Camelia, the Perl 6 bug

IRC log for #mojo, 2011-12-28

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

All times shown according to UTC.

Time Nick Message
00:00 Quai tempire: I switched to Plack, but im still seeing the same thing. Ill try to debug it a bit more myself. Just wondering if this is something someone had seen before
00:00 tempire what about hypnotoad directly?
00:05 Quai im not able to reproduce it directly under hypnotoad
00:05 Quai strange
00:06 jnap joined #mojo
00:06 bvierra left #mojo
00:15 alnewkirk joined #mojo
00:30 jnap joined #mojo
00:34 GitHub28 joined #mojo
00:34 GitHub28 [mojo] kraih pushed 1 new commit to master: http://git.io/1ktTRQ
00:34 GitHub28 [mojo/master] improved session expiration by allowing it to be disabled - Sebastian Riedel
00:34 GitHub28 left #mojo
00:42 tempire huh.  penny arcade is down.  the drama continues!
00:42 * sri blames reddit
00:44 alnewkirk tempire: wtf, one comment as of yesterday on mojocast?
00:44 alnewkirk GREAT episode
00:44 tempire :(
00:44 * tempire fades away into obscurity
00:45 tempire thanks, though :)
00:46 tempire makes me wonder who's watching.  I presume people just don't have anything to say.
00:47 sri number of watcher not lower than usual?
00:47 sri *+s
00:47 sri it's the worst week of the year
00:47 tempire Number of views is actually up from last episode.
00:48 sri Oo
00:48 tempire But views doesn't really say much, I don't know who's really watching at the moment.
00:48 tempire could be because the last episode was about sessions, probably the most boring topic of them all
00:48 sri but it's true, i can't think of things to comment anymore either
00:49 sri the screencasts are just great
00:49 tempire I'm not a commenter myself, so I can't knock people for staying silent
00:49 tempire next episode will be the premiere episode on final cut
00:49 tempire and wide screen
00:49 sri :o
00:49 tempire this one took FOREVER just trying to get around iMovie's limitations
00:50 tempire and final cut has a 30 free trial now
00:50 tempire so I have 30 days to finish the episode :)
00:50 sri tick tock
00:52 tempire not sure what to do it on
00:52 tempire testing or database connectivity
00:54 sri websockets and events :)
00:54 alnewkirk yup, what he said
00:54 sri only problem, websockets are still experimental
00:55 alnewkirk focus on the $#!+ devs aren't already fluent with
00:55 tempire are web sockets working in every browser yet?
00:56 alnewkirk I'd like to see a screencast on it
00:56 tempire the goal is to get everyone up to speed, though.  sessions, stash, and flash, for example, are super simple, but if a new web developer can be pointed to mojocasts and be up to speed in 60 minutes, that's super valuable
00:57 tempire I don't have the tools yet to do a proper explanation of events.
00:57 sri firefox and chrome
00:57 purl firefox and chrome are already using the latest
00:57 * sri pats purl
00:57 purl don't mess with me
00:57 alnewkirk tempire++
00:57 tempire going to require some hip graphics and animation
00:57 tempire that will take a couple rounds with final cut & motion, I think
00:58 sri i think websockets are the key to explaining events
00:59 alnewkirk MongoDBI is rounding out (sorta), I think putting together a simple yet viable app would be awesome
00:59 alnewkirk shameless plug
00:59 tempire :)
01:02 foomator joined #mojo
01:07 foomator joined #mojo
01:22 sri really wonder who else requested 0 sessions
01:23 sri could have been marcus and diegok
01:41 mire__ joined #mojo
01:41 jnap joined #mojo
01:41 xaka 0 means till user close browser's window?
01:52 foomator joined #mojo
01:52 jnap joined #mojo
02:05 tempire the vim casts guy is releasing a book, and is going to run a series of traveling vim cast workshops
02:10 jnap joined #mojo
02:17 mire__ joined #mojo
02:19 cfedde vim cast?  me googles
02:25 d4rkie joined #mojo
02:33 d4rkie joined #mojo
02:46 marty If you mac guys were going to buy a macbook for your "not-so-computer-smart" friend.  Would you buy a macbook pro or air?
02:48 tempire air
02:48 marty Or maybe you guys do not even use the macbooks and instead use desktops?  dunno
02:48 marty Ok.   Thanks, tempire
02:48 tempire assuming they don't make heavy use of dvds
02:49 marty Nope, just some surfing and word processing and such.  Occassional light image editing?
02:49 tempire air is fine
02:49 Debolaz marty: It really depends on use. And by use, I don't mean what you do on the laptop, but where its likely to be put.
02:49 Debolaz marty: If its to act as a moveable desktop of some sort, then you want as large as possible.
02:49 marty ok.   what kind do you recommend for heavy uses like us?   (in case I am tempted away from linux and go over to the dark side)  :)
02:50 Debolaz marty: If one intends to actually carry it around, then as small as possible. Don't fall into the midway trap.
02:50 marty It's primarily going to be used on the couch and an occasionally in an office.
02:50 tempire air air air
02:51 Debolaz If its a desktop that can be moved on occasion, then a 17" macbook pro is a good choice. If you want to carry it around, go for the air.
02:51 marty I'm assuming we can hook an air up to a large monitor and keyboard if needed.
02:52 tempire with an adaptor (for the monitor), yes
02:53 Debolaz Personally, I have an iMac as desktop because I like the extra power. And just use a very cheap ASUS with Linux as my laptop.
02:53 Debolaz I'm very happy with this combination.
02:54 marty Thank you Mac guys.  I really appreciate the help.   Now, off to guy shop for a new Mac.   :)
03:16 hshong joined #mojo
03:38 bvierra joined #mojo
03:42 Psyche^ joined #mojo
04:17 mire__ joined #mojo
04:22 tempire interesting, regarding html5 video playback in firefox: http://support.mozilla.com/en-US​/questions/864137#answer-234133
04:58 wheelnotkiss joined #mojo
05:53 tempire lulz. THe hip new feature in c# is dynamics and try invocation, which is essentially dynamic function names and autoload
06:05 osprey joined #mojo
06:06 sugar joined #mojo
06:11 tempire hmm
06:11 tempire sri: heard of c# massive orm?
06:40 spleenjack joined #mojo
06:42 Vandal joined #mojo
07:22 koban joined #mojo
08:35 D4RK-PH0ENiX joined #mojo
08:58 judofyr joined #mojo
09:34 judofyr any reason why http://pastie.org/private/awlw1xfqgatgrod1kwkia gives me: Event "close" failed: Can't use an undefined value as a symbol reference at /Users/magnus/perl5/perlbrew/perls/perl-5.14.​1/bin/../lib/5.14.1/darwin-2level/IO/Poll.pm line 48.
09:40 Vandal maybe it because of wrong route declaration?
10:21 osprey left #mojo
10:23 judofyr_ joined #mojo
10:25 cstamas tempire: thats mobile firefox
10:26 cstamas tempire: mobile firefox was a joke half a year ago last time I checked
10:31 d4rkie joined #mojo
11:02 chansen joined #mojo
11:13 d4rkie joined #mojo
11:13 sri moin
11:20 judofyr_ can anyone tell me why this doesn't work? http://i.imgur.com/txgvz.png
11:21 judofyr_ I get a "Connection refused"
11:21 judofyr_ and sorry for the image
11:21 judofyr_ but I'm lazzzzzy
11:21 judofyr I tried to mimic test_server in UserAgent, but apparently I didn't do it right :/
11:21 judofyr sri: ^
11:24 judofyr damn ioloop
12:09 netfeed i think i read somewhere that you can inject stuff from the controller into the head-tag, but i can't find it anymore, is this possible?
12:19 sri content_for
12:19 purl hmmm... content_for is awesome
12:21 netfeed so i add a "%= content_for 'foo'" in my template and then use stash to put stuff there for it?
12:22 sri there's documentation for it in the tutorial and guides with examples
12:22 netfeed kk, thanks
12:40 non joined #mojo
13:42 koban joined #mojo
13:46 judofyr sri: got any ideas for http://i.imgur.com/txgvz.png?
13:46 judofyr it doesn't work :(
13:47 sri once someone starts swearing at mojolicious components i stop looking
13:47 judofyr poor ioloop
13:48 judofyr stuck with a single instance; no friends
13:49 judofyr I feel like there's just some simple timing issue
13:49 judofyr but if I add the $ua->get inside a Mojo::IOLoop->timer(…) it doesn't work either
13:50 * judofyr realizes he doesn't know how Mojo::UA works together with the IOLoop
13:51 judofyr and I can't telnet to it either :S
13:52 judofyr but if I do "$server->ioloop->start" right after prepare_ioloop
13:52 judofyr it works fine (in telnet)
13:54 judofyr but if I do "Mojo::IOLoop->start" *later* it doesn't work :S
13:58 jnap joined #mojo
13:59 judofyr sri: does Mojo::IOLoop->start behave different depending on which package you call it from? http://i.imgur.com/d5eq6.png :S
14:01 judofyr and by "not working", I mean I'll get: "telnet: Unable to connect to remote host"
14:11 crab hi from bangkok.
14:12 judofyr hi crab
14:13 * sri waves
14:18 spleenjack joined #mojo
14:26 marty Good morning
14:26 purl For you maybe.
14:39 abra_ joined #mojo
14:51 deniszh joined #mojo
14:58 sri latest cookie spec is rather interesting
14:58 sri removed a lot of the rfc 2965 garbage
14:58 sri i'm gonna deprecate version, port and comment
15:00 cstamas sri: can you give a link?
15:00 cstamas to the newest spec
15:00 sri looks like they realized that they can never ever get rid of the netscape spec, and just brought back expires -.-
15:00 sri http://tools.ietf.org/html/rfc6265
15:00 cstamas thx
15:01 sri it's very pragmatic
15:01 sri they removed everything they couldn't get people to adopt
15:02 crab following in the footsteps of html5? good.
15:03 sri yea, cookie2 is also dead, YAY!
15:08 al has anyone here gotten mojo websockets to work with iphone's safari?
15:10 hshong joined #mojo
15:12 GitHub96 joined #mojo
15:12 GitHub96 [mojo] kraih pushed 1 new commit to master: http://git.io/wQNKpg
15:12 GitHub96 [mojo/master] updated cookie implementation for RFC 6265 and in turn deprecated Mojo::Cookie->version, Mojo::Cookie::Response->comment and Mojo::Cookie::Response->port - Sebastian Riedel
15:12 GitHub96 left #mojo
15:12 amoore joined #mojo
15:13 sri not gonna mention RFCs though, since we still need quite a few real world hacks
15:14 abra joined #mojo
15:16 sri deprecating the attributes is also mostly a joke, if they are present or not is completely irrelevant
15:17 sri so i just made them do nothing :)
15:21 sri also interesting that HttpOnly made it into the spec
15:32 sri oh, path also needs to be moved from Mojo::Cookie to Mojo::Cookie::Response
15:37 inokenty joined #mojo
15:38 KayT3 joined #mojo
15:41 Andrei666 joined #mojo
15:41 Andrei666 happy holidays everybody :)
15:42 Andrei666 how can I route the  '/' to the public/index.html file in a new non_lite mojo app?
15:43 Andrei666 i tried with $r->route('/')->(public/index); but it doesn't work....that way...
15:46 tempire Andrei666: route to an action and use redirect_to
15:52 GitHub98 joined #mojo
15:52 GitHub98 [mojo] kraih pushed 2 new commits to master: http://git.io/0Ow0xA
15:52 GitHub98 [mojo/master] moved path attribute of Mojo::Cookie to Mojo::Cookie::Response for better RFC 6265 compliance - Sebastian Riedel
15:52 GitHub98 [mojo/master] updated Changes - Sebastian Riedel
15:52 GitHub98 left #mojo
15:53 sri would be interesting to know how many browsers and web frameworks actually support RFC 2965
15:54 Andrei666 tempire: thanks :)
15:55 Andrei666 tempire: shoulde I use redirect_to inside the action or in the main app pm?
15:55 tempire inside the action
15:55 tempire it only makes sense within an action
15:55 Andrei666 o
15:55 Andrei666 oh
15:55 Andrei666 ok :)
15:59 sri tempire: how do you feel about uploading the last release of the year?
15:59 tempire \o/
16:05 sri tempire: i think master is in a good state for release now
16:05 asim_ joined #mojo
16:08 tempire done
16:08 sri \o/
16:08 * tempire constructs a tweet
16:09 GitHub50 joined #mojo
16:09 GitHub50 [mojo] kraih pushed 1 new commit to master: http://git.io/q_9FXg
16:09 GitHub50 [mojo/master] fixed a few small typos in cookie test - Sebastian Riedel
16:09 GitHub50 left #mojo
16:11 tempire hmm
16:11 tempire indexing failed
16:11 sri oh
16:12 sri the new module
16:12 purl the new module is, like, see perlnewmod or h2xs -A -X -n Module::Name or ExtUtils::ModuleMaker
16:12 tempire yes
16:12 tempire good thing too, I forgot the disable env variable for the mac
16:13 sri ok, you all should have permission now
16:13 sri :O
16:13 sri that's very bad
16:14 MojoGuest509 joined #mojo
16:14 MojoGuest509 From: http://networkedblogs.com/33Knp (32 hits)
16:14 tempire what, the env variable?
16:14 sri ye
16:14 tempire I did it 'cause marcus said so, but I'm not sure what the problem is
16:15 sri you get garbage files included in the tarball otherwise
16:15 tempire more than .DS_store?
16:15 sri much more
16:16 sri that tarball needs to be thrown away asap
16:17 sri COPYFILE_DISABLE=1 is the env var btw
16:17 tempire yeah, done
16:18 sri you'll have to release 2.42 i think
16:20 sri doesn't look like there is garbage though
16:21 sri http://api.metacpan.org/sour​ce/TEMPIRE/Mojolicious-2.41/
16:21 sri maybe you're lucky
16:21 tempire wait…doesn't indexing failure mean I have to do it again?
16:21 tempire or will it retry
16:21 sri i don't know
16:22 tempire oh, I guess everything indexed except for pointer
16:23 sri tempire: you can force reindexing
16:23 sri pause has an option for that
16:24 sri look for garbage files though, they usually have the file name and some underscores
16:25 xaka joined #mojo
16:26 tempire "Released #mojolicious 2.41 with session improvements; stablized collections, app mounts, and etag/dnt headers https://github.com/kraih/mojo/blob/master/Changes #perl"
16:26 tempire er, guess I should link to the commit
16:26 sri not master changes, tag changes
16:26 sri ye
16:26 sri otherwise good
16:26 KayT3 joined #mojo
16:27 sri are you using vim exclusively?
16:27 tempire yes
16:27 inokenty Vim rocks!
16:27 sri ah, that might be what saved you from garbage files
16:27 sri textmate uses resource fork magic to store editing state
16:27 KayT3 joined #mojo
16:28 tempire I also have the lock files disabled
16:28 tempire vim wins again!
16:28 sri well, you just have to open the directory with an os x gui program and there would be garbage files ;p
16:29 sri so better use the env var in the future
16:29 tempire is there any disadvantage to exporting it permanently?
16:29 sri don't know
16:31 * tempire exports boldly in the face of uncertainty
16:31 tempire hooray for a real ssl cert
16:31 tempire now all pause needs is a COMPLETE rewrite and redesign from scratch.
16:32 sri :D
16:32 ispy_ joined #mojo
16:33 inokenty Does anyone uses emacs?
16:33 tempire there was on person at one point
16:38 tempire my klout score increased significantly after I made the first mojolicious release to cpan.
16:38 tempire I'm pretty sure that means I'm part of the internet now
16:44 hshong joined #mojo
16:46 * sri does not know what klout score is
16:46 noganex_ joined #mojo
16:55 tempire klout.com
16:56 plu bah, now you made me curious and sign up too :)
16:57 batman joined #mojo
16:57 sri my score is only 41 :(
16:58 plu that's more than twice of mine :)
16:58 plu I have 20, only
17:09 tarski joined #mojo
17:19 mercutioviz joined #mojo
17:21 kaare joined #mojo
17:21 GitHub154 joined #mojo
17:21 GitHub154 [mojo] kraih pushed 1 new commit to master: http://git.io/_2jyDQ
17:21 GitHub154 [mojo/master] removed obsolete deprecation warning - Sebastian Riedel
17:21 GitHub154 left #mojo
17:24 sri removing deprecated code always makes me happy :)
17:29 jegade joined #mojo
17:32 ispy_ joined #mojo
17:39 jegade hi, i like to split html into each child nodes and group this nodes by tag-name, how could i access the tag-name for an element with Mojo::DOM?
17:42 jegade Mojo::DOM->new('<p>First</p><video src="xx"></video>')->children->each( sub { my ( $e, $count) = @_; ACCESS THE TAG NAME OF $e HERE } );
17:53 bvierra joined #mojo
17:54 GitHub190 joined #mojo
17:54 GitHub190 [mojo] kraih pushed 1 new commit to master: http://git.io/lE-NGg
17:54 GitHub190 [mojo/master] better Mojo::DOM type example - Sebastian Riedel
17:54 GitHub190 left #mojo
17:56 sri that should answer the question
17:56 jegade ;-) ty
17:59 metaperl joined #mojo
17:59 GitHub178 joined #mojo
17:59 GitHub178 [mojo] kraih pushed 1 new commit to master: http://git.io/fSAOQQ
17:59 GitHub178 [mojo/master] slightly better type example - Sebastian Riedel
17:59 GitHub178 left #mojo
18:04 mib_0nlxtg joined #mojo
18:12 judofyr_ joined #mojo
18:27 mateu joined #mojo
18:28 mateu left #mojo
18:50 metaperl|2 joined #mojo
18:54 gshank_ joined #mojo
19:07 gshank joined #mojo
19:10 tempire sri: I presume you're familiar with how event machine works
19:11 tempire as far as I know, it's just a loop, similar to ioloop
19:12 gshank_ joined #mojo
19:15 sri tempire: mostly true
19:16 tempire apparently there's a non-blocking mysqlplus gem that sends a query to mysql, and then polls for the answer
19:17 sri how is that surprising?
19:18 tempire it's not, I'm just trying to understand exactly what was done, and why it hasn't been created for perl.
19:18 sri you have to recreate the binary wire protocol
19:18 sri in the end it's all just simple binary protocols over a socket
19:19 sri mysql doesn't do multiplexing though to my knowledge, so you need a pool of connections to do stuff parallel
19:19 tempire yeah, that's what the gem expects
19:20 sri ruby has more advantages though
19:20 sri green threads that release the GIL when waiting for io
19:21 tempire didn't know ruby had green threads
19:28 marcus Yo
19:28 sri yo yo yo
19:28 * purl walks the dog
19:30 sri marcus: was it you who wanted sessions that persist until the browser window is closed?
19:30 marcus No
19:31 marcus Btw, there's an async anyevent pg driver?
19:31 marcus For Perl.
19:31 tempire for realz?
19:31 sri DBD::Pg does async on its own
19:32 sri you just have to poll every now and then
19:32 tempire it's high time I got the postgres peep code episode
19:33 sri however… before starting with non-blocking db queries you should think about if it actually makes sense for you
19:33 tempire ah, a pure perl implementation of the postgres wire protocol
19:34 tempire it's the design/tech that I'm interested in, mostly
19:34 sri the facebook folks behind tornado prefer blocking db queries for automatic throttling
19:34 marcus http://bit.ly/vGke5V
19:34 marcus Sax json parsing
19:36 sri who streams json?
19:38 marcus Sri: any api that supports chunked?
19:39 sri no
19:39 sri twitter for example sends full json objects in a chunk
19:39 marcus If you're doing mobile apps streamparsing is a big deal.
19:40 sri i don't see how that relates
19:40 sri json is about the worst framing method i could think of
19:40 marcus Sri: why is it any worse than XML?
19:40 sri websockets and chunked encoding *are* framing methods
19:41 sri not the point
19:41 sri you already have framing built into the transport protocol
19:41 sri wrapping another json array around it only makes things complicated
19:42 sri for no gain
19:42 sri take twitter for example…. <header>{…}<header>{…}<header>{…}
19:43 sri they abuse chunked encoding very well
19:43 sri small json objects are parsed much faster with a blocking parser
19:43 sri and you reuse the framing of the transport protocol
19:44 sri and with websockets you don't have to abuse anything, framing is a protocol feature
19:45 marcus But depends on The api actually supporting it.
19:46 sri now you're changing the problem
19:47 sri if your api send huge ass json objects and you have memory constraints… of course a sax parser makes sense
19:48 sri but that's not usually the case for well designed apis
19:48 sri *+s
19:53 marcus I am not changing the problem, i was just pointing out it is useful for some cases.
19:58 wheelnotkiss joined #mojo
20:11 judofyr_ how do you get the user's IP?
20:11 judofyr_ and user agent?
20:11 purl user agent is sent as part of the HTTP request
20:11 judofyr_ I can't find anything in Mojo::Message::Request…
20:12 judofyr_ found it!
20:25 GitHub32 joined #mojo
20:25 GitHub32 [mojo] kraih pushed 1 new commit to master: http://git.io/RGl2MQ
20:25 GitHub32 [mojo/master] improved cookie header generation slightly - Sebastian Riedel
20:25 GitHub32 left #mojo
20:35 ccushing what is Mojolicious::Plugin::Mount?
20:35 ccushing I read the docs and I don't get what it does
20:36 ccushing why would I need it?
20:40 plu you can run two different (mojolicious) applications in the same process (e.g. hypnotoad) and serve them from a different path (e.g. /app1 and /app2)
20:40 plu and more than two of course :)
20:42 plu so let's say I wrote a guestbook application and you need a guestbook inside your app, you could just "mount" my guestbook application on /guestbook directly in your app
20:56 GitHub91 joined #mojo
20:56 GitHub91 [mojo] kraih pushed 1 new commit to master: http://git.io/KeQyWg
20:56 GitHub91 [mojo/master] improved RFC 6265 compliance of generated request headers - Sebastian Riedel
20:56 GitHub91 left #mojo
21:08 judofyr_ hm… how can I UserAgent to talk to a Lite-HTTPS-proxy
21:09 judofyr_ right now I get "Proxy connection failed."
21:09 GitHub43 joined #mojo
21:09 GitHub43 [mojo] kraih pushed 1 new commit to master: http://git.io/QGrABQ
21:09 GitHub43 [mojo/master] better cookie tests - Sebastian Riedel
21:09 GitHub43 left #mojo
21:09 judofyr_ even though I'm listening on https://127.0.0.1:13113:t/server.crt:t/server.key
21:09 judofyr_ and the client is connected with MOJO_KEY_FILE and MOJO_CERT_FILE
21:10 judofyr sri: any ideas?
21:13 judofyr there's probably not a "secure" way to intercept SSL connections anyway…
21:23 marcus Judofyr;
21:23 marcus No
21:25 judofyr is there a way to get Mojo::UserAgent to use a "less secure" HTTPS proxy? :P
21:25 judofyr i.e. a proxy which can modify the stuff…
21:26 judofyr Mojo::UserAgent, WHY U SO SECURE?
21:27 marcus You mean like, make it not check the certificate?
21:28 judofyr and not use HTTP CONNECT
21:28 marcus I am not sure I understand your use case.
21:39 GitHub183 joined #mojo
21:39 GitHub183 [mojo] kraih pushed 1 new commit to master: http://git.io/zin0Zg
21:39 GitHub183 [mojo/master] cookie micro optimizations - Sebastian Riedel
21:39 GitHub183 left #mojo
21:57 tempire marcus: you mentioned once that you could remove sub links from google results
21:57 tempire where would one go to do that
21:57 marcus Tempire: webmaster tools
23:01 alnewkirk joined #mojo
23:04 mattastrophe joined #mojo
23:11 GitHub127 joined #mojo
23:11 GitHub127 [mojo] kraih pushed 1 new commit to master: http://git.io/22f_vA
23:11 GitHub127 [mojo/master] cleaned up tests that change the environment - Sebastian Riedel
23:11 GitHub127 left #mojo
23:15 * sri yawns
23:36 wheelnotkiss sri: I think there's a bug in $app->sessions->default_expiration(0)
23:37 sri wheelnotkiss: i can confirm your observations though, there is an expires value
23:37 wheelnotkiss I don't understand why the tests work.
23:37 sri luck
23:37 sri it's the same second
23:37 wheelnotkiss ah
23:38 sri i would guess
23:38 wheelnotkiss but it's even luckier: in M::Sessions::load() it says return if $expires && $expires <= time; so even right now is expired.
23:38 sri bet there's a very very simple fix
23:39 wheelnotkiss sleep 1 :-)
23:39 sri just ran the tests a few fundred times, it sometimes fails
23:40 sri but ye, sleep 1 makes it fail reliably
23:43 sri haha
23:43 sri return if $expiration && !(my $expires = delete $session->{expires})
23:43 sri second part doesn't run
23:44 sri since $expiration is false
23:44 wheelnotkiss oops
23:46 wheelnotkiss that means the next line (that skips expired sessions) won't run either, since $expires will be undefined.
23:49 sri damn
23:49 sri i know what went wrong
23:50 sri think i have it fixed
23:50 wheelnotkiss nice!
23:50 sri the delete happens after the session is encoded *facepalm*
23:51 wheelnotkiss right, totally. In the patch, I only set $session->{expires} if $expiration || $default.
23:52 wheelnotkiss left #mojo
23:52 wheelnotkiss joined #mojo
23:56 wheelnotkiss the strange thing is it works sometimes in tls_lite_app.t. I'm still trying to figure out why it doesn't *always* fail.

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