Camelia, the Perl 6 bug

IRC log for #mojo, 2011-05-06

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

All times shown according to UTC.

Time Nick Message
00:02 ispy__ joined #mojo
00:04 ispy_ left #mojo
00:21 mattastrophe joined #mojo
00:24 perlite left #mojo
00:27 tempire I'm still waiting on my dotcloud beta invite
00:27 tempire :(
00:38 perlite joined #mojo
00:43 Akron joined #mojo
00:47 mattastrophe left #mojo
00:53 Akron left #mojo
01:01 mattastrophe joined #mojo
01:11 tardisx tempire: how long ago did you sign up?
01:13 tempire apr 28
01:13 fhelmber_ joined #mojo
01:19 gizzlon left #mojo
01:19 gizzlon joined #mojo
01:20 abra left #mojo
01:20 omega left #mojo
01:20 abra joined #mojo
01:20 omega joined #mojo
01:29 tardisx hmm, odd, I got mine maybe 72 hours after signup, a colleague here got his same day
01:29 tardisx sorry no, his was next day
01:30 tardisx maybe dotcloud likes the cuts of our jibs better :-)
01:31 * tempire coddles his jibs
01:44 j3nnn1 joined #mojo
02:00 miyagawa tempire: i can give you an invite
02:00 tempire \o/
02:03 vel joined #mojo
02:42 perlrocks Twitter: "I'm going to do what I can to try and convince the VMware CloudFoundry team to support perl, dancer, and mojolicious." --jbobbylopez http://twitter.com/jbobbylop​ez/status/66331872623476737
02:48 vel left #mojo
03:32 perlrocks Twitter: "Got a ticket saying Mojolicious app not work with POST on uWSGI - talked and fixed on both sides: http://bit.ly/jy1Oey http://bit.ly/lVoyQH" --miyagawa http://twitter.com/miyagaw​a/status/66344534036070400
03:51 zakame joined #mojo
03:51 zakame miyagawa: sri: thanks, that was my ticket :D
03:52 miyagawa zakame: yep :)
03:53 zakame actually I first saw it happen when I was trying to get tweetylicious running off dotcloud, as I couldn't register on my instance
04:13 Foxcool joined #mojo
04:16 Foxcool left #mojo
04:29 zakame left #mojo
04:41 jpablobr joined #mojo
04:51 j3nnn1 left #mojo
05:12 koban` joined #mojo
05:34 Foxcool joined #mojo
06:02 Christian joined #mojo
06:06 ysyrota joined #mojo
06:11 AmeliePoulain joined #mojo
06:16 Christian morning
06:20 marcus hai
06:24 koban` left #mojo
06:24 koban` joined #mojo
06:29 sri good morning fellow dinosaurs
06:30 jpablobr left #mojo
06:42 koban` left #mojo
06:42 koban` joined #mojo
06:45 koban` left #mojo
06:46 k-man morning sri
06:48 koban` joined #mojo
06:51 * yko waves
07:07 Foxcool left #mojo
07:12 GabrielVieira2 left #mojo
07:14 abra left #mojo
07:18 Foxcool joined #mojo
07:25 abra joined #mojo
07:25 Sugar joined #mojo
07:50 arthas joined #mojo
08:04 cyril joined #mojo
08:08 ajgb joined #mojo
08:09 perlrocks Twitter: "Etnogenez – Russian Literary Series uses Mojolicious v.1.3 http://www.etnogenez.ru" --sharifulin http://twitter.com/shariful​in/status/66414294719401984
08:10 koban` left #mojo
08:10 koban` joined #mojo
08:17 perlrocks Twitter: "New version of perl-Mojolicious packaged. # # Mandriva Mojolicious # perl" (fr) --Kharec http://twitter.com/Kharec/status/66416325404262400
08:20 perlrocks Twitter: "Fixed a macro. # # Mojolicious Mandriva" (fr) --Kharec http://twitter.com/Kharec/status/66417119037886464
08:21 GitHub159 joined #mojo
08:21 GitHub159 mojo: master Sebastian Riedel * bd177b3 (3 files in 3 dirs): answered new frequently asked question - http://bit.ly/j4nHgA
08:21 GitHub159 left #mojo
08:26 cosmincx joined #mojo
08:37 sri https://github.com/kraih/mojo/issues # opened some issues, in case there are interested volunteers
08:42 sri moving templates is a really tricky one
08:48 jwang left #mojo
08:56 perlrocks Twitter: "perl-1.300.0-1mdv2011.0 Mojolicious-uploaded in contrib / release, cooker: Mandriva http://bit.ly/imWY9k #" (fr) --Kharec http://twitter.com/Kharec/status/66426173751693312
09:00 Foxcool left #mojo
09:11 Foxcool joined #mojo
09:38 koban` left #mojo
09:56 zakame joined #mojo
09:59 spleenjack joined #mojo
10:05 PerlNinja joined #mojo
10:06 mateu left #mojo
10:13 koban` joined #mojo
10:17 sh4 joined #mojo
10:30 perlrocks Twitter: "I really wish @dot_cloud gave me an invitation, I tried flattery, I'll resort to bribes if I have to! I want my #mojolicious on @dot_cloud" --perlninja http://twitter.com/perlninj​a/status/66449835871580160
10:31 * sri pokes PerlNinja
10:32 sri PerlNinja: i think if you ask miyagawa nicely he will give you an invite code
10:33 sri http://hackerne.ws/item?id=2518546 # maybe one of these even still works
10:39 Foxcool left #mojo
10:42 koban` left #mojo
10:47 PerlNinja nah those won't - everyone's got them now hahaha
10:47 PerlNinja I'll wait and see :)
10:48 moritz sri: they didn't work 4 hours after it was posted
10:48 sri :o
10:50 Foxcool joined #mojo
11:05 Akron joined #mojo
11:10 Foxcool left #mojo
11:22 Foxcool joined #mojo
11:28 perlrocks Twitter: "@ Dot_cloud I'll use to deploy DotCloud Mojolicious awesome apps. Plus I'd like to say: you're incredible smart! Awesome! :)" (de) --memowe http://twitter.com/memowe/status/66464407558946816
11:52 zakame left #mojo
12:00 vel joined #mojo
12:00 vel left #mojo
12:07 perlrocks Twitter: "Working with query string / Mojolicious Tips http://dlvr.it/Qqrhr" (ja) --hiraba_reader http://twitter.com/hiraba_rea​der/status/66474083621150720
12:17 AmeliePoulain left #mojo
12:20 marcus left #mojo
12:22 marcus joined #mojo
12:28 marcus_ joined #mojo
12:28 marcus left #mojo
12:28 marcus_ is now known as marcus
12:29 koban` joined #mojo
12:31 AmeliePoulain joined #mojo
12:36 mateu joined #mojo
12:46 zakame joined #mojo
12:51 marcus left #mojo
12:58 Foxcool left #mojo
12:59 marcus_ joined #mojo
13:00 sri Akron: what would you do differently about xml mode?
13:00 marcus_ is now known as marcus
13:02 Akron I would decouple the SAX from the DOM and make the HTML treatment a filter for the DOM. For (more or less) strict XML parsing a total simple tree building filter would be applied. The decision on the filter could be made before parsing. Explicitely or with autodetection.
13:03 sri umm, how is that different from what we have now apart from implementation details?
13:04 Akron Yes - it's more or less implementation detail, but you would be able to completely decouple the html magic from the parsing.
13:05 Akron However - the ->xml flag is okay, but the treatment is not consequently enough I think.
13:07 sri hmm, now we are back to where we started
13:08 Akron I should'nt mention the "I don't know about ->xml" thing in the issue. ;)
13:08 sri i've been playing with a separate sax module btw. but it was only like 20 lines of code.... didn't seem worth it in the end
13:09 Akron Can I send you a suggestion for the module tonight?
13:09 sri what kind of suggestion?
13:09 marcus I like sax
13:09 sri i'm only really interested in solutions, not problems ;)
13:10 sri when you parse html sax kinda sucks... since you have all the special cases on top
13:10 Akron For Mojo::SAX. I experimented with it as well - it has optional namespace treating and only one "special" method called "skip" - for skipping process contents, so the DOM parser for HTML5 can skip script-section etc.
13:12 sri good example, i'm not even sure Mojo::DOM is finished yet, there can still be special html5 cases we have not covered yet
13:12 Akron I think a SAX parser would also be nice for things like the OpenGraph protocol.
13:12 Foxcool joined #mojo
13:12 Akron Yep. So a decoupled filter may be good - even for maintenance. But I love your css selectors - I definitely want to treat XML with it as well! ;)
13:13 sri selectors are in trouble if you go case sensitive
13:13 Akron The selectors aren't case sensitive atm?
13:14 sri everything is lowercase atm
13:14 sri internally
13:14 Akron Oh - I thought they were sensitive ...
13:14 sri by spec css selectors can't be case sensitive i think
13:15 sri right, spec forbids it
13:16 sri hmm, or not
13:16 sri http://www.w3.org/TR/css3-selectors/#casesens # bit confusing
13:16 sri seems it is up to us
13:17 Akron I think they mean the selectors are case insensitive - the tag names depend on the language. And ... well ... the namespace prefixes are underdefined I think. ;)
13:18 Akron But the current implementation seems to be okay with it - or not?
13:18 sri the biggest problem with optional case sensitivity will be the amount of new tests required
13:18 Akron I see no LCing of arguments.
13:19 sri at least a few hundred new unit tests
13:20 Akron Some old ones would break (the ->xml(1) ones) - but I see no real problem with it.
13:20 Akron DOM.t would also benefit from decoupling! ;)
13:20 metaperl_ left #mojo
13:20 sri how?
13:21 sri not like we are going to delete old test cases
13:22 Akron Hm - okay, yeah, you are right. It's not easy to separate the tests in filter, sax and selector tests ...
13:22 sri i still don't know what you mean with filter, and how it could work
13:23 Akron I will try to hack something down tonight.
13:23 sri didn't you promise the same a few months back? ;p
13:23 Akron Did I? Damn ...
13:23 sri i vaguely remember ok-ing a sax module back then
13:24 Akron Oh - yeah. Right. Okay. Tonight! Really!
13:24 Akron well - the SAX was ready then - but not the coupling with the DOM.
13:24 * sri cracks the whip
13:24 Akron \me greets the aligators.
13:24 sri the code has also changed a few times since then
13:25 sri html healing and stuff
13:25 Akron Yepp - but not the principally parts.
13:25 Akron I mean - I did not touch the healing.
13:26 sri well, script/style handling would atm be at the sax layer, i would imagine that makes things rather tricky
13:26 amoore joined #mojo
13:26 sri it's a huge problem with the html5 spec
13:27 sri it was never meant to be parsable with sax
13:27 sri style/script are defined as "raw" content
13:27 sri outside of any html/xml rules
13:29 sri the whole spec is written that way, exceptions over exceptions
13:34 sri interesting
13:35 mattastrophe left #mojo
13:40 dekimsey joined #mojo
13:51 GitHub198 joined #mojo
13:51 GitHub198 mojo: master Sebastian Riedel * 7015005 (3 files in 3 dirs): fixed case sensitivity of Mojo::DOM in XML mode - http://bit.ly/kOdZNG
13:51 GitHub198 left #mojo
13:51 sri Akron: it could be as easy as that
13:53 Akron sri: :) That was my purpose with the issue.
13:53 Akron sri++
13:54 sri attributes were not lowercased before, huge bug
13:54 sri bit surprised the selectors just work
13:55 Akron I was wondering why you said they don't - I saw no lc or /i anywhere.
13:55 sri i was so sure i used /i a lot
13:55 sri well... i'm almost 30 :,(
13:56 sri brain is not what it used to be
13:56 Akron script and style handling is in fact not possible with a simple SAX parser - that's why I implemented a "skip" method for that.
13:56 Akron :)
14:00 AmeliePoulain left #mojo
14:00 sri looking forward to your solution :)
14:01 Akron :)
14:02 * PerlNinja pushes sri into the alligator pit
14:02 sri ouch!
14:02 PerlNinja that's for making fun of us 30+ people :P
14:02 * PerlNinja has 29 days to go for 34 :(
14:02 sri the elderly?
14:02 purl the elderly is, like, Mhoram's favorite I think, kinda like daughter i'd say, but try it, and try yellow ledbetter, if still no go, give it up
14:02 Akron Oh - in the pseudoclass section of _match_selector you could use some lc and /i to be up to the spec.
14:04 Sugar|2 joined #mojo
14:06 Akron https://github.com/kraih/mojo/b​lob/master/lib/Mojo/DOM.pm#L743 seems to be a good line for my $class = lc($c->[1]);
14:09 Sugar left #mojo
14:10 sri makes sense i guess
14:12 Foxcool left #mojo
14:15 amoore left #mojo
14:15 sri Akron: that means "send a pull request when you;re done" ;)
14:15 perlrocks Twitter: "@ Yurigaoka_man implementation Perl + SQLite. Mojolicious framework, CGI heavy, so the best. I want to run fast Perl is a cloud service." (ja) --flo_jack http://twitter.com/flo_jac​k/status/66506408501321728
14:16 Akron Hey - forked mojo on GitHub yesterday! I am now totally in that thing ... sort of ... "pull request" you say ... okay ... whatever ... ;)
14:16 sri \o/
14:16 sri http://help.github.com/pull-requests/
14:17 sri just don't forget the tests!
14:17 * sri cracks the whip
14:17 Akron :)
14:18 Akron Okay - I will try to do a pull request on this lc thing now ... hmm ...
14:18 sri we are targeting 6000 unit tests for summer
14:20 koban` left #mojo
14:21 koban` joined #mojo
14:21 amoore joined #mojo
14:22 koban` left #mojo
14:23 koban` joined #mojo
14:24 mattastrophe joined #mojo
14:25 koban` left #mojo
14:30 Christian left #mojo
14:44 PerlNinja left #mojo
14:56 Akron Okay ... first pull request type thing tested ... fingers crossed.
14:57 sri looks ok, but in the future please squash commits
14:57 sri now i can't use the green button to auto merge :/
14:58 Akron What's that? I was so proud of doing things I don't understand while having the feeling it works.
14:59 sri pull request was ok, you usually do it for single commits only though
14:59 ZadYree left #mojo
14:59 Akron Ah - okay. Can I merge commits afterwards?
15:00 Akron (just for this purpose I mean.)
15:00 sri not anymore
15:01 Akron Okay. sorry.
15:03 GitHub94 joined #mojo
15:03 GitHub94 mojo: master Akron * 1ee0754 (1 files in 1 dirs): Lowercasing pseudoclass descriptors
15:03 GitHub94 mojo: master Sebastian Riedel * 5b710f9 (2 files in 2 dirs): added selector tests
15:03 GitHub94 mojo: master commits 7015005...5b710f9 - http://bit.ly/liLzix
15:03 GitHub94 left #mojo
15:03 kaare joined #mojo
15:04 Akron \o/ Pull request ftw!
15:04 dekimsey left #mojo
15:10 metaperl joined #mojo
15:11 dekimsey joined #mojo
15:14 GitHub182 joined #mojo
15:14 GitHub182 mojo: master Sebastian Riedel * 561a013 (2 files in 2 dirs): fixed css equation case sensitivity bug - http://bit.ly/mecAur
15:14 GitHub182 left #mojo
15:15 sri phew
15:21 Akron Oh - I overlooked those.
15:26 tempire left #mojo
15:34 cyril left #mojo
15:48 Foxcool joined #mojo
15:57 ysyrota left #mojo
16:16 cyril joined #mojo
16:25 mattastrophe left #mojo
16:26 mattastrophe joined #mojo
16:28 ajgb left #mojo
16:37 sri http://www.cpantesters.org​/distro/M/Mojolicious.html # looks like we still have a problem with freebsd
16:37 sri sadly i can't replicate it
16:55 cyril left #mojo
16:55 cosmincx left #mojo
17:08 Akron How do I force overwrite my forked repo with the mother repo?
17:10 sri your repo is broken, just delete it
17:10 sri and in the future use feature branches for pull requests ;)
17:11 sri then you can keep master clean
17:14 Akron Okay - thanks. I found a new test case, I think.
17:16 Sugar|2 left #mojo
17:24 vel joined #mojo
17:48 cyril joined #mojo
17:50 xaka joined #mojo
17:57 fhelmber_ left #mojo
18:11 Akron Next try. ;)
18:12 sri Akron: are you sure that tiny test covers all cases?
18:13 cyril left #mojo
18:13 Akron It covers one embedding example and the xml test. Of course you can embed in some other ways - but I didn't want to pollute the testsuite.
18:14 cyril joined #mojo
18:14 sri no, i mean you changed 5 cases
18:14 Akron But if you want me to I can cover all ways of embedding.
18:14 sri are they all covered?
18:14 Akron No.
18:15 sri yea, changes without tests have no chance
18:15 Akron Okay - I will cover all.
18:16 sri please use perltidy too
18:16 sri we have strict formatting rules ;)
18:16 sri .perltidyrc is included
18:17 Akron Damn - this is all so complicated ... okay okay.
18:18 sri in the end it will make your life easier though :)
18:18 sri (and mine)
18:19 Akron Do I have to create a new branch from master to commit the last one + extended test cases?
18:19 Akron I mean - for your green button. ;)
18:20 sri you can make a second commit and squash both into one
18:20 sri git has this wonderful feature where you can rewrite history
18:20 sri http://gitready.com/advanced/2009/02/​10/squashing-commits-with-rebase.html
18:21 Akron Thanks.
18:33 Akron left #mojo
18:42 non joined #mojo
18:42 spleenjack left #mojo
18:56 snearch joined #mojo
18:59 ispy__ left #mojo
19:11 Akron joined #mojo
19:19 Foxcool left #mojo
19:20 Foxcool joined #mojo
19:35 tempire joined #mojo
19:52 marty joined #mojo
19:56 metaperl Is there some reason that my post '/proxy' action is not forwarding POST parameters to my post '/newkey' action? https://gist.github.com/959662
20:07 daviddelikat left #mojo
20:08 metaperl oh I see
20:09 metaperl $self->req->params() is returning a Mojo::Parameters class, which is a blessed arrayref
20:09 metaperl I need all the POST parameters back in a hashref
20:10 metaperl $self->req->body_params is what I need
20:11 metaperl my $param = $self->req->body_params->to_hash;
20:11 metaperl we had this discussion a few days ago
20:14 sh4 left #mojo
20:15 metaperl How do you set the Mojo::UserAgent to not timeout for a request expected to last 120 seconds
20:16 Foxcool left #mojo
20:16 sri first of all why do you use a custom useragent?
20:16 sri $self->ua is a readily prepared one
20:17 sri what you want is $self->ua->keep_alive_timeout(120);
20:19 Foxcool joined #mojo
20:22 metaperl sri - I just copied some code from the wiki. I didnt know there was a prepared one. I read the keep_alive_timeout docs, but it said keep alive _between_ requests, so I didnt think it pertained to _during_ a request
20:25 sri yea, i suppose that's a bit misleading
20:25 GitHub133 joined #mojo
20:25 GitHub133 mojo: master Sebastian Riedel * 7f33a54 (2 files in 1 dirs): improved timeout description - http://bit.ly/lJH8R3
20:25 GitHub133 left #mojo
20:29 snearch left #mojo
20:30 GitHub125 joined #mojo
20:30 GitHub125 mojo: master Sebastian Riedel * bcf9f72 (3 files in 2 dirs): updated more descriptions - http://bit.ly/kKfiep
20:30 GitHub125 left #mojo
20:32 GitHub145 joined #mojo
20:32 GitHub145 mojo: master Sebastian Riedel * 792a9f9 (1 files in 1 dirs): fixed typo - http://bit.ly/ivxhdc
20:32 GitHub145 left #mojo
20:32 sri phew
20:40 non left #mojo
20:41 kaare left #mojo
20:45 GitHub111 joined #mojo
20:45 GitHub111 mojo: master Akron * 5472f3c (2 files in 2 dirs): Fixed a small xml flag bug
20:45 GitHub111 mojo: master Sebastian Riedel * 453df53 (2 files in 2 dirs): Merge pull request #141 from Akron/5472f3c62236bd7f62bba28fe8c564ac860b231d ...
20:45 GitHub111 mojo: master commits 792a9f9...453df53 - http://bit.ly/j1zYCS
20:45 GitHub111 left #mojo
20:46 sri yay for the green button
20:46 Akron This commit merging is voodoo!
20:47 sri and perfectly safe :)
20:47 sri since git has two histories, the public and the private one
20:48 perlrocks Twitter: "Mojolicious deprecating 5.8 support? Good for them! http://cpansearch.perl.org/src​/KRAIH/Mojolicious-1.3/Changes #perl" --chromatic_x http://twitter.com/chromatic​_x/status/66605201976602624
20:48 sri you just change the public one
20:48 Akron Ah ... okay. GitHub has the private one for my eyes only?
20:49 sri the private one is mostly invisible
20:49 sri you would have to dig deep
20:49 Akron Cool.
20:50 Akron The last test is really tricky. In fact - it's wrong. But it's better than totally wrong I guess.
20:51 DaTa sri: i can reproduce the failing tests on freebsd
20:51 Akron It could be fixed however when checking if there's a replace in root in advance - before parsing. But I don't know if there's a reason you didn't do that.
20:52 sri DaTa: yay... and damn!
20:53 sri DaTa: i don't even know where to start looking for a bug
20:56 * DaTa too :)
21:00 sri i have a hunch that this bug hides somewhere in Mojo::IOLoop
21:00 sri possibly around the write or read code
21:01 vel left #mojo
21:02 sri a previous bug showed up because the kernel send buffer was smaller than usual on freebsd... so it could be something rather weird
21:04 DaTa it's not IO::Poll vs. IO::KQueue, it fails with both
21:04 sri you'll forever be my hero if you find it :)
21:05 dekimsey left #mojo
21:07 perlrocks Twitter: "Vuln: Mojolicious 'link_to helper' HTML Injection Vulnerability http://goo.gl/fb/VgCqX" --BackTrack5 http://twitter.com/BackTrac​k5/status/66610053481906176
21:09 amoore left #mojo
21:12 sri Akron: hmm, i'm not sure the last test is wrong
21:13 sri thing is, if you really wanted to start fresh with detection you could just use a fresh dom instance
21:14 Akron That's right - but atm the xml prolog is thrown away.
21:15 sri of course
21:15 Foxcool left #mojo
21:16 Akron Okay. So it's your decision. :) I think replacing intendedly an xml document with an html document (and not starting a new instance) is rather ridiculous.
21:17 sri what decision did i make?
21:17 sri it's not like there are choices really
21:17 Akron That the test is correct. :)
21:17 sri what is the other option then? :)
21:17 moritz sri++ # verbose render tracing
21:18 sri \o/
21:18 Akron Well - there is the choice to change the behaviour in replace to detect xml behaviour again.
21:18 moritz didn't see it before because I spent the last two days entirely on the backend
21:25 DaTa strange, some tests don't fail sometimes
21:36 GitHub102 joined #mojo
21:36 GitHub102 mojo: master Sebastian Riedel * 87eb1e7 (2 files in 2 dirs): fixed XML detection in Mojo::DOM - http://bit.ly/k8K7fU
21:36 GitHub102 left #mojo
21:37 sri Akron: changed it a bit
21:37 sri not sure which solution is better though
21:39 Akron This one is better.
21:42 sri i'm really undecided
21:42 fmerges joined #mojo
21:43 sri since i can totally see people replace the root to inherit the xml flag and charset
21:44 Akron In that case I think they also expect to inherit the prologue.
21:53 GitHub13 joined #mojo
21:53 GitHub13 mojo: master Sebastian Riedel * 8918a6a (1 files in 1 dirs): micro optimization - http://bit.ly/lE11y7
21:53 GitHub13 left #mojo
22:22 marty left #mojo
22:27 metaperl left #mojo
22:40 vel joined #mojo
22:52 metaperl joined #mojo
23:18 baton8 left #mojo
23:18 baton8 joined #mojo
23:22 jwang joined #mojo

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