Camelia, the Perl 6 bug

IRC log for #mojo, 2012-05-03

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

All times shown according to UTC.

Time Nick Message
00:02 GitHub101 joined #mojo
00:02 GitHub101 [mojo] kraih pushed 1 new commit to master: http://git.io/Aa5GYg
00:02 GitHub101 [mojo/master] improved 32bit Perl support of Mojo::Transaction::WebSocket - Sebastian Riedel
00:02 GitHub101 left #mojo
00:02 sri psimanx1: that appears to be a harmless change
00:03 psimanx1 that works for me
00:05 jmmills joined #mojo
00:05 psimanx1 sri++  :)
00:06 sri psimanx1++
00:06 sri don't let this stop you from searching for a proper solution :)
00:08 psimanx1 Nope - Just been playing around with some code on 32 bit that is handling the number 90 quadrillion (9000000000000000).  Anything much bigger than that it can't handle.
00:09 psimanx1 That's 9 petabytes btw - quite a big payload for a websocket... :)
00:10 psimanx1 Would be a bit hard to test however...
00:11 sri you say that now... just wait another century...
00:11 psimanx1 They'll have microsd cards that size by then...
00:14 jnap joined #mojo
00:15 GitHub117 joined #mojo
00:15 GitHub117 [mojo] kraih pushed 1 new commit to master: http://git.io/hW1SkQ
00:15 GitHub117 [mojo/master] improved 64bit detection slightly - Sebastian Riedel
00:15 GitHub117 left #mojo
00:23 sri 64bit does seem a bit pointless, considering that you can just send multiple frames
00:24 psimanx1 Yeah - a colleague suggested I do that earlier - wished I coded it up now.. :)
00:24 psimanx1 Thought it would be easier to get this working.. ;)
00:35 psimanx1 sri: would it be acceptable to write tests that take various numbers as the input and then put them through the same code and make sure the same numbers come out correct the other side?  Or do the tests need to reflect actual use cases (which is going to be as close to impossible to achieve as you can get...)
00:36 sri i think many of my test cases currently look like ('hi' x 20000)... go figure :)
00:37 psimanx1 So it would be acceptable to replicate the code in a test and make sure it's working there, with the logic that it will also then work when used in actuality?
00:38 sri you mean cut-n-paste code from Mojo::Transaction::WebSocket to a test?
00:38 psimanx1 Yeah - the bits that deal with converting to a quad and then back again - test the 64 bit and 32 bit and make sure they accept, encode and decode the input/output correctly.
00:39 sri hmm, i don't like that
00:39 psimanx1 That's the only way I can see of properly testing.  I know it's not ideal - but it's not going to be practical to create a 2 petabyte payload... :)
00:40 psimanx1 Or put em into methods and call them during the test - to avoid the cut and paste
00:40 sri what are you trying to fix atm?
00:41 sri actually, i'm not sure this is worth fixing
00:41 psimanx1 Well - I basically have the length of the payload accepting up to 9 petabytes - anything more than that in 32 bit perl and it dies with an error.
00:41 psimanx1 I can leave it if you're happy with a 4Gb limit.
00:41 sri once 2 petabyte payloads become usable we won't have to worry about 32bit perl :)
00:42 sri 4gb is still a lot for one message
00:42 psimanx1 Very true.  And I think 32 bit perl runs into issues in 2038 anyways...
00:42 sri i'll just document the limitation for now
00:42 psimanx1 Great - sounds like a plan!
00:51 GitHub197 joined #mojo
00:51 GitHub197 [mojo] kraih pushed 1 new commit to master: http://git.io/UaS6Rw
00:51 GitHub197 [mojo/master] mention 32bit limit - Sebastian Riedel
00:51 GitHub197 left #mojo
01:02 lukep joined #mojo
01:06 scott joined #mojo
01:33 preflex_ joined #mojo
01:36 GitHub65 joined #mojo
01:36 GitHub65 [mojo] kraih pushed 1 new commit to master: http://git.io/d_MmSw
01:36 GitHub65 [mojo/master] fixed empty frame handling in Mojo::Transaction::WebSocket - Sebastian Riedel
01:36 GitHub65 left #mojo
01:54 mire joined #mojo
01:59 MichaelRpdx joined #mojo
02:00 GitHub13 joined #mojo
02:00 GitHub13 [mojo] kraih pushed 1 new commit to master: http://git.io/CjUyrg
02:00 GitHub13 [mojo/master] more WebSocket tests - Sebastian Riedel
02:00 GitHub13 left #mojo
02:06 xaka1 joined #mojo
02:08 scott_ joined #mojo
02:30 michale joined #mojo
03:01 lukep joined #mojo
03:20 lukep joined #mojo
03:32 lukep joined #mojo
03:32 scott joined #mojo
03:49 scott_ joined #mojo
04:07 tempire http://www.bloc.io/ <- thinking about the design course
04:10 tempire oh, $3k
04:10 tempire nevermind
04:18 lukep joined #mojo
04:21 lukep joined #mojo
04:42 ispy_ joined #mojo
04:53 AmeliePoulain joined #mojo
04:58 rem_lex joined #mojo
05:12 mattastrophe joined #mojo
05:17 jwang joined #mojo
05:52 Foxcool joined #mojo
06:07 Vandal joined #mojo
06:22 Foxcool joined #mojo
06:24 spleenjack joined #mojo
06:37 marcus :D
06:38 marcus I guess it's not *that* expensive compared to many courses.
06:51 lukep joined #mojo
07:11 batman joined #mojo
07:12 kanishka joined #mojo
07:13 tempire it's expensive compared to the internet, though.
07:13 tempire then again, I'm not sure how to learn "design"
07:13 tempire but I suspect I could figure it out for less than 3k
07:14 tempire 2995, max!
07:18 d4rkie joined #mojo
07:19 ladnaV joined #mojo
07:25 marcus sweet
07:27 fhelmber_ joined #mojo
07:29 Htbaa joined #mojo
07:31 spleenjack joined #mojo
07:32 taryk joined #mojo
07:44 cosmincx joined #mojo
07:53 cosmincx joined #mojo
07:58 Vandal joined #mojo
08:29 Htbaa1 joined #mojo
08:36 cosimo joined #mojo
08:46 bird joined #mojo
09:26 bird_ joined #mojo
09:26 Foxcool joined #mojo
10:04 michale joined #mojo
10:20 d4rkie joined #mojo
10:24 d4rkie joined #mojo
10:33 Akron joined #mojo
11:28 amirite can i use test::mojo in non-mojo apps
11:31 leon joined #mojo
11:34 amirite wtf is DNT => 1
11:35 psimanx1 amirite: http://dnt.mozilla.org/ ?
11:36 amirite no, from the documentation
11:36 psimanx1 ahh k
11:36 amirite $t = $t->put_ok('/foo' => {DNT => 1} => 'Hi!');
11:37 psimanx1 See Mojo::Headers
11:38 amirite i wish the documentation would elude to things like that
11:38 amirite shitty ass documentation
11:38 psimanx1 grep is your friend.. :)
11:38 stephan48 dnt ist just a random header
11:39 Akron amirite: It's exactly what psimanx1 says. http://tools.ietf.org/id/dra​ft-mayer-do-not-track-00.txt
11:44 amirite how do i prevent an inactivity timeout
11:44 daxim s/elude/elucidate/
11:45 athenot joined #mojo
11:46 * amirite poops on mojolicious
11:48 Akron amirite: What do you mean with "prevent"?
11:54 GabrielVieira joined #mojo
12:20 espent joined #mojo
12:49 nic Who should be credited as author(s) for the Mojolicious documentation?
12:49 nic I don't see any credits within the PODs themselves
12:50 nic Should it be "Sebastian Riedel" (as if such a person really existed) or "Mojolicious Cast" or "Various Artists" or a list or what?
12:54 tempire nic: it's primarily sri
12:54 tempire was written under a perl foundation grant
12:54 nic Oh, perhaps that needs to be mentioned too?
12:55 tempire if it's relevant
12:55 nic (In that scenario in academia you usually need to credit the grant body)
12:55 nic Thanks
12:55 tempire actually, it's only kind of relevant
12:55 tempire there's quite a bit of work that's been done since that time
12:56 tempire refinements and reorgs and such
12:57 nic I've just been enjoying Mojocast #3 again
12:57 nic I love how upbeat they are :)
12:57 tempire \o
12:57 asarch joined #mojo
13:06 tempire I'm happy to see from the google analytics that the mojocasts are being used as reference tools
13:06 tempire as opposed to watch-once-and-forget
13:09 sri \o/
13:10 nic They're nice and short and snappy and the signal to noise ratio is very high
13:10 nic Not like an oracle tutorial where you have to drag through 20 mins to get to something useful
13:11 * sri sneezes
13:11 tempire I've been thinking about having some fast track ones
13:11 tempire sort of like this style: http://library.creativecow.net/articles/mc​auliffe_kevin/Motion-5_2-1/video-tutorial
13:11 tempire they would be for more advanced topics
13:12 sri don't forget to update relaxed plaxeholders, the new ones appear to be successful :o
13:13 * tempire curses sri
13:13 * sri curses back
13:15 * tempire ncurses sri
13:15 * sri flags tempire
13:15 * tempire interrupts sri
13:16 * sri throws signals
13:17 nic "You know, 'L' as in Replicator"
13:18 tempire futurama is upon us: http://www.daytonaelevator.com/Pneumatic​%20Vacuum%20Elevator%20Main%20Page2.htm
13:18 sri WANT!
13:20 nic "Your sentence is to be placed inside a sealed tube at which time one of the seals will rupture..."
13:22 nic "If the upper seal ruptures, your air pressure will instantly plummet till your insides splat against the clear walls..."
13:23 nic "If however the lower seal ruptures, your air pressure will instantly rocket till your insides splat against the clear walls..."
13:23 nic "And nobody will know in advance which seal will be ruptured."
13:25 GitHub185 joined #mojo
13:25 GitHub185 [mojo] kraih pushed 2 new commits to master: http://git.io/ZNz8cQ
13:25 GitHub185 [mojo/master] Bugfix end.* helpers - Nils Diewald
13:25 GitHub185 [mojo/master] Merge pull request #325 from Akron/endpoints - Sebastian Riedel
13:25 GitHub185 left #mojo
13:26 tempire Akron++
13:26 jnap joined #mojo
13:26 sri the test case is a little underwhelming, but that can be fixed
13:39 GitHub64 joined #mojo
13:39 GitHub64 [mojo] kraih pushed 1 new commit to master: http://git.io/55FGFA
13:39 GitHub64 [mojo/master] better helper tests - Sebastian Riedel
13:39 GitHub64 left #mojo
13:39 sri even found and unused token :)
13:39 sri s/d//
13:51 Htbaa joined #mojo
14:01 inokenty joined #mojo
14:06 nic In my menubar.html.ep I have "<%= $self->stash('user') %>" which works
14:07 nic I thought that "<%= $user %>" would also work, but it complains that $user needs to be qualified
14:08 inokenty You can use <%= stash 'user' %>
14:08 nic ah, thanks, yep, that's an improvement
14:09 inokenty <%= $user %> probably doesn't work because of 'use strict' in templates
14:10 inokenty And $user is not passed to template render
14:10 nic I see.  Mojolicious::Guides::Rendering might need some clarity on that
14:28 sri no, what you said makes no sense
14:28 sri if it's in the stash it gets a variable
14:29 sri if you pass it to render or not is irrelevant
14:29 sri your problem is something you did not mention
14:36 jnap joined #mojo
14:40 ispy_ joined #mojo
14:52 GitHub12 joined #mojo
14:52 GitHub12 [mojo] kraih pushed 1 new commit to master: http://git.io/-55RoA
14:52 GitHub12 [mojo/master] fixed bug that prevented helper names from ending with "begin" - Sebastian Riedel
14:52 GitHub12 left #mojo
15:02 GitHub172 joined #mojo
15:02 GitHub172 [mojo] kraih pushed 1 new commit to master: http://git.io/qZATtA
15:02 GitHub172 [mojo/master] better helper tests - Sebastian Riedel
15:02 GitHub172 left #mojo
15:03 andrefs joined #mojo
15:05 andrefs joined #mojo
15:08 Akron joined #mojo
15:09 Akron sri: Oh - I didn't check for .*begin. Nice!
15:12 sri :)
15:18 marty Mojolicious is AMAZING!  sri++
15:18 sri \o/
15:21 xaka1 joined #mojo
15:50 mire joined #mojo
16:16 batman joined #mojo
16:19 d4rkie joined #mojo
16:21 D4RK-PH0ENiX joined #mojo
16:42 Patterner joined #mojo
17:23 * sri looks around
17:23 * sri yawns
17:27 whee one day.
17:27 purl i think one day is 60 * 60 * 24. or 86400 or 60 * 60 * 24 * x or 1440 minutes
17:28 mire joined #mojo
17:31 sri http://26.media.tumblr.com/tumb​lr_lrb5erpfBw1qzrlhgo1_500.jpg
17:48 mire joined #mojo
17:50 batman i would like to specify a custom route object when loading PODRenderer, instead of attaching it to $app->routes->any
17:51 batman would you mind adding that sri?
17:52 sri batman: yes, i would mind
17:52 batman would you mind taking a patch for it? :)
17:52 sri yes, i would mind that too
17:52 batman not my day, i guess :)
17:52 batman hehe
17:53 sri first you would have to convince me that it is absolutely necessary and that the majority of our users benefits from it
17:54 batman i guess nothing is necessary. it's just that i want to add the endpoint to a route with authentication
17:55 batman like /admin/documentation instead of just /perldoc
17:55 sri that plugin is not a general purpose documentation browser
17:56 sri i believe it has been forked a few times already for that reason
17:57 batman then i will do the same
17:58 sri not that what you want requires it
17:58 batman sorry. i don't understand that sentence
17:58 sri you could just relocate the route
17:58 batman i can?
17:59 batman how?
17:59 sri yes, hack the route
18:00 sri my $route = $r->find('whatever_name');, remove it from its parent, and reattach it to another one
18:00 batman that's aweoms
18:00 batman *awesome :)
18:00 sri few weeks ago i wanted to make that easier, but nobody thought it was a good idea
18:01 batman i'm not sure if it's a required idea, but it's cool :)
18:01 sri like my $route->find('some_name')->detach;
18:01 sri umm
18:01 sri my $route = $r->...
18:02 sri and $r->route('/some/prefix')->add_child($route);
18:02 batman so how do i remove it from the parent?
18:03 sri the hard way ;p
18:03 sri which i'm afraid, you'll have to find out yourself
18:03 batman @{ $self->children } = grep { $_ ne $r } @{ $self->children }; ?
18:05 batman sounds freaky that one route can be the child of two parents...
18:06 batman maybe insteead of detach() you should just make it auto-detach if a route gets a new parent
18:06 batman like in DOM
18:06 sri it can't really
18:06 sri chilren have a link back to their one parent
18:06 Htbaa joined #mojo
18:06 sri multiple parents only goes one way
18:06 batman hm... then it SHOULD have auto detach functionality
18:07 sri ?
18:07 batman $r = $routes->find(...); $r2 = $routes->route("/foo"); $r2->add_child($r);
18:07 batman ^ i think both the first $routes and the second $r2 object will have $r as child
18:07 batman or..?
18:08 sri only the second will truly be the parent
18:08 sri first is pretty much broken
18:08 batman yes.
18:08 sri this is not documented core functionality
18:08 sri you're just hacking around
18:09 batman that's my point .when you add a $route object to a new $parent_route, then $previous_route_parent should get the child auto removed
18:09 batman but it's perl! :)
18:09 sri i proposed that a few weeks ago
18:09 batman nothing is "hacking" in perl. it's just perling ;)
18:09 batman go for it!
18:09 sri but if i remember correctly nobody cared
18:10 sri batman: get a core dev to agree and i'll make it happen ;p
18:10 * sri is a little more reluctant this time
18:10 batman marcus: do what i say or i won't talk to sjn about the hackathon!
18:10 batman :D
18:10 sri :D
18:11 batman or...sorry... that wasn't right...
18:11 batman marcus: agree to what sri suggested or i won't talk to sjn about the hackathon!
18:11 batman sri: there might be another issue attached to this solution though...
18:12 batman maybe the code base will get over 10k of lines :/
18:12 tempire batman: are you in oslo.pm?
18:12 batman occasionally
18:12 purl occasionally is a strong word for it
18:12 batman purl is righ :P
18:12 batman tempire^
18:13 tempire moar vikings!
18:14 batman :)
19:00 andrefs joined #mojo
19:03 sri batman: even found my old implementation in a branch
19:04 batman ok?
19:04 sri so it could definitely happen if there is interest
19:18 batman nice
19:23 sri here's the full diff from my branch http://hastebin.com/raw/suyugiqado
19:23 * sri still wants marcus, tempire or crab to agree with the change
19:23 rem_lex|pivo joined #mojo
19:25 sri side effects should be minimal, but i'm not sure if this is the direction we want to go with routes
19:33 fibo joined #mojo
19:38 koban joined #mojo
19:42 GitHub29 joined #mojo
19:42 GitHub29 [mojo] kraih pushed 1 new commit to master: http://git.io/e7gKuw
19:42 GitHub29 [mojo/master] added remove method to Mojolicious::Routes::Route - Sebastian Riedel
19:42 GitHub29 left #mojo
19:43 sri as usual, it will get removed if there's no feedback ;p
19:43 sri batman: but now you can at least play with it
19:44 sri it also does $r->add_route($r->find('some_name'));
19:44 sri (automatic removal from old parent)
19:45 sri so $r->route('/some/prefix')->ad​d_route($r->find('perldoc')); should just work
19:45 batman are you sure you want to set the ->name() ?
19:46 sri why?
19:46 batman just curious if it could come in conflict of other routes...
19:46 sri first one always wins
19:46 sri if you don't like a name, rename it
19:46 batman ok
19:46 sri $r->find('perldoc')->name('something_else');
19:47 batman btw: i don't mind
19:47 sri you could just as well ask if /perldoc/*module would conflict with other routes
19:47 batman true
19:49 batman what about detach() instead of remove() ?
19:49 sri please elaborate
19:50 batman i think $route->detach means detach me from parent, while remove() is... not the same
19:50 sri i don't really care and am fine with both
19:50 batman or maybe ->parent(undef) could do it, so you don't need another method
19:50 sri but it is a removal
19:51 sri if you don't reattach the route it is lost
19:51 sri no, ->parent(undef) would not do it
19:51 batman but it could
19:51 purl if you patch it
19:52 sri $r->find('perldoc')->parent(undef) is better than $r->find('perldoc')->detach?
19:52 batman i think i would keep the remove() internal for now
19:52 batman no. it's just one method less :) i like as few methods as possible
19:53 sri you're doing a good job of convincing me to remove the whole feature :)
19:53 batman i think i would put the remove() code inside add_child()
19:53 batman just giving you feedback :)
19:53 tempire I really should have priced out plane tickets before submitting a yapc talk :(
19:53 sri but then it couldn't be used to remove routes
19:53 * tempire curses small town venues
19:54 batman sri. true. but when do you want to remove routes?
19:54 sri maybe you have a plugin that defined multiple routes and you just want to get rid of one
19:54 batman because of dispatch speed?
19:54 batman ok. i'm sold :)
19:55 batman sri: commit, f5, compile! :)
19:59 GitHub92 joined #mojo
19:59 GitHub92 [mojo] kraih pushed 1 new commit to master: http://git.io/AFBTtA
19:59 GitHub92 [mojo/master] renamed remove method to detach - Sebastian Riedel
19:59 GitHub92 left #mojo
19:59 batman me like
20:04 Gedge joined #mojo
20:06 GitHub26 joined #mojo
20:06 GitHub26 [mojo] kraih pushed 1 new commit to master: http://git.io/x8s60g
20:06 GitHub26 [mojo/master] removed route name from documentation browser again - Sebastian Riedel
20:06 GitHub26 left #mojo
20:08 mire joined #mojo
20:09 * sri bets marcus is busy with rubymotion
20:27 Htbaa1 joined #mojo
20:32 xaka1 joined #mojo
20:36 GitHub26 joined #mojo
20:36 GitHub26 [mojo] kraih pushed 1 new commit to master: http://git.io/J9YRpA
20:36 GitHub26 [mojo/master] documentation tweaks - Sebastian Riedel
20:36 GitHub26 left #mojo
20:49 Gedge_ joined #mojo
20:49 GitHub129 joined #mojo
20:49 GitHub129 [mojo] kraih pushed 1 new commit to master: http://git.io/f_6EMA
20:49 GitHub129 [mojo/master] better examples for reattaching routes - Sebastian Riedel
20:49 GitHub129 left #mojo
20:49 ashleydev joined #mojo
21:00 diegok tempire: you should have bought your tickets sooner, don't blame small towns :-p
21:01 tempire I don't think the ticket prices don't change much before month in
21:01 tempire it's just the regional airports that kill it
21:01 diegok tempire: but, yeah, I'm always tempted to assist a yapc::na from europe and this year, being there was a dream stopper :)
21:03 diegok tempire: but it sold out!, what's the logic in it?, a talk on mojolicious!? :)
21:03 diegok sri: I like detach (after reading why it can be useful on the backlog) ;)
21:22 perlite joined #mojo
21:25 sri i'm still unsure
21:26 sri but giving plugins more freedom with routes might be a good thing
21:27 sri you don't have to go crazy with plugin options, but can instead just use sensible defaults and let the user reattach if necessary
21:29 diegok sri: sure. Still unsure about 'detach' vs 'remove' as first one can be confusing for someone coming from catalyst...
21:29 sri remove could be confusing for someone coming from Mojo::IOLoop ;p
21:30 diegok sri: and, remove tells more about your purpose of removing plugin routes...
21:30 diegok eheh
21:30 diegok sure :)
21:30 diegok but, you know. I'm not the one for naming :(
21:32 sri well, the list of possible names is basically endless
21:33 sri remove, detach, eliminate, dislodge, extract, unplug clear, take_out, expurgate, strip_off...
21:34 sri deduct, withdraw, subtract, flush, pull, move_away...
21:34 diegok oh my!
21:34 diegok unplug! :p
21:35 sri http://www.youtube.com/watch?v=Vsapab5LwcY
21:49 lukep joined #mojo
22:07 jnap_ joined #mojo
22:25 GitHub61 joined #mojo
22:25 GitHub61 [mojo] kraih pushed 1 new commit to master: http://git.io/rPwS5A
22:25 GitHub61 [mojo/master] better plugin examples - Sebastian Riedel
22:25 GitHub61 left #mojo
22:25 sri tempire: remove or detach?
22:29 andrefs joined #mojo
22:30 jnap joined #mojo
22:32 andrefs joined #mojo
22:38 Gedge joined #mojo
22:58 Gedge joined #mojo
23:00 tempire what are we talking about?
23:01 tempire oh
23:01 tempire nay on detach
23:01 sri :o
23:01 sri as a whole or just the name?
23:01 tempire name
23:02 tempire detach is the first thing catalyst people look for
23:02 * sri nods
23:02 sri which name do you like?
23:02 tempire ->annihilate
23:03 tempire remove works
23:03 sri i'm ok with that
23:03 sri annihilate it is
23:03 tempire sweet
23:09 Gedge joined #mojo
23:09 GitHub22 joined #mojo
23:09 GitHub22 [mojo] kraih pushed 1 new commit to master: http://git.io/UDx5Wg
23:09 GitHub22 [mojo/master] renamed detach method to remove - Sebastian Riedel
23:09 GitHub22 left #mojo
23:14 Gedge joined #mojo
23:15 diegok ^ eheh
23:41 GitHub44 joined #mojo
23:41 GitHub44 [mojo] kraih pushed 1 new commit to master: http://git.io/Betd9Q
23:41 GitHub44 [mojo/master] fixed small rendering bug in Mojolicious::Plugin::PODRenderer - Sebastian Riedel
23:41 GitHub44 left #mojo

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