll bet is it gorgeous.
23:06 marcus I was visiting my parents at Lillehammer this weekend. She was complaining about a really big elk running around on the fields and in her garden and stuff.
23:07 marcus (my mother)
23:07 marcus We don't see much bear around here tho :)
23:08 * marty thinks he would like Norway
23:13 * marcus thinks he would like the northern parts of the US.
23:14 * marcus thinks he would like it if Texas became a separate state :)
23:20 marty Maybe we can swap Texas for Puerto Rico.  :)
23:22 marcus marty: seems like a good deal!
23:23 marcus and you could keep the flag as it is as well.
23:31 SmokeMachine_ joined #mojo
23:35 rem_lex|pivo joined #mojo
23:35 mattastrophe joined #mojo
23:47 cafe_ joined #mojo

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