Camelia, the Perl 6 bug

IRC log for #mojo, 2011-08-29

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

All times shown according to UTC.

Time Nick Message
00:02 chorny joined #mojo
00:03 alnewkirk joined #mojo
00:40 chansen joined #mojo
01:10 d4rkie joined #mojo
01:32 rwstauner joined #mojo
01:33 gwui joined #mojo
01:43 d4rkie joined #mojo
02:02 jkitazawa joined #mojo
04:12 tempire TO THE MOJOLICIOUS!!!!!1!!!
04:18 kaare joined #mojo
04:56 vel joined #mojo
04:57 arpadszasz_ joined #mojo
05:03 perlrocks joined #mojo
05:04 irs marcus: yes, perlrocks crashed
05:13 Eugene joined #mojo
05:48 alpha6 joined #mojo
05:58 irs marcus: those are the people that ruin release early release often for everyone else...
06:05 eugen joined #mojo
06:08 * irs goes back to mongodb hacking
06:14 koban` joined #mojo
06:14 marcus irs: yupp
06:15 marcus irs++ # mongodb hacking
06:15 marcus irs: have you considered that you are in fact giving the IRS a good name?
06:37 irs hmmm
06:38 marcus :D
06:50 sri so many ideas for hack of the day articles :)
06:50 sri think next i'll show the css4 selector
06:57 tempire sri: I've been reading hacker news a bit
06:57 tempire and it's clear that for real up votes, you need an article
06:58 tempire a lot of folks were probably seeing the screencast and then moving on
07:00 tempire so I'm working on a galvanizing article that will tie into mojocasts and what I'm going for the next episode
07:00 tempire *doing
07:01 sri smart
07:02 tempire assuming it can get first place using the same technique as last time, it's likely it will hook
07:15 eugen joined #mojo
07:42 gizzlon tempire: you should post the hn link here later :)
07:42 tempire I will.  But it's not going to be until next week :)
07:43 gizzlon k
07:44 marcus o/' be mine , be mine , be my little rock rock and roll queen o/'
07:48 cosmincx joined #mojo
07:54 alpha6 joined #mojo
08:15 sugar joined #mojo
08:19 spleenjack joined #mojo
08:26 arthas joined #mojo
08:30 sugar joined #mojo
08:33 sri https://github.com/joyent/node/pull/1580 # can we please have a pull request with this much drama too?
08:33 crab hi
08:33 purl what's up, crab.
08:33 * sri waves
08:34 sri in fact, our pull requests are extremely low on memes :(
08:43 black joined #mojo
08:54 d4rkie joined #mojo
08:56 GitHub67 joined #mojo
08:56 GitHub67 [mojo] kraih pushed 1 new commit to master: http://git.io/t1A4Ag
08:56 GitHub67 [mojo/master] removed obsolete FAQ answers - Sebastian Riedel
08:56 GitHub67 left #mojo
09:00 cosmincx joined #mojo
09:06 marcus sri: found my actual worst thing about vico. No replace mode.
09:07 marcus (R)
09:07 * sri is back on textmate
09:09 sri i would love a textmate with vi mode
09:09 sri that's really all i want :/
09:10 marcus sri: vico seems close tho
09:11 marcus sri: but I agree, won't buy it until it's actually good enough
09:11 sri macvim with smooth scrolling and a buffer browser sidebar would be ok too
09:11 marcus yeah
09:17 ghandi joined #mojo
09:23 sri marcus: did chocolat get any better recently?
09:24 marcus sri: a little
09:24 marcus still not quite there tho
09:34 judofyr joined #mojo
09:48 sromanov joined #mojo
09:50 Firkin joined #mojo
09:54 Eugene joined #mojo
09:55 noganex mh.. why would you need a gui for editing perl code? o_O
09:55 noganex i'm still using vim on mac os x.
09:55 sugar joined #mojo
09:57 noganex (and for anything else there is eclipse)
09:58 judofyr when did Java become "everything else"?
09:58 plu joined #mojo
09:58 noganex judofyr: eclipse isn't just for java ;)
10:05 plu joined #mojo
10:11 d4rkie joined #mojo
10:13 jwang joined #mojo
10:16 sri noganex: vim is great for editing single files, but sucks for projects
10:17 sugar nerd tree
10:17 sri sucks
10:17 jwang joined #mojo
10:18 sri not to mention buffer management
10:18 sri just considering every file to be open doesn't work for me
10:20 sri my most common workflow is… search for something in a whole project, open 20 files, make random changes in all files, test, commit
10:20 sugar i used vim in paste, but now i am at Padre? it's good for me
10:20 sri navigating between those 20 files is absolutely miserable in vim
10:20 sugar yes
10:21 sugar but it has shorcut for every buffeer
10:21 sri and after the first 20 files?
10:21 sri if you open new files it gets a higher buffer number
10:21 sugar shortcut+number
10:21 sugar or aother sorcut + shift
10:22 sri :b1000000045
10:22 sri fun!
10:22 sugar *another shortcut
10:22 sri that sucks
10:22 purl The rock is now off.
10:22 sri seriously, why would i have to make my own shortcuts just to edit a few files?
10:23 noganex because timtowtdi ;)
10:24 sri i'd rather use a better editor when i have more than a shell
10:24 sri i still use vim all the time for single file edits, but for project editing it's unusable for me
10:27 marcus I'm with sri
10:28 marcus the project bit works well in vico tho
10:28 sri ye, the general workflow is great in vico
10:28 sri it just lacks editing features
10:29 marcus yupp
10:34 judofyr sri: I use Command-T for opening buffers/files
10:35 sri heh, i've actually just thought about how command-t would work for buffers
10:35 judofyr <Leader>b
10:35 judofyr it works just like <Leader>t just for buffers
10:35 judofyr other than that, I just use split screens everywhere
10:36 judofyr and Ctrl-6 to jump between current buffer and the previous open buffer
10:39 koban joined #mojo
10:52 perlrocks Twitter: "www/p5-Mojolicious - 1.92: - Update to 1.92  Changes:        http://t.co/ToyQrlu PR:... http://t.co/8dV8NRd" --freshports_org http://twitter.com/freshports_​org/status/108129777768079361
10:58 sri command-t for buffer management is actually not too bad
11:02 d4rkie joined #mojo
11:09 sri eeeep
11:09 sri but macvim looks horrible in lion
11:10 sri visible scrollbar, broken fullscreen, no smooth scrolling...
11:10 * sri pokes his eyes out
11:30 marcus sri: agree
12:07 tabbi joined #mojo
12:07 tabbi left #mojo
12:34 plu hm. That vico editor looks really promising :)
12:36 noganex mh... my dog chased a squirrel - and caught it.
12:36 noganex mad beast...
12:36 plu so you got new options for dinner :)
12:38 noganex what? squirrel au vin?
12:47 marty Good morning ya'll
12:50 crab what's the easiest/shortest way to ask "did the request have a cookie named 'x'?"
12:54 sri cookie or cookie value?
12:55 sri i think defined($c->cookie('foo')) should work, but am not 100% sure
12:56 * sri waves
12:57 crab seems good.
13:01 marcus I'm still using vico
13:29 crab my $HOST_RE = qr/^([^\:]*)\:?(.*)$/;
13:29 crab what is that supposed to match? (M::Message::Request.pm)
13:30 crab oh i see. anything followed by optional port specification.
13:31 crab sri: M::M::R::parse should handle multiple-valued X-Forwarded-Host, shouldn't it?
13:35 rwstauner joined #mojo
13:38 crab because the common proxy modules seem to append, though it's stupid
13:41 tholen joined #mojo
13:41 sri crab: there is no standard
13:41 crab i know
13:42 sri if it doesn't work properly i actually wouldn't mind just removing it again
13:42 crab if there were one, i'd say just parse that. it's because there's no standard that i'm thinking about making it more cooperative...
13:42 sri considering that good proxies like nginx just send the right host header
13:42 crab but that said, between when i asked and now, i have read some nginx source and i think i can make it work ok without any risk from clients who try to fake those headers.
13:43 crab so i've suddenly stopped caring. ;-)
13:43 GabrielVieira joined #mojo
13:44 sri huh?
13:44 sri fake a host header? that makes no sense
13:46 crab sri: fake X-Forwarded-HTTPS
13:46 crab or X-Forwarded-For
13:46 purl X-Forwarded-For is the client IP, I think
13:47 sri yes, you need to be aware of those, but we were talking X-Forwarded-Host
13:52 GitHub13 joined #mojo
13:52 GitHub13 [mojo] kraih pushed 1 new commit to master: http://git.io/EVNbzQ
13:52 GitHub13 [mojo/master] improved nginx recipe - Sebastian Riedel
13:52 GitHub13 left #mojo
13:52 sri guess mentioning it doesn't hurt
14:03 crab sri: i'd suggest changing $proxy_add_x_forwarded_for to $remote_addr on the line above also
14:03 d4rkie joined #mojo
14:03 crab sri: and also $http_host above to $host (that will send the right value even if the client doesn't send a Host; header in the request)
14:04 crab (the $remote_addr change ensures that X-Forwarded-For is a single ip, not a list)
14:04 sri what's the point?
14:05 crab of which?
14:05 sri why would we suggest to handle x-forwarded-for incorrectly?
14:05 crab incorrectly? i thought you said there was no standard.
14:06 sri we were talking about X-Forwarded-Host!
14:07 crab <sri> why would we suggest to handle x-forwarded-for incorrectly?
14:07 sri "incorrectly? i thought you said there was no standard."
14:07 sri i never said anything like that about x-forwarded-for
14:09 crab the nginx recipe says "proxy_set_header X-Forwarded-For $proxy_add_x_forwarded_for;"
14:10 crab anyway, obviously i misunderstood something you said, or you misunderstood something i said, or both. but it's not important enough to dissect. my problem is solved, and the recipe is ok. you need to understand nginx's proxy module to use it anyway, so anyone can make their own decisions.
14:11 sri you made suggestions i don't understand
14:11 sri especially the x-forwarded-for one seems just plain wrong
14:11 stephen joined #mojo
14:12 GitHub22 joined #mojo
14:12 GitHub22 [mojo] kraih pushed 1 new commit to master: http://git.io/wianUg
14:12 GitHub22 [mojo/master] improved nginx recipe some more - Sebastian Riedel
14:12 GitHub22 left #mojo
14:13 sri $host seems reasonable however
14:15 spleenjack1 joined #mojo
14:15 crab i am using "proxy_set_header X-Forwarded-For $remote_addr" because i (a) know that my nginx talks to the clients directly, and (b) i want to discard any X-Forwarded-For values that the client provides in its request.
14:16 sri but why would we suggest that to our users?
14:16 stephen1 joined #mojo
14:16 crab it just seems a more common situation than someone who legitimately wants a chain of X-Forwarded-For ips.
14:17 sri i think the common situation is that the user just doesn't care
14:18 sri but i don't see why we shouldn't suggest handling it according to standard
14:18 sri for the user it's the same, they just get a little more information a potential plugin could use for example
14:19 crab now i can see where i misunderstood you.
14:19 crab when you said "we were talking about x-f-host", i did not read that to mean "but x-f-for DOES have a standard (of sorts)"
14:20 sri http://en.wikipedia.org/wiki/X-Forwarded-For # it has a long history
14:21 crab so i see. thanks.
14:22 sri it's the only X-Forwarded header that's not a total PITA :)
14:23 sri especially the protocol forwarding is a total mess
14:26 vel joined #mojo
14:28 tholen joined #mojo
14:34 ispy_ joined #mojo
14:42 koban left #mojo
14:42 perlrocks Twitter: "Mojolicious Web Framework - Join the Perl revolution! http://t.co/NuZkKMf" --rbucker http://twitter.com/rbucker​/status/108187887211786241
14:51 abra joined #mojo
14:51 sugar joined #mojo
15:00 judofyr joined #mojo
15:09 sri tempire: http://iam.richardbucker.com/​2011/08/29/get-your-mojo-on/ # :)
15:11 hernan win 18
15:11 sri fail
15:12 vel joined #mojo
15:13 arpadszasz joined #mojo
15:14 arpadszasz hello
15:15 * sri waves
15:16 arpadszasz i'm playing with eventsource in a mojo app, using a modified version of this paste i found: http://pastebin.com/YQ4L47bk
15:17 arpadszasz setting the Mojo::IOLoop time parameter to grater than 20 seconds still runs the loop every 20 seconds
15:18 arpadszasz for ex. Mojo::IOLoop->recurring(30 => sub {})
15:19 arpadszasz is this known behavior or i should test deeper?
15:21 sri arpadszasz: perl -MMojo::IOLoop -E'say time; Mojo::IOLoop->recurring(30 => sub { say time }); Mojo::IOLoop->start;'
15:22 sri (works flawless here)
15:22 arpadszasz sri: that runs ok 1314631297 ... 1314631327... 1314631357
15:23 arpadszasz sri: thanks!
15:23 sri wait
15:23 sri does your browser maybe reconnect?
15:24 arpadszasz might be, Firefox 6
15:24 sri default inactivity timeout is 15 seconds
15:24 arpadszasz just a second
15:24 sri the browser prolly waits 5 seconds for reconnect
15:24 sri :D
15:24 sri your options are to send a ping every < 15 seconds, or to increase the timeout
15:25 sri Mojo::IOLoop->singleton->connecti​on_timeout($self->tx->connection => 300);
15:25 sri but ping is not such a bad idea
15:26 sri the 15 seconds in the example are not chosen arbitrarily
15:26 arpadszasz sri: yes, morbo logs show that the browser reconnects
15:26 arpadszasz sorry for missing that
15:26 arpadszasz :)
15:27 sri i would expect that to become an FAQ if eventsource goes mainstream ;)
15:27 eugen joined #mojo
15:28 sri for websockets we default to 300 seconds, so people usually don't stumble over it
15:31 sri really wonder if eventsource has a chance to go mainstream before everybody uses websockets
15:33 arpadszasz i guess the majority of people only need server-push functionality and are running on Apache (no default websoscket)
15:34 sri well, you can't really have server push with apache either
15:36 arpadszasz sri: dammit you're right :(
15:36 sri theoretically you can stream data through it, but there are many obstacles
15:37 sri the cgi/fastcgi layer might cache, chunked transfer encoding might interfere with the event stream somehow, one process per client is very expensive
15:38 sri i've not actually tried server sent events on apache yet though
15:41 arpadszasz sri: thanks for the tips, i'll test it and report back
15:41 sri just be aware that timers don't work with the cgi/fastcgi backends (since there's no ioloop running)
15:43 sri basic tools like the drain callback should work however
15:52 spleenjack joined #mojo
16:06 Akron joined #mojo
17:07 sugar joined #mojo
17:13 arpadszasz sri: no go with eventsource and Apache CGI/FCGI
17:19 tempire bam!
17:20 tempire why bother with buffers in macvim
17:20 tempire just use tabs and be done with it
17:28 mattastrophe joined #mojo
17:29 tempire also, long live VIM!
17:29 tempire death to the mouse
17:29 arpadszasz tempire++
17:29 arpadszasz :)
17:30 arpadszasz sri: just wondering is there a way to share eventsource data between hypnotoad workers ?
17:31 tempire the problem with 'vim' modes of different editors is that they just have the hjkl navigation.  I'd like to a see the full vim engine with my entire .vimrc supported.
17:32 tempire arpadszasz: the hypnotoad workers are completely different processes.  you would have to communicate out-of-band
17:32 arpadszasz tempire: that's what i've feared
17:44 j-v-e joined #mojo
17:47 arpadszasz joined #mojo
17:52 judofyr joined #mojo
17:56 alpha6 joined #mojo
18:04 sri marcus: getting reports that mojolicious doesn't install because cpanmin.us is down :(
18:23 Firkin left #mojo
18:55 black joined #mojo
18:56 ispy_ joined #mojo
18:59 trone_ joined #mojo
19:07 noganex_ joined #mojo
19:13 tabbi joined #mojo
19:13 tabbi left #mojo
19:49 judofyr joined #mojo
19:50 perlrocks Twitter: "Mojolicious hack of the day: HTML5 EventSource http://t.co/s5LRArn #perl" --perl_m8y http://twitter.com/perl_m8y​/status/108265201698213888
19:51 noganex joined #mojo
20:16 Eugene joined #mojo
20:17 Eugene joined #mojo
21:11 perlrocks Twitter: "Internationalization (i18n) with Mojolicious and Template Toolkit http://t.co/JmQtWgY" --cstrep http://twitter.com/cstrep/​status/108285784251039745
21:20 rwstauner joined #mojo
21:29 trone joined #mojo
21:51 perlrocks Twitter: "Cosimo Streppone (cosimo):      Internationalization (i18n) with Mojolicious and Template Toolkit http://t.co/uL6elYC" --perlironman http://twitter.com/perlironm​an/status/108295667880960000
21:58 Shaeto joined #mojo
22:00 Shaeto is it possible to change controllers name space for subroutes ? tried to change namespace in 1.94 and it still trying to search using root router namespace
22:02 perlrocks Twitter: "This is a pretty interesting post about Python apps, Python versions, and Perl based frameworks http://t.co/XHSiC6d" --DrKnucklehead http://twitter.com/DrKnuckleh​ead/status/108298487145308160
22:47 gshank joined #mojo
23:14 black joined #mojo
23:25 noganex joined #mojo
23:45 perlrocks Twitter: "Comparing #perldancer and #mojolicious. Article should be ready in a few days." --rbucker http://twitter.com/rbucker​/status/108324429259472896
23:55 rwstauner joined #mojo

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