Camelia, the Perl 6 bug

IRC log for #mojo, 2011-04-17

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

All times shown according to UTC.

Time Nick Message
00:39 trcjr left #mojo
00:47 Alias joined #mojo
00:50 Alias__ left #mojo
02:06 alnewkirk joined #mojo
02:18 daviddelikat1 left #mojo
02:50 daviddelikat joined #mojo
03:05 estranho joined #mojo
03:47 perlrocks Twitter: "CVE request: Mojolicious directory traversal vulnerability: Could a CVE be assigned to the directory traversal f... http://bit.ly/dUknC1" --oss_security http://twitter.com/oss_secur​ity/status/59462908693839873
04:58 avkhozov joined #mojo
06:02 baton8 joined #mojo
07:00 Foxcool joined #mojo
07:39 esskar2 joined #mojo
07:40 esskar left #mojo
07:53 arpadszasz joined #mojo
08:14 daviddelikat1 joined #mojo
08:18 daviddelikat left #mojo
08:28 arpadszasz left #mojo
09:02 esskar joined #mojo
09:03 esskar2 left #mojo
10:04 abra left #mojo
10:14 * sri yawns
10:15 abra joined #mojo
11:13 GitHub145 joined #mojo
11:13 GitHub145 mojo: master Sebastian Riedel * 8e17df1 (1 files in 1 dirs): pod tweak - http://bit.ly/eoBsSb
11:13 GitHub145 left #mojo
12:25 TheCritic left #mojo
12:33 GitHub92 joined #mojo
12:33 GitHub92 mojo: master Sebastian Riedel * c8670bd (2 files in 2 dirs): added experimental dnt (Do Not Track) method to Mojo::Headers - http://bit.ly/gQTNj6
12:33 GitHub92 left #mojo
12:39 GitHub185 joined #mojo
12:39 GitHub185 mojo: master Sebastian Riedel * 40fd31b (1 files in 1 dirs): added test - http://bit.ly/fetJjX
12:39 GitHub185 left #mojo
12:52 Akron joined #mojo
13:14 Akron left #mojo
13:14 chansen left #mojo
13:22 chansen joined #mojo
13:33 yegor left #mojo
13:43 avkhozov left #mojo
14:03 yegor joined #mojo
14:31 Foxcool left #mojo
14:35 Foxcool joined #mojo
14:42 uniejo left #mojo
14:56 Foxcool_ joined #mojo
14:56 Foxcool left #mojo
14:58 Foxcool_ is now known as Foxcool
15:01 stephanj What is the correct way of attaching STDIN to a Mojo::IOLoop? I have this code http://pastebin.com/ZVC0iCdU , which gives me this error: http://pastebin.com/z3FTqJSk but when i call the script with something on stdin i get no error: http://pastebin.com/jTpU2Rai
15:05 yko stephanj: i think yo ucan find dsamples at MojoX::Run
15:05 yegor left #mojo
15:07 sri you also need non blocking handles
15:07 sri it's a rather complicated topic
15:09 yegor joined #mojo
15:10 sri also make sure to use a recent version, none of your errors relate to line numbers in the current version
15:14 kaare joined #mojo
15:17 sri anyway, the whole thing is much much more complicated than you might expect
15:18 stephanj okey
15:35 Akron joined #mojo
15:40 sri actually, none of the errors seem to relate to any recent version, so nobody will be able to help you
15:43 stephanj http://pastebin.com/gYZfx7s7
15:44 stephanj yes i had 1.12 upgraded now to 1.16(i thought i upgraded before but had only done it in a local::lib dir)
15:48 mattastrophe left #mojo
15:51 mattastrophe joined #mojo
15:51 perlrocks Twitter: "YAPP # 3: Interview with Anatoly Sharifulin http://j.mp/ganmwt # perl # yapp # podcast # interview # Mojolicious" (ru) --yapp_ru http://twitter.com/yapp_ru​/status/59645160170078208
15:53 sri stephanj: there is so much wrong i don't know where to start, stuff like on_write doesn't even exist
15:54 stephanj okey
15:55 sri stephanj: i think it might be best if you start with a general purpose event loop that actually has the right documentation
15:55 sri POE or so
15:55 tabbi joined #mojo
15:56 tabbi left #mojo
15:56 sri we are not aiming to be a general purpose event loop and therefore don't have the tutorials
16:04 perlrocks Twitter: ""@ Yapp_ru: YAPP # 3: Interview with Anatoly Sharifulin http://t.co/Nc559HY # perl # yapp # podcast # interview # Mojolicious"" (ru) --sharifulin http://twitter.com/shariful​in/status/59648391910662144
16:06 sri wow, the russians have their own perl podcast :o
16:07 yko :D
16:07 yko why shouldn't we?
16:11 arpadszasz joined #mojo
16:21 sri stephanj: there is also an ioloop bug though
16:21 stephanj mh?
16:21 purl mh is a unix commandline-based mail user agent or at http://www.ics.uci.edu/~mh/ or Men's Health magazine or a "women's magazine" for men or Getty's way of saying hm or stephan48's way of saying hm
16:21 stephanj it always writes to the handle?
16:21 sri actually two
16:21 purl actually two are .tar.gz but that's not relevant to his question
16:21 sri right
16:22 sri sockets start writing so we know when they are connected
16:22 stephanj ye
16:22 sri for other handles that is not such a good idea
16:23 sri i can fix it, but i don't see a way to test this portably
16:23 stephanj ye
16:23 stephanj maybe a flag readonly?
16:23 sri isa IO::Socket check
16:23 sri i mean an actual unit test
16:23 sri so it doesn't break again
16:23 stephanj k
16:24 sri sockets should always be isa IO::Socket, that's no problem
16:25 stephanj maybe throw a  RO IO::Handle at it, and when it throws a bad file descriptor error, and write appears in stacktrace, fail the test?
16:26 sri bit naive, but better than nothing i guess
16:26 sri the fact that we support poll, epoll and kqueue doesn't make testing any easier
16:27 stephanj ye
16:27 stephanj #
16:27 estranho left #mojo
16:33 stephanj and the second one?
16:33 purl the second one is, like, better - the first could be confused with trying to execute an unknown opcode or something
16:36 sri ?
16:36 stephanj you saied that there are two bugs
16:37 stephanj but named only one
16:39 sri complicated
16:39 purl hmmm... complicated is for winner!
16:39 sri much more important is a test case now
16:49 estranho joined #mojo
17:01 GitHub3 joined #mojo
17:01 GitHub3 mojo: master Sebastian Riedel * e9c1928 (3 files in 3 dirs): fixed readonly handle support in Mojo::IOLoop - http://bit.ly/htof7e
17:01 GitHub3 left #mojo
17:03 janus uh, nice one!
17:03 purl geezer!
17:03 sri it's all experimental
17:03 janus stephanj++ # anyways
17:04 sri if you're interested in stuff like that make sure to send test cases! ;)
17:04 sri we are also still missing a udp client/server test case
17:04 purl okay, sri.
17:04 * sri pokes purl
17:04 purl stop that you naughty sri
17:31 sri actually a portable open3 example controlling a second ioloop example would be a great test case
17:43 stephanj janus++ # also responsible for this :)
17:48 dku this is driving me crazy... does anyone see anything wrong with this? "perl app.pl daemon --listen http://*:13000/"
17:48 sri haha
17:48 sri trailing slash
17:48 purl i think trailing slash is converted to line continuation char on win
17:48 sri github master throws an error for that i think
17:49 dku ahh works now, thanks :)
17:49 sri "Invalid listen value "http://*:13000/"..."
17:49 sri that one has become soooo common, i might even add a faq entry
17:49 sri or actually...
17:49 sri just support it -.-
17:50 stephanj or make the server warn on invalid values...
17:50 sri see above ;p
17:55 stephanj ah
17:55 stephanj last time i checked it didnt but also this might be on a really old version
17:56 dku see above again? ;)
17:56 stephanj :)
17:56 sri :D
17:57 * stephanj should learn to read
18:08 estranho left #mojo
18:08 perlrocks Twitter: "@kraih russian perl podcast about Perl, I told about #Mojolicious :) http://yapp.rpod.ru/232126.html" --sharifulin http://twitter.com/shariful​in/status/59679653245358080
18:15 perlrocks Twitter: "We talked about the * web framework - Mojolicious * MVC in Mojolicious * ORM for Mojolicious * NoSQL and perl * Project ... http://goo.gl/CHHWX" (ru) --volcanotonga http://twitter.com/volcanoto​nga/status/59681307566931968
18:28 perlrocks Twitter: "Perl / [PODCAST] Yet Another Perl Podcast # 3: Интервью с Анатолием Шарифулиным: Мы говорили оweb framework - Mojolicious MVC Mojoli в ..." (it) --GreenDayItalia http://twitter.com/GreenDayIt​alia/status/59684784137441280
18:44 Foxcool left #mojo
18:48 daviddelikat joined #mojo
18:52 daviddelikat1 left #mojo
18:53 Akron left #mojo
18:54 TheCritic joined #mojo
19:06 Alias_ joined #mojo
19:10 Alias left #mojo
19:57 tabbi joined #mojo
19:57 tabbi left #mojo
20:15 marcus wasssup
20:15 marcus anything exciting happened the last week?
20:18 sri just the usual
20:28 tempire marcus: some people accused sri of something, sri used it for marketing, some japanese people tweeted weird stuff.
20:29 tempire indeed, the usual
20:29 sri oh and we trolled chromatic
20:33 Akron joined #mojo
20:55 perlrocks Twitter: "perl-Mojolicious-0.999929-2.fc14 - http://rkrh.kr/1l6" --fedorasecurity http://twitter.com/fedorasecu​rity/status/59721558750535680
20:55 sri oO
20:55 sri seriously?
20:55 purl is that for real?
20:56 sri wow... just wow
21:05 GitHub4 joined #mojo
21:05 GitHub4 mojo: master Sebastian Riedel * 3a3867a (2 files in 2 dirs): added experimental UDP support to Mojo::IOLoop - http://bit.ly/flxYbf
21:05 GitHub4 left #mojo
21:06 sri for better test coverage :)
21:06 sri since the resolver is here to stay
21:13 arpadszasz left #mojo
21:15 stigtsp left #mojo
21:21 stigtsp joined #mojo
21:30 sri hmm, think i'll undo the commit, api is just too wrong
21:34 kaare left #mojo
21:48 sri yea
21:53 metaperl1 joined #mojo
21:57 metaperl left #mojo
21:57 stephanj sri: wtf? thes still have pre 1.0 versions?!
22:21 marcus sri: Any plans to make your mongodb work public?
22:21 sri marcus: of course
22:22 sri but there are so many distractions
22:26 TheCritic left #mojo
22:26 marcus but I've been offline for a week!
22:26 marcus :)
22:27 marcus where did you troll chromatic? Twitter?
22:27 sri right now for example i'm tempted to tear ioloop apart
22:28 marcus sri: you keep doing that
22:28 marcus :)
22:28 sri that module needs to be 5 or so modules
22:28 marcus won't that make it slower?
22:28 sri not unlikely
22:29 marcus what's the gain? Sanity?
22:29 sri but also easier to extend and replace with different event loops
22:30 sri short term gain is rather slim
22:33 marcus sounds boring
22:33 sri i know i should be hacking on shiny new high level stuff :S
22:33 marcus yepp
22:35 marcus I started on a mojolicious port for iusethis, but it's a pretty big task.
22:37 marcus will probably do iwatchthis first.
22:37 sri neat
22:39 marcus would like to switch iwt over to mongodb from dbic while I'm at it tho.
22:39 marcus but maybe better to do framework port first
22:40 * sri nods
22:49 sri marcus: did you see my Coro example?
22:50 sri https://gist.github.com/917702 # that's some really fancy stuff
22:50 sri non-blocking requests and database queries from within templates without callbacks
22:51 marcus sri: that is fucking fancy
22:51 sri blocking 6 rps, non-blocking 60 rps
22:52 abra left #mojo
22:55 marcus sri: I guess that depends on the number of children tho ? :)
22:55 marcus sri: with 60 children, I imagine blocking would get better results :p
22:56 sri unlikely, and it would cost a lot more resources
22:56 sri a coroutine costs about 4k
22:57 marcus the api is very clean
22:57 marcus what's the cpu wise costs of a coro?
22:57 sri minimal according to my tests so far
22:58 sri i've had a pretty bad crash though
22:58 sri Coro took down the server
22:58 marcus ouch
22:59 sri it's fiddling with perl internals from C pretty bad
22:59 sri multiple call stacks and stuff
22:59 sri i think that's the biggest risk
23:01 sri aside from that i'm a huge fan of coroutines, they make non-blocking io awesome
23:08 sri hmm, talking about coroutines makes me want to work on high level stuff again :D
23:09 sri first sleep though, nn

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