Camelia, the Perl 6 bug

IRC log for #mojo, 2012-09-12

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

All times shown according to UTC.

Time Nick Message
00:01 * dross ponders a project name
00:25 crab "MyApp"
00:28 dross pfft
00:28 dross then something non-related to the program or IT
00:29 dross kangaroos not into feeding elephants (knife)
00:30 dross kangaroo native [to] infrastructure frameworks [and] enterprise   (knife)
00:33 dross kangaroo native [to] infrastructure frameworks [and the] enterprise   (knife)
00:45 laouji joined #mojo
00:50 xaka joined #mojo
01:09 mandreacchio joined #mojo
01:49 mandreacchio joined #mojo
01:52 jnap joined #mojo
01:58 mandreacchio joined #mojo
01:58 mandreacchio joined #mojo
02:00 d4rkie joined #mojo
02:10 d4rkie joined #mojo
02:12 marty_ joined #mojo
02:33 marty joined #mojo
02:34 marty joined #mojo
02:51 noganex_ joined #mojo
02:55 mandreacchio joined #mojo
03:33 BlueT_ joined #mojo
03:44 vel joined #mojo
03:49 pingup3rl joined #mojo
03:56 mandreacchio joined #mojo
03:57 pingup3rl joined #mojo
04:00 mandreacchio joined #mojo
04:58 good_news_everyone joined #mojo
04:58 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/dPP9YA
04:58 good_news_everyone [mojo/master] fixed small CGI bug - Sebastian Riedel
04:58 good_news_everyone left #mojo
05:04 mandreacchio joined #mojo
05:16 xaka1 joined #mojo
05:31 mandreacchio joined #mojo
05:46 Averna joined #mojo
05:49 fhelmber_ joined #mojo
05:49 marty joined #mojo
05:53 pingup3rl joined #mojo
06:03 mandreacchio_ joined #mojo
06:05 spleenjack joined #mojo
06:08 Vandal joined #mojo
06:24 mandreacchio joined #mojo
06:49 abra joined #mojo
06:50 mandreacchio joined #mojo
07:08 mandreacchio joined #mojo
07:18 mandreacchio joined #mojo
07:19 mandreacchio joined #mojo
07:23 batman joined #mojo
07:24 jamesw joined #mojo
07:43 cosmincx joined #mojo
07:52 rem_lex| joined #mojo
07:52 mandreacchio joined #mojo
07:53 pingup3rl joined #mojo
07:54 sri o/
08:05 AmeliePoulain joined #mojo
08:08 pau4o joined #mojo
08:18 Kwa joined #mojo
08:51 geira joined #mojo
09:08 geira joined #mojo
09:14 geira joined #mojo
09:19 andrefs joined #mojo
09:19 tempire \o
09:20 * tempire dances to kpop
09:35 sri \o\
09:35 sri /o/
09:37 andrefs joined #mojo
09:42 pingup3rl joined #mojo
09:48 asarch joined #mojo
09:53 Lucas1 joined #mojo
09:56 basic6 joined #mojo
10:33 nic tried to install Hopkins
10:34 nic after 24 mins and 116 distributions... it failed (cos 2 distributions won't build)
10:34 nic the world of numerous dependencies sucks
10:38 Kwa :/
10:43 Kwa Anyone written an application which renders pages depending on whether the useragent is a web browser or a mobile?
10:44 Kwa Or know which doc I should be reading?
10:49 memowe Kwa: http://mojolicio.us/perldoc/Moj​olicious/Plugin/HeaderCondition with 'agent' might help. :)
10:51 sri god i hate sites that have a different site for mobile
10:51 sri s/site/version/
10:53 sri a clever design that scales down with the browser window is awesome... but special dumbed down mobile sites annoy the hell out of me :)
10:53 Kwa Ah, I was just about to ask why you didn't like them. :)
10:54 xxtjaxx joined #mojo
10:55 Kwa sri: Well the 'bossman' wants an app, but I didn't fancy the maintenance of that on top of the current website, so I compromised by making it look like an app at least.
10:55 Kwa sri: It's for a panel to view messages, and update contact details. It's not the whole site.
10:55 sri then at least allow people to opt out!
10:56 Kwa Haha. Yeah good point. I don't like sites that force you to mobile versions of sites which are so badly broken you cannot use them.
10:57 Kwa I'll probably have it as opt-in, thinking about it.
10:58 sri good choice
11:07 nic Anyone able to load github pages on an android browser?
11:07 * memowe .
11:07 nic which browser?
11:09 memowe Um - the one on my Wildfire S :P
11:09 nic :)
11:10 memowe It's called "Internet", version 2.3.3
11:12 nic hmm, I might need to borrow your Wildfire S
11:12 nic I'd give it back for weekends of course
11:12 memowe Thanks, that's awesome.
11:13 nic anyway, thanks, that tells me it's time to go browser hunting again
11:13 memowe Good luck!
11:13 purl You'll need it.
11:15 nic opera mobile is usually the bees nostrils, but it completely chokes on github pages
11:18 ryozi joined #mojo
11:23 daxim joined #mojo
11:29 mandreacchio joined #mojo
11:30 Lucas11 joined #mojo
11:31 Kwa sri: Would you say the best way for me to have the opt-in mobile design would be to simply set the format in the session? If the mobile format exists, it renders x.mobi.ep instead of x.html.ep?
11:32 good_news_everyone joined #mojo
11:32 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/SiwvDQ
11:32 good_news_everyone [mojo/master] better CGI tests - Sebastian Riedel
11:32 good_news_everyone left #mojo
11:33 sri Kwa: i have no opinion about that
11:34 Kwa Fair enough.
11:34 * sri wonders how realistic our CGI tests are these days https://github.com/kraih/mojo/b​lob/master/t/mojo/request_cgi.t
11:34 sri would be nice to have them reviewed and checked again
11:40 good_news_everyone joined #mojo
11:40 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/Z6LNJQ
11:40 good_news_everyone [mojo/master] mention browser versions - Sebastian Riedel
11:40 good_news_everyone left #mojo
11:43 sri or some updated replacement test cases would be nice too
11:43 pingup3rl joined #mojo
12:05 pingup3rl joined #mojo
12:42 fhelmber_ joined #mojo
12:53 Lucas1 joined #mojo
12:56 jnap joined #mojo
12:58 mire_ joined #mojo
13:31 d4rkie joined #mojo
13:33 sri *crickets*
13:40 bluescreen joined #mojo
13:50 kenjiskywalker joined #mojo
13:52 gryphon joined #mojo
13:52 pingup3rl joined #mojo
14:06 vel joined #mojo
14:08 marty joined #mojo
14:11 marty joined #mojo
14:39 knshaum joined #mojo
14:42 nic (I might be able to help with that, but didn't want to say so cos it'll be at least three weeks from now)
14:42 nic next month I'm supposed to look at replacing some old cgi nonsense, so it could work out
14:42 sri \o/
14:44 nic I was chuckling earlier when I counted how many times in the last 18 months you've mentioned that it might be nice if someone would modernise MojoX::Run :D
14:46 sri it would be nice if someone modernized MojoX::Run
14:48 nic :D
14:49 nic I see zpmorgan has attacked it, but his pull request hasn't been acted on yet
14:49 nic zpmorgan++
14:53 marty joined #mojo
15:28 Adura joined #mojo
15:46 Vandal joined #mojo
15:49 sri ohoh, looks like there will be a successor for the BEAST attack soon :S
15:50 * sri wonders if IO::Socket::SSL allows compression to be disabled at all
15:50 sri http://hackerne.ws/item?id=4510829
15:52 xaka1 joined #mojo
16:05 osprey_ joined #mojo
16:07 Adura Change the cookie value every request.
16:07 purl Adura: that doesn't look right
16:07 Adura Not sure of the value of such a script, purl.
16:07 purl Adura: huh?
16:07 sri hmm, sessions actually do that :)
16:08 Adura Thought they did.
16:08 sri we might be immune \o/
16:08 sri purl isa bot ;)
16:08 sri botsnack!
16:08 purl :)
16:13 vel joined #mojo
16:17 augensalat Addendum to the file upload discussion: IE might put absolute paths in the filename field, i.e. http://www.coderanch.com/t/477285/Servlets/​java/Uploading-file-FileUpload-control-was
16:17 jzawodn joined #mojo
16:23 labrown joined #mojo
16:42 jnap joined #mojo
16:43 osprey_ sri: can you help with SOCKS proxy? I don't want use LWP for only socks
17:41 nferraz joined #mojo
17:41 nferraz Hey guys,
17:41 nferraz Quick question that I've seen in a perl mailing list...
17:42 tempire the channel is silent during the apple keynote
17:43 tempire *crickets*
17:43 nferraz Actually, two questions: 1) What's the correct way to find the IP address of the remote user? 2) $self->tx->remote_address can be trusted, or can it be forged?
17:43 nferraz ah... Apple keynote... I completely forgot that... :)
17:56 sri sssht!
17:58 xaka1 can't believe the screen size is still so small :( google nexus has 1280x720 and the phone was released long time before iphone 5. every next presentation apple fails more and more
17:59 tempire I don't think so.  Apple has never been focused on features, but instead the product itself
18:00 tempire free iphone 4 on contract, that's pretty cool
18:01 xaka1 no new iPads i think?
18:21 knshaum joined #mojo
18:27 * sri is disappointed because everything interesting has been leaked before :(
18:29 osprey_ left #mojo
18:29 sri apple is losing its magic
18:30 sri lucky for them nobody else has figured out how to make quality products
18:32 nic I guess Vandal's ACC quote means that any technology distinguishable from magic is insufficiently advanced
18:49 tempire they'll be on top until they start slipping
18:50 tempire we're just used to them chugging away
18:50 tempire what I find interesting is that they're pushing a lot of boundaries forward in tech that usually just sits there
18:51 tempire computer construction, earphone tech, call quality on phones
18:51 tempire it's not the holy crap-consumer variety, but it's something that no other company has been doing.
18:51 sri that's not true
18:52 sri Nokia is doing it too atm
18:52 sri wireless charging is pretty damn awesome
18:52 sri their software and marketing may be garbage, but the engineering is top
18:53 sri i'm really surprised apple doesn't do wireless charging
18:54 tempire I'm not.  They're always last to the table in terms of the latest hardware hipness.
18:54 tempire feature-wise, at least
18:55 pingup3rl joined #mojo
19:02 rhaen hm.
19:03 nic zpmorgan: In your fork of mojox-run, does "make tests" work for you?
19:10 zpmorgan nic, I think so. Lemme test.
19:10 zpmorgan i'm certain that 'prove' worked
19:12 zpmorgan yeah, 'make test' works for me, under linux.
19:18 nic weird.  Which version of perl & Mojolicious is that with?
19:18 zpmorgan mojolicious 3.40, perl 5.16.0
19:19 zpmorgan how's it failing?
19:28 nic hm, it actually looks much less bad than I thought
19:30 nic "make test" gave what appeared to be a page of errors and the a 'FAIL'
19:30 nic but running each one as "prove t/??.t"
19:31 nic -perlsub.t gives a lot of errors but then ends with 'PASS', so that may be by design
19:31 zpmorgan oh, it spewed that stuff before I touched it. I know for a fact that it functions though.
19:32 nic then -exit-signal.t ends in 'FAIL'
19:32 nic but probably just cos it can't find common::sense
19:32 zpmorgan ah :)
19:32 nic so I'm guessing if I install that then the whole thing may give a 'PASS'
19:33 nic thanks for helping with that
19:33 nic btw, I saw that bfg posted on twitter (to sri) about 10 hrs ago
19:33 knshaum joined #mojo
19:34 zpmorgan my pleasure :)
19:34 nic so I've asked him if he's still actively maintaining this
19:34 nic I'm guessing xaka1 and others on here will be interested to pick up your fork in the meantime
19:36 andrefs joined #mojo
19:42 nic yay
19:43 nic "make test" still hangs till I <ctrl>+<c>
19:43 zpmorgan nic, requiring common::sense :)
19:43 nic but "prove t" ends in 'PASS'
19:43 nic got that now
19:43 rhaen hm, good evening mojo.
19:44 rhaen here is a little question. I just looked into Mojo::IOLoop and I am a little bit confused.
19:44 rhaen https://gist.github.com/aa9d12ab055f6f8347ee
19:44 rhaen First question: The client never gets to the "say - inside client part", why?
19:45 rhaen turning on the debugging on shows that it's looping around.
19:45 rhaen 2nd - the client will fail as there is no port 3000 on my server (correct?)
19:45 zpmorgan nic, that's a bit troubling. Does the child process have a log? Maybe you can tell if it's getting stuck.
19:45 rhaen how do I start the mojo loop correct?
19:46 marcus sri: getting a new phone?
19:47 zpmorgan nic, but I suspect that it's listening for sigchld or something.
19:47 sri marcus: dunno
19:47 nic yeah, was thinking that
19:48 sri rhaen: works for me
19:48 rhaen huh.
19:48 rhaen Does your client say "Inside client"?
19:48 sri yes
19:49 rhaen hm. Let me check my mojo version, my perl and my OS.
19:49 rhaen sri: and then go shopping a new ipod nano.
19:49 rhaen :)
19:49 rhaen hm. perl-5.10.1
19:50 sri awkward, it fires twice for me
19:52 GabrielVieira joined #mojo
19:52 GabrielVieira Hey, how can I use write_chunk to print the layout header first?
19:53 GabrielVieira Can I just specify the header.html.ep or something like that? Or do I need to open the file and send its content to write_chunk?
19:55 zpmorgan nic, what's the difference between 'prove' & 'make test'? why would one hang like that?
19:55 nic I'm going to pastebin both outputs
19:56 nic I've modified Makefile so that the only test is t/mojox-run-exit-signal.t
19:56 zpmorgan & 'prove -v'..
19:56 nic and "make test" still hangs
19:56 nic and "prove t" still passes
19:59 nic http://hastebin.com/wusowebave.vhdl is the "make test" and you can see where I've hit <ctrl>+<c>
20:00 zpmorgan what a mess.
20:00 purl PLEASE DO NOT MAKE A MESS IN HERE
20:01 rhaen :)
20:02 rhaen The bot is still something awesome.
20:02 rhaen tail -f ~/perl5/perlbrew/build.perl-5.16.1.log
20:03 nic and http://hastebin.com/muniqegaja.vhdl is the output from "prove -v t/mojox-run-exit-signal.t"
20:04 tempire marcus: does nordaaker provide staff with new iphones on every release?
20:04 * tempire moves to norway
20:05 nic Is the clue right at the beginning, the signal handler difference?
20:05 rhaen tempire: oh, it's cold there. And always dark.
20:05 tempire a friend of mine got a new iphone last week when he started his new job
20:05 tempire *facepalm*
20:05 rhaen tempire: oh, same here.
20:05 marcus tempire: this time we'll only offer to pay half, for employees who want to upgrade, and pull the rest from the paycheck.
20:06 rhaen tempire: move to Germany!
20:06 tempire no, I meant on every mojolicious release
20:06 zpmorgan nic, I can't tell if exit_cb is even being called. On line 21 of the test, maybe you have it print $i and $elem...
20:06 marcus tempire: that's crazy talk :)
20:06 tempire after about a week these things get scratched, and I'm too fabulus for scratches :)
20:06 nic zpmorgan: k, I can do that; need to grab something to eat first :)
20:06 tempire that's a pretty sweet policy though, if you were serious.
20:07 tempire way to keep the kids motivated
20:07 tempire marcus++
20:07 marcus tempire: my daughter tried to polish our wood floor with my iphone. :-/
20:07 marcus tempire: yeah, I was serious.
20:07 tempire did it work?
20:07 rhaen haha.
20:07 marcus no, it left a lot of scratches in the phone.
20:07 marcus so I'll be happy to get a new one.
20:07 rhaen :(
20:07 zpmorgan iplones are great. if you stand on one it'll tell you how much you weigh.
20:08 marcus zpmorgan: what does the cracking sound mean?
20:08 marcus 'too much' ?
20:08 xaka1 joined #mojo
20:08 tempire a new kid or phone?
20:08 tempire 'cause both seems to be handled
20:08 marcus tempire: I already got a new kid, I meant the new phone.
20:08 marcus 16:9 seems ideal for porn.. I mean movies.
20:09 xaka1 especially with those new ears, you gonna have perfect porn sound
20:10 nour funna gun
20:18 nferraz hi guys... now that Apple's distortion field was reduced...
20:19 sri rhaen: i think you might have found a leak
20:19 nferraz I had two questions from a new user in a perl mailing list: 1) What's the correct way to find the IP address of the remote user? 2) $self->tx->remote_address can be trusted, or can it be forged?
20:23 nferraz (actually question 2 partially answers question 1... just wanted to know if $self->tx->remote_address can't be falsified)
20:24 rhaen sri: hm, I have a talent to break everything. :(
20:24 xxtjaxx_ joined #mojo
20:24 edenc nferraz: why don't you point him to the channel instead of asking questions on his behalf? the double-hop in communication can be confusing
20:25 rhaen sri: Mojo version is 3.35 (here) - I am updating everything now.
20:26 nferraz hi eden - I did that, but I'm also curious about the answer :)
20:29 rhaen sri: tell me when you need more infos. I am not sure if this is relating, this host has IPv6 enabled and makes a use of it.
20:29 sri rhaen: i can already replicate it
20:30 rhaen k
20:31 rhaen sri: ok, after updating to 5.16.1 and Mojolicious 3.40 - everything is working here.
20:32 sri rhaen: i very much doubt that
20:32 rhaen hm. well, now it says - inside client.
20:33 sri it only says that with IO::Socket::IP
20:33 sri and only because it leaks
20:33 rhaen ah! got it.
20:33 sri i can't find the leak though
20:37 rhaen hm.
20:43 nic zpmorgan: http://hastebin.com/qagodigeco.sm is the output
20:46 zpmorgan nic, does that make any sense to you?
20:46 nic not yet :)
20:47 zpmorgan does it make sense under prove?
20:48 nic but it does highlight a difference between the two
20:49 nic with prove the output doesn't finish on 1..5 but on 1..6
20:49 zpmorgan there should be 7 lines instead of 22
20:50 zpmorgan nic, that could just be a race condition. maybe you could have it print $i after it's incremented.
20:50 nic yeah, i'll do that
20:51 good_news_everyone joined #mojo
20:51 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/Ey4nfw
20:51 good_news_everyone [mojo/master] improved resilience of Mojo::IOLoop exception handling - Sebastian Riedel
20:51 good_news_everyone left #mojo
20:51 sri it was not a leak, just some awkward side effects
20:54 xxtjaxx_ joined #mojo
20:55 mattp joined #mojo
20:57 pingup3rl joined #mojo
20:57 nic http://hastebin.com/pecoqawuqo.sm is from 'make test'
20:57 nic and http://hastebin.com/degokomaji.sm is from 'prove'
20:58 zpmorgan nic, my guess is that 'make test' is interfering with your child processes' ability to exit(0). If they don't exit, they'd behave like a fork after their 'cmd' finishes, and each child will continue spawning children.
20:58 zpmorgan that's why there are 22 lines. 1+1+2+3+4+5+6=22
20:59 zpmorgan maybe you could see if it's running after exit(0)
20:59 zpmorgan or if 'cmd' is being called at all
21:00 nic would that explain why the last bundle of output is missing for make?
21:01 berov joined #mojo
21:02 zpmorgan that would explain why it hangs. each child would be waiting for 7 sigchld's
21:02 zpmorgan I can't imagine what would interfere with exit(0). You're on your own there.
21:04 sri hmmmmm
21:04 sri complicated design problem
21:04 sri perl -Ilib -Mojo -E 'Mojo::IOLoop->client({port => 3000} => sub { die }); Mojo::IOLoop->start'
21:04 sri the error from the die call gets eaten by Mojo::EventEmitter atm
21:05 sri since the invocation of the callback was done from an error event, which can't recurse
21:06 nic crumbs, the more I work on this the more confused I get; I moved the open/close and added another print; now the output looks _completely_ different
21:07 zpmorgan nic, you're not on windows, are yoy?
21:08 mattp joined #mojo
21:08 zpmorgan oh, i see that you're not
21:08 nic heh, that would make sense, but no, linux
21:09 nic it would be nice if it worked this much on Win
21:13 GabrielVieira Hi... write_chunk makes the enconding a mess.. With normal rendering it returns UTF-8, but with write_chunk its I think its returning ISO-8859-1. Even if I set $self->res->headers->content_ty​pe('text/html;charset=UTF-8'); the browser shows a mess.
21:13 nic if I can believe the output, it's suggesting that exit_cb doesn't get called under 'make'
21:16 zpmorgan where does it print $i and $elem?
21:24 good_news_everyone joined #mojo
21:24 good_news_everyone [mojo] kraih pushed 1 new commit to master: http://git.io/34XbBA
21:25 good_news_everyone [mojo/master] improved Mojo::EventEmitter to warn about failed error events - Sebastian Riedel
21:25 good_news_everyone left #mojo
21:25 vervain Is there a way to address a form in a Mojo::Dom and use it's data to generate a Mojo::UserAgent POST without extracting and rebuilding the 'data map'?
21:26 zpmorgan nic, my test passes, but it's still completely wrong: http://hastebin.com/pojotohiki.hs
21:26 zpmorgan children shouldn't be spawning children.
21:28 nic http://hastebin.com/mixumadoko.coffee shows code & output
21:28 nic but for 'make' only the first 7 lines of output appear; the ones from exit_cb don't appear
21:29 nic What I don't get is that previously we were seeing output from exit_cb, and all I've changed is the debugging prints
21:29 sri rhaen: those two commits should solve all problems
21:32 * sri sets nic on fire
21:33 GabrielVieira Hi... write_chunk makes the enconding a mess.. With normal rendering it returns UTF-8, but with write_chunk its I think its returning ISO-8859-1. Even if I set $self->res->headers->content_ty​pe('text/html;charset=UTF-8'); the browser shows a mess.
21:39 lukep joined #mojo
21:53 nic zpmorgan: If you grep ' spawned ' in your debug output, does it appear 7 times?
21:53 nic I'm wondering if the qty proc spawned is actually correct, and it's just the output which is messed up
21:54 nic I thought it might be a confusion of EV.pm, so I moved to a fresh box rather than my over-installed dev box
21:54 nic expected success
21:54 nic got worse failure; now it hangs both ways, the same
21:55 nic The "Bad file descriptor" messages seem to correlate with overall failure
21:55 nic so maybe the output is missing cos the spawned proc lost contact with file descriptors
21:56 * nic clutches at more straws
21:56 zpmorgan nic, thanks for the confusion :)
21:56 nic :D
21:57 GabrielVieira joined #mojo
21:57 zpmorgan can sigchld accidentally signal other children?
21:58 zpmorgan or are you only seeing 'spawned' 7 times because the childrens' logs are collected with their stdouts?
21:59 zpmorgan I'm leaning towards the former
21:59 nic well, we do know that sig handling is dodgy
21:59 zpmorgan http://hastebin.com/yifuwibuji.hs
22:01 nic interesting.  i'm thinking to give up on this for today, and tomorrow try ripping the code apart
22:02 Jason_ joined #mojo
22:03 zpmorgan thanks, nic.
22:03 nic zpmorgan: In your debug output (the stuff put there by the module, not the stuff we've written) do you get 'spawned' more than 7 times?
22:03 zpmorgan nope, but debug goes to stdout, and the parent process collects stdout
22:03 nic i see
22:04 * nic &
22:05 Jason_ Hi all, first time listener, first time caller...I'm trying to set some universal session data in the startup of my app, but I can't seem to figure out how to make it work. Is there a good way to set session data at startup in Mojo?
22:23 Adura I made a route that all other routes use to do just that.
22:24 Adura With under.
22:24 Adura No idea if that's the best way to do it, but it works...!
22:24 Jason_ I'll take a look into that, thanks.
22:31 Averna joined #mojo
22:31 jnap joined #mojo
23:01 berov left #mojo
23:49 mandreacchio joined #mojo
23:53 zpmorgan joined #mojo
23:54 asarch joined #mojo

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