Camelia, the Perl 6 bug

IRC log for #mojo, 2012-02-14

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

All times shown according to UTC.

Time Nick Message
00:30 marcus Sri: hypnotoad works afaik.
00:31 * marcus just set a highscore on theater of magic. 2,4 billion.
00:31 marcus 43. All time so far.
00:32 marcus Pinball arcade.
00:38 omf_ joined #mojo
00:45 jnap joined #mojo
01:02 d4rkie joined #mojo
01:06 scott joined #mojo
01:18 bird joined #mojo
01:40 tempire are you guys not paying attention?
01:40 tempire https://github.com/tempire/Mo​joExample/blob/heroku/Perloku
01:41 tempire it's running in foreground mode; apparently the forking into the background makes heroku think the process has crashed.
01:41 * tempire cracks the whip
01:41 tempire :o
02:21 omf_ So I wrote a test mojo app that does a simple ajax command. The funny thing is the ajax response works with morbo and hypnotoad but not with apache2 mod_perl. The app runs but the response is not sent. Is there anything special I need to be using.
02:27 jnap joined #mojo
03:00 mire joined #mojo
03:35 gshank_ joined #mojo
03:41 MacAtack joined #mojo
03:48 mercutioviz joined #mojo
04:12 MacAtack how does one send to the debug log from my own perl module, which I am attempting to use as a Model?
04:23 xaka joined #mojo
04:23 MacAtack heh - odd... now that I am actually doing tasks, print works.  must have not kept around a reference long enough  :)
05:02 vlixes joined #mojo
05:23 memowe joined #mojo
05:24 jr joined #mojo
05:25 al joined #mojo
05:25 augensalat joined #mojo
05:26 sven^ joined #mojo
05:27 rhaen joined #mojo
05:27 plu joined #mojo
05:27 Kallisti joined #mojo
05:28 stephan48 joined #mojo
05:29 sri joined #mojo
05:29 chansen joined #mojo
05:32 rafl joined #mojo
06:28 Vandal joined #mojo
06:45 mercutioviz joined #mojo
07:00 marcus purl: jerk it!
07:00 purl marcus: Must we lick the dick like so ?
07:05 Foxcool joined #mojo
07:20 ki0 joined #mojo
07:29 anparker hi. how can I check, is there helper with given name?
07:40 spleenjack joined #mojo
07:50 melmothX joined #mojo
07:54 tempire judofyr: heroku create is working consistently now
07:54 tempire 3 times in a row
07:55 tempire this is great.
07:55 tempire "my_app deploy heroku" just works
07:55 sri \o/
07:55 tempire INORITE
07:55 tempire no git knowledge required
07:58 * tempire waits for mongodb driver
07:59 tempire the ability to create an app, deploy it to heroku, deploy schema to free mongodb hosted service, all in a couple commands
08:02 sri mongodb has no schema
08:07 tempire account
08:07 tempire whatever
08:07 purl i heard whatever was going on there is going to be a bitch to figure out :)
08:09 tempire there's still a schema
08:09 tempire organization of collections
08:09 tempire I think of that as a schema
08:09 sri collections are created on the fly if they don't exist
08:10 tempire oh fine.
08:11 Foxcool joined #mojo
08:15 batman joined #mojo
08:16 mercutioviz joined #mojo
08:32 cosmincx joined #mojo
08:32 judofyr joined #mojo
08:32 memowe I'm happy now.
08:36 memowe http://www.pipingrad.com/sht/sht29.html
08:37 sri \o/
08:47 berov joined #mojo
08:48 sri tempire/crab: btw. https://github.com/kraih/mojo/issues/288
08:55 sri well, crab already said no on irc… so it's just tempire left
09:05 berov joined #mojo
09:18 black joined #mojo
09:19 ki0 joined #mojo
09:36 marcus sri: didn't see any tests indicating that layout should be applied to text=>, so I think it could be considered a bug. Should I open an issue for discussion?
09:36 marcus (looked through renderer.t)
09:38 sri marcus: two tests break
09:39 sri text with layout is specifically tested
09:43 sri in lite_app.t
09:44 sri line 811 and 1128
09:50 marcus oh.
09:51 marcus so it is. oh well.
09:53 vlixes left #mojo
09:59 noganex joined #mojo
10:20 osprey_ joined #mojo
10:24 osprey_ Hi
10:24 purl hey, osprey_.
10:24 osprey_ I still have problem with syswrite
10:24 osprey_ http://pastebin.com/90gwrEUk .
10:25 osprey_ Only if one param have utf8 chars and {file=>..}
10:28 osprey_ oh, wait, it was old mojo (1.99)
10:29 marcus ancient :)
10:29 osprey_ yeah
10:29 osprey_ my local sitelib
10:32 daxim joined #mojo
10:33 osprey_ >Can't locate object method "is_done"
10:33 osprey_ that is why im lazy to uptodate :)
10:53 sri tempire will decide the vote on #288
11:00 marcus osprey_: is_finished
11:05 sri if something has been deprecated and removed just search Changes, new alternatives are always mentioned
11:06 marcus sri: so if tempire is yay, do we have to have a tag team match to decide the outcome?
11:07 sri nope, i'll change my vote to neutral then
11:08 osprey_ marcus: thank
11:08 sri although a tag team match sounds tempting
11:08 osprey_ Well, still have problem with syswrite
11:09 osprey_ but now I have wipe of error Write failed: Wide character in syswrite at .. /IO/Handle.pm line 476. until Inactivity timeout
11:10 sri perhaps you shouldn't write wide characters
11:10 osprey_ hm
11:10 sri purl: perl unicode?
11:10 purl perl unicode is http://www.slideshare.net/Penfold/perl-and-unicode and http://stackoverflow.com/questions​/6162484/why-does-modern-perl-avoi​d-utf-8-by-default/6163129#6163129
11:11 sri you should read up on that
11:11 osprey_ but its ok when wide characters sends w/o file =>
11:11 sri it's characters vs bytes
11:12 osprey_ yes
11:15 cosmincx joined #mojo
11:17 osprey_ ok, done with 'UTF-8': my $tx = $ua->post_form($murl->to_string(), 'UTF-8' => \%params, {Referer => $url});
11:25 D4RK-PH0ENiX joined #mojo
11:28 briang joined #mojo
11:54 cosmincx joined #mojo
12:04 graf joined #mojo
12:04 sven joined #mojo
12:06 rhaen joined #mojo
12:07 sri joined #mojo
12:07 rafl joined #mojo
12:07 plu joined #mojo
12:11 Kallisti joined #mojo
12:12 chansen joined #mojo
12:13 stephan48 joined #mojo
12:15 al joined #mojo
12:20 alnewkirk joined #mojo
12:42 black joined #mojo
12:47 d4rkie joined #mojo
12:47 gabriel__ joined #mojo
13:34 jnap joined #mojo
13:41 gshank joined #mojo
13:50 vel joined #mojo
14:21 tabbi joined #mojo
14:21 tabbi left #mojo
14:28 sven^ joined #mojo
14:28 Kallisti joined #mojo
14:29 augensalat joined #mojo
14:30 plu joined #mojo
14:30 sri joined #mojo
14:30 al joined #mojo
14:32 rhaen joined #mojo
14:35 stephan48 joined #mojo
15:04 ver joined #mojo
15:09 geira joined #mojo
15:09 MacAtack joined #mojo
15:20 MacAtack joined #mojo
15:29 amoore joined #mojo
15:31 ver joined #mojo
15:41 inokenty joined #mojo
15:53 osprey_ left #mojo
15:57 d4rkie joined #mojo
16:04 Foxcool joined #mojo
16:19 chansen joined #mojo
16:34 MacAtack joined #mojo
16:44 Vandal joined #mojo
16:49 MacAtack i cannot seem to get any of the websocket examples to operate as I'd expect.  perhaps has something changed within mojo?
16:51 MacAtack So far all i've tried is safari - grabing chrome now
16:51 sri safari doesn't support rfc 6455
16:53 MacAtack safari doesn't attempt to use the flash policy server as a fallback, so either the examples are bad or there is some kind of support there... no ?
16:55 sri it uses an old insecure legacy version of the protocol
16:55 MacAtack ugh :(
16:55 sri apple is not very good at html5
16:56 sri worst part is, webkit apparently already includes proper rfc 6455 support, apple just doesn't enable the compile flag for safari
17:00 MacAtack guess i'll stick to ajax/json with mojo for a new project ... i have previously written several catalyst apps for customers
17:01 sri chrome and firefox have excellent rfc 6455 support though
17:01 sri ie and safari have it, but aren't shipping it yet
17:02 MacAtack what's with the flash garbage though on firefox ?
17:02 MacAtack on the exampels
17:03 sri what examples?
17:03 purl examples is http://shorl.com/bynygruhofrebru
17:04 MacAtack this one seems old:  http://d.hatena.ne.jp/yuki​kimoto/20100418/1271601822
17:04 sri lol 2010
17:04 mercutioviz joined #mojo
17:04 sri the spec was finished a few weeks ago
17:05 sri and the browser api spec is still getting polished
17:05 sri go figure
17:05 MacAtack heh - nice ... i've always been an early adopter
17:07 KayT3 joined #mojo
17:08 mire joined #mojo
17:09 sri i have a feeling old websocket examples will be coming back to haunt us once all browsers support them
17:09 abra joined #mojo
17:10 MacAtack yeah, bad example code is not fun
17:10 sri i guess we could say "ignore everything before 2012"
17:11 * sri wonders if firefox has removed the -moz- prefix yet from the api
17:12 berov joined #mojo
17:14 xaka joined #mojo
17:16 tempire I think 288 is a good idea…but only if it can be tested.
17:17 tempire tests tests tests
17:17 tempire the dude seems to be fine with the wrapper script
17:18 sri how could it be tested?
17:21 kaare joined #mojo
17:23 tempire I can't think of a way.
17:23 tempire wonder why brian d foy deleted his article about chomp
17:29 abra_ joined #mojo
17:30 sri tempire: soooo… your vote is a nay too?
17:31 tempire I suppose so.  Unless marcus can come up with a way to test it.
17:34 tempire I don't think gabor really understood the complaint.
17:35 tempire but at least things have remained friendly.
18:12 gabriel__ joined #mojo
18:19 mire joined #mojo
18:27 noganex_ joined #mojo
18:35 sri tempire: btw. why do you think 288 is a good idea?
18:36 tempire It's something I've run into a couple of times.  would have been convenient.
18:36 sri i don't get what common problem is solves
18:36 sri when?
18:38 tempire testing things with multiple mojolicious versions
18:38 sri wouldn't you use the hypnotoad/morbo command from those respective versions anyway?
18:38 sri after all… they do contain code too
18:39 tempire yes.  you're right.  as soon as you asked the question, I realized I hadn't really thought that through.
18:39 * tempire is caught
18:39 sri i'm actually pretty sure those scripts have had incompatible changes a few times
18:42 * sri wonders if marcus overlooked that too
19:09 ki0 joined #mojo
19:12 mercutioviz how hard have you guys pushed hypnotoad? just curious if you know where the breaking point is
19:12 * mercutioviz likes to break things
19:12 sri what breaking point?
19:13 mercutioviz i mean simultaneous requests
19:13 mercutioviz load testing, that sort of thing
19:14 sri at some point the backlog will be full and you get rejected connections or your server runs out of memory if you allow a gazillion descriptors
19:15 jwang joined #mojo
19:16 judofyr joined #mojo
19:21 sri i've actually not profiled memory requirements per connection in quite some time
19:26 marcus sri: no, most of the time I've needed to include libs that aren't installed in the perl INC, not mojolicious itself.
19:26 marcus sri: I think it's fair enough about the testing tho.
19:26 sri marcus: but that sound like you could have used "use lib" from the app or config too, no?
19:27 marcus sri: developers have different setups, so the paths vary.
19:27 MacAtack joined #mojo
19:27 sri can you give an example?
19:30 sri like hypnotoad -I../../repo/Mojolicious-Plugin-Nordaaker/lib -Ilib ./script/my_app?
19:30 marcus like tvguide project is split up into meetv-common, meetv-backend, meetv-ws meetv-web5 I actually have a PERL5LIB thingie in my .bashprofile to include $dir/lib in each of the ~/Source/TVGuide directories tho.
19:30 sri hmm
19:33 marcus not sure it's a good example. We also do use lib in another project, where we had a bunch of subprojects in vendor/
19:34 marcus like the parallol and solarbeam plugin judofyr wrote, and a dbic schema.
19:35 judofyr woot, I'm on IRC \o/
19:35 marcus judofyr: famous.
19:38 marcus judofyr: seems your sync thingie doesn't generate iso8601 timestamps btw?
19:38 marcus or is that just the default stringification working?
19:38 judofyr default stringification
19:38 judofyr I thought it was iso8601 by default…
19:40 marcus seems you're right.
19:41 judofyr \o/
19:42 sri marcus: how do you test such apps?
19:43 sri don't you need the PERL5LIB hacks anyway?
19:43 marcus sri: prove supports -I
20:03 Netfeed \o/ my site works in newest version
20:08 sri hmmm
20:08 sri i might have a way to actually test it
20:08 sri marcus/tempire: does this change the situation?
20:09 marcus sri: if we're adding this, I suggest using '-I' like perl/prove rather than --lib
20:10 sri -l is not an option
20:10 sri already in use for --listen
20:10 marcus yes, I know
20:10 marcus I didn't suggest using -l ? :)
20:10 sri oh crap… bad irc font
20:11 sri -I and --include?
20:11 * marcus ircs with Menlo. accept no substitute
20:11 marcus sri: yupp
20:11 marcus sri: not sure we need the long form, but I guess it can't hurt.
20:12 sri for some reason my textual uses lucida grande
20:12 marcus I think that's the default for some stupid reason.
20:12 marcus it looks pretty bad.
20:12 marcus hmm, kinda seems like textual development has stalled :-/
20:13 sri damn… my test works… so -I is back on the table
20:15 marcus sri: another bright point of this is it makes it easier to write init / cron scripts that use local libs
20:15 sri how so?
20:15 marcus cron doesn't run with your normal ENV
20:15 sri what would you cron with morbo/hypnotoad?
20:16 marcus so you can have a crontab like '@reboot ~/perl5/bin/hypnotoad -I ~/perl5/lib ~/Source/myapp.pl'
20:16 sri oh, right
20:17 sri wait a minute
20:18 sri @INC manipulation doesn't survive the exec
20:18 marcus use lib works tho?
20:18 sri it's on the command line, first restart would have lost that data
20:18 sri because use lib is actual code
20:18 sri it runs again
20:19 sri but command line options are lost
20:19 marcus in the case above, a restart would have the correct env
20:19 marcus since you do the restart from a logged in shell
20:19 sri nope, it wouldn't
20:19 marcus oh,you mean the restart doesn't get the env from the restart either?
20:20 sri nono
20:20 sri we manipulate @INC
20:20 marcus from the shell running the restart I mean
20:20 sri not %ENV
20:20 sri after restart we get a fresh @INC without the manipulations, since the command line arguments don't exist anymore
20:21 marcus but don't we get the @INC from the command line running the restart?
20:21 sri PERL5LIB on the other hand works
20:21 sri huh?
20:21 marcus you start a new perl, right?
20:21 sri entirely new, only %ENV and a few descriptors survive
20:22 marcus doesn't it get the env from the shell that issues the restart?
20:22 marcus oh, I guess you can restart with a signal as well...
20:22 sri how is @INC related to the shell?
20:22 sri or tied to the shell
20:22 marcus sri: if your shell has PERL5LIB set
20:22 marcus that changes @INC
20:22 sri but -I foo/bar doesn't change PERL5LIB
20:23 marcus no, it doesn't.
20:23 sri that's the problem
20:24 sri first hypnotoad start would have the right @INC, but after restart all -I directories are lost
20:24 marcus sri: what about --config, before we deprecated that?
20:24 marcus was that lost on a restart as well?
20:24 sri all those settings are preserved in ENV vars
20:25 sri we could theoretically presrve -I directories in an env var too… but that's not testable again
20:25 marcus guess we would have to set an ENV var to support it then. seems messy.
20:26 sri and very messy
20:26 marcus yeah
20:26 sri :)
20:26 marcus not worth it, probably.
20:26 marcus I'll write a comment on the ticket and close it
20:26 sri marcus++
20:28 abra__ joined #mojo
20:32 sri that's also the main reason i exec on first start too
20:32 sri to ensure the env has been initialized right
20:32 sri no bad surprises on first restart ;)
20:39 sri i think the -l problem is what rvm solves with their evil cd hack
20:40 sri right, project specific .rvmrc files
20:41 judofyr sri: hm? you don't need the cd-hack for RVM to work
20:42 judofyr it's only for getting "cd into project, automatically choose Ruby-version" to work
20:42 sri you can't use it for setting env vars?
20:42 judofyr yes, you can
20:43 sri that would solve the problem marcus described
20:44 sri arguable though what's more sane, bash hack or rvm hack
20:44 judofyr see also: https://github.com/zimbatm/direnv
20:44 sri oh, that looks pretty neat
20:49 MacAtack joined #mojo
21:23 lukep joined #mojo
21:26 vervain joined #mojo
21:29 MacAtack_ joined #mojo
21:31 mire joined #mojo
21:31 baton8 joined #mojo
21:41 MacAtack joined #mojo
21:54 abra_ joined #mojo
22:06 batman joined #mojo
22:45 marcus Hmm, who is http://iam.richardbucker.com/, and when did we scold him?
22:47 marcus The comments on that Why I Left Perl piece are turning pretty vicious..
22:49 abra__ joined #mojo
23:07 sri marcus: that's the guy that called tempire rude
23:07 marcus Ah. I guess we'll let that one go.
23:08 tempire it turns out he was furious that I said no dependencies when the cpan list clearly showed deps
23:08 sri oh, right
23:08 marcus He has some stupid post about accessors being a dumb idea whem you have hundreds of them in your class on his blog....
23:08 sri that was so weird
23:08 marcus When.

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