Camelia, the Perl 6 bug

IRC log for #mojo, 2011-05-12

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

All times shown according to UTC.

Time Nick Message
00:03 k-man_ left #mojo
00:03 k-man_ joined #mojo
00:04 k-man_ left #mojo
00:04 k-man joined #mojo
00:06 roman left #mojo
00:39 Alias joined #mojo
00:46 tempire say pop;
01:02 metaperl_ joined #mojo
01:09 xaka left #mojo
01:45 alnewkirk ok, so wtf
01:45 alnewkirk where do i put the FastCgiIpcDir directive?
01:46 alnewkirk VirtualHost ... no, Directory .. no, ... ? um
02:43 daviddelikat joined #mojo
02:56 tempire dernit
02:56 tempire http://googlefight.com/index.php?l​ang=en_GB&word1=three+letter+n​icks&word2=other+letter+nicks
03:05 Alias left #mojo
03:07 PerlNinja joined #mojo
04:05 dwierenga joined #mojo
04:14 dwierenga left #mojo
04:14 dwierenga joined #mojo
04:22 dwierenga left #mojo
05:21 ysyrota joined #mojo
05:36 edenc left #mojo
05:37 edenc joined #mojo
05:42 robinsmidsrod left #mojo
05:42 diegok left #mojo
05:43 diegok joined #mojo
05:44 robinsmidsrod joined #mojo
05:55 Christian joined #mojo
06:05 gshank_ joined #mojo
06:05 Artrey joined #mojo
06:05 gizzlon_ joined #mojo
06:05 cxreg2 joined #mojo
06:05 edenc left #mojo
06:05 koban` joined #mojo
06:06 cxreg left #mojo
06:06 gshank left #mojo
06:06 gizzlon left #mojo
06:06 Kallisti left #mojo
06:12 edenc joined #mojo
06:31 cxreg2 is now known as cxreg
06:46 AmeliePoulain joined #mojo
07:00 fhelmber_ joined #mojo
07:22 arthas joined #mojo
07:36 PerlNinja left #mojo
07:54 marcus hurgle
07:56 crab sounds painful.
08:00 mattastrophe joined #mojo
08:06 mattastrophe left #mojo
08:35 AmeliePoulain left #mojo
08:36 AmeliePoulain joined #mojo
08:37 AmeliePoulain left #mojo
08:38 AmeliePoulain joined #mojo
08:38 AmeliePoulain left #mojo
08:39 AmeliePoulain joined #mojo
08:40 AmeliePoulain left #mojo
08:40 tabbi joined #mojo
08:40 tabbi left #mojo
08:40 AmeliePoulain joined #mojo
08:41 AmeliePoulain left #mojo
08:42 ajgb joined #mojo
08:42 AmeliePoulain joined #mojo
08:42 AmeliePoulain left #mojo
08:43 AmeliePoulain joined #mojo
08:43 AmeliePoulain left #mojo
08:44 AmeliePoulain joined #mojo
08:44 AmeliePoulain left #mojo
08:56 memowe A HOI
09:14 yko tempire: no, they don't. i don't see way how to load epl snippets into perl filetype at the moment
09:15 yko but if you find good not too hackish way just let me know
09:20 marcus YO HO HO AND A BOTTLE OF RUM!
09:21 memowe Whoops. Typo.
09:21 memowe O HAI
09:21 zakame joined #mojo
09:22 * sri yawns
09:42 AmeliePoulain joined #mojo
09:43 AmeliePoulain left #mojo
09:46 tabbi joined #mojo
09:58 cosmincx joined #mojo
10:03 tabbi left #mojo
10:16 crab i wish i'd killed the wasp that stung me on my chin.
10:18 yko ouch
10:38 sh4 joined #mojo
11:23 koban` left #mojo
11:37 GitHub197 joined #mojo
11:37 GitHub197 mojo: master Sebastian Riedel * 61c3a79 (3 files in 3 dirs): improved documentation for Mojo::DOM collections - http://bit.ly/k0R8GV
11:37 GitHub197 left #mojo
11:38 Akron joined #mojo
11:42 GitHub124 joined #mojo
11:42 GitHub124 mojo: master Sebastian Riedel * 6e39351 (1 files in 1 dirs): fixed typo - http://bit.ly/ll7gAb
11:42 GitHub124 left #mojo
12:22 yegor joined #mojo
12:22 yko sri: what meaning has pair_separator in Mojo::Parameters? Looks like it's not used anywhere
12:23 sri yko: just look where it is used ;p
12:24 yko maybe it's better to interpret as separator '&' or ';' but not both that characters?
12:24 moritz $self->pair_separator(';') if $string =~ /\;/ && $string !~ /\&/;
12:24 yko i think pair_separator supposed to be used here -> https://github.com/kraih/mojo/blob/​master/lib/Mojo/Parameters.pm#L111
12:25 yko but it doesn't
12:25 sri no it's not
12:25 sri it works as intended
12:25 sri as far as i can see
12:25 yko "/?foo=bar;baz&x=2" => "/?foo=bar&baz=&x=2"
12:26 yko is this behavior intended?
12:26 sri yes
12:26 yko ah. than ok
12:27 sri feel free to dive into the specs and come up with a better solution though
12:27 yko by the way, could you tell me where & and ; specified? just wanna read that
12:27 yko it's just looks not very logically
12:28 sri good luck finding those specs ;p
12:28 yko ok
12:28 yko searched already, but didn't found
12:28 sri afraid i don't remember
12:28 yko ok, np
12:28 sri somewhere in the old w3c drafts or early url rfcs maybe
12:29 sri finding specs for urlencoded and multipart formdata is just as horrible
12:32 elb0w oh really you can use ; opposed to &?
12:33 moritz CGI.pm supports it, for example
12:33 moritz I find it convenient because you have to escape & in URLs in xhtml
12:33 moritz but not all server software supports it
12:37 yko ok. i did some searches before talking to you. and did after. all what i found on this topic was that both '&' and ';' are 'reserved' in query component and one interesting thing
12:38 yko http://www.w3.org/TR/1999/REC-html401-19991​224/appendix/notes.html#ampersands-in-uris recommends to use ';' in place of '&'
12:40 yko i'd rather interpret that as tr/&/;/
12:41 sri i actually don't think there is a "proper" way to support both, just recommendations here and there
12:41 sri the pair separator argument is prolly rather useless as long as we parse both
12:42 sri since there is no proper roundtrip integrity anyway
12:42 sri make sure to check out the old url rfcs though
12:42 yko and if you look one more time into M::Parameters you may found that it actually doesn't affect anything
12:42 yko sure. i'm working on it
12:43 sri those that came before uri... back in the early 90s or so
12:44 sri yko: that's not true
12:44 sri https://github.com/kraih/mojo/blob/​master/lib/Mojo/Parameters.pm#L215
12:44 sri how is that affecting nothing?
12:45 yko damn... right you are. i was blind
12:45 * sri pokes yko in the eye
12:46 * yko sparkles with another one
12:46 yko actually all what i want is client doesn't doesn't change passed url
12:47 yko when you pass one url and client requesting server with another one its not very good
12:47 sri you want the param module to not touch ";"?
12:48 yko i want url to be exactly the same on input and on output
12:48 sri don't think the specs allow that
12:49 yko there could be some workarround with serving undefined values of parameters but not sure you would like it
12:49 yko spec? what spec? show me one? :)
12:49 yko jsut kidding
12:49 sri the one you quoted earlier for example ;p
12:51 yko it defines query string as something that should be analyzed by the service. there's few characters reserved as subdelims. and few words that semicolon is better to use in place of ampersand
12:51 yko and nothing more. that's the spec :p
12:52 sri http://stackoverflow.com/questions/348​1664/semicolon-as-url-query-separator
12:53 sri good discussion about the topic
12:53 arthas left #mojo
12:54 Akron left #mojo
12:56 arthas joined #mojo
12:56 yko i don't see answers there. only problems :p
12:58 moritz well, there aren't any one-size-fits-all answers
12:59 moritz if the discussion gave answers where there are none, it wouldn't be good :-)
13:00 metaperl_ left #mojo
13:00 sri there is no easy answer ;p
13:01 Akron joined #mojo
13:05 yko interpreting '/?1;2&x=2' with Mojo::URL i'd say that my original url is damaged
13:16 dekimsey joined #mojo
13:18 gshank_ is now known as gshank
13:20 sri yko: or repaired ;p
13:20 sri i would say your original url was broken
13:22 sri it's a reserved character after all
13:22 yko that makes sense much more :)
13:26 arthas left #mojo
13:26 sri in fact, even if 1;2 was interpreted as a value it would not survive spec compliant escaping
13:27 arthas joined #mojo
13:27 sri perl -Mojo -E'say Mojo::Parameters->new(qw/la la foo b;ar/)'
13:28 sri la=la&foo=b%3Bar
13:29 koban` joined #mojo
13:29 tabbi joined #mojo
13:29 tabbi left #mojo
13:50 metaperl_ joined #mojo
13:51 yko ok, i wrote a lot in notepad but dont think it worth pasing, hehe. just last remark about this discussion, i think we both don't want to continue
13:52 yko /?foo=1;2;3;4&bar=baz '&' could delimit pairs, '=' could delimit key/values and ';' could delimit array elements and should not be escaped
13:52 cosmincx left #mojo
13:52 cosmincx joined #mojo
13:55 sri yko: are you refering to any existing spec or are you just loud thinking?
14:03 zakame left #mojo
14:11 toebu left #mojo
14:13 yko ok... i lean on http://tools.ietf.org/html/rfc3986 i you don't mind
14:13 yko It defines reserved characters "!" "$" "&"  "'"  "(" ")" "*"  "+"  ","  ";"  "=" as a part of query on page 49
14:13 yko 2.2.  Reserved Characters of same spec explains nature of reserved characters:
14:14 yko URIs include components and subcomponents that are delimited by characters in the "reserved" set
14:14 sri relying on a spec from 2005 is not very realistic ;p
14:14 yko and 2.4.  "When to Encode or Decode" tells that components and subcomponents of uri should be parsed and separated before decoding
14:16 yko well at least i think you can not find any spec that allows you to replace one separator character with another within query string
14:16 * sri doesn't follow
14:17 yko that's current behavior, we started with it
14:17 yko but you know, i think we bothjsut should go and have some beer
14:17 sri i don't know what you're refering to
14:18 yko perl -MMojo::URL -e 'print Mojo::URL->new("/?foo=1;2&bar=3")'
14:18 sri seems like a sensible result
14:18 yko ; is replaced with & but & already present in query string and it separates different level of values
14:19 sri your query string is invalid in the first place
14:20 sri really, i don't see your point
14:20 sri your problem is that an invalid query string gets turned into a valid one without dieing?
14:21 yko which part of query is invalid and which part of which spec is voilated?
14:24 sri quite sure you quoted it yourself
14:26 sri maybe this should be discussed in a github issue or the on the mailing list (or both)
14:27 sri s/the//
14:28 sri issue if you already have a solution in mind
14:28 koban` left #mojo
14:29 sri yko: suppose your proposal is to remove ";" support entirely?
14:31 Christian left #mojo
14:32 yko i don't seek for solutions until there's problem. since you don't consider this as problem i don't like to spent time on finding solution
14:32 amoore joined #mojo
14:33 sri yko: well, convincing me that there's a problem is your job ;p
14:34 yko looks like i failed
14:34 sri i think a mailing list discussion could help with that
14:35 sri well, i don't think there is a problem because there is no right or wrong in this case
14:35 sri we either support ";" as a separator, or we don't
14:35 sri for me it's just a strategic decision at this point
14:36 sri i would very much welcome a third solution if there was one and discuss it
14:36 yko maybe. i think i should dig more specs and formulate my question better
14:36 sri sounds good
14:37 yko it's really not a problem atm, i found solution for myself.
14:37 yko i just don't like way to find one day that i can not send query string formatted exactly as i need
14:38 sri i wonder of pair_separator is even tested
14:41 sri yea, it's even tested
14:44 abra joined #mojo
14:46 moritz yko: don't wonder, embrace the power of ack instead
14:47 sri yko: URI does the same btw
14:47 sri perl -MURI -MURI::QueryParam -MData::Dumper -E'say Dumper([URI->new("/?foo=1;​2&bar=3")->query_param])'
14:47 sri it considers 2 a key without value
14:49 sri python version too
14:51 stephan48 sri: is ; in RFC?#
14:52 sri as reserved character
14:52 stephan48 i mean as query param seperator?
14:52 sri http://www.w3.org/TR/1999/REC-html4​01-19991224/appendix/notes.html#h-B.2.2
14:53 sri w3c recommends it
14:53 stephan48 k
14:53 sri i think in the rfcs it is an unused reserved character
14:55 sri it seems like a very bad idea to remove support for it
14:56 sri at least from the parser
14:56 sri generating that kinda query string seems much less important
15:00 non joined #mojo
15:03 sh4 left #mojo
15:08 yegor left #mojo
15:09 kaare joined #mojo
15:10 perlrocks Twitter: "Mojolicious in the cloud: Hello DotCloud! - Sebastian Riedel about Perl and the Web http://icio.us/ABktsw" (ja) --mix3 http://twitter.com/mix3/status/68694544052523008
15:18 tabbi joined #mojo
15:32 tabbi left #mojo
15:37 fmerges joined #mojo
15:49 metaperl_ is now known as metaperl
15:53 xaka joined #mojo
16:10 ysyrota left #mojo
16:20 Foxcool joined #mojo
16:31 ajgb left #mojo
16:38 cosmincx left #mojo
16:43 sh4 joined #mojo
16:44 Akron left #mojo
17:11 fmerges left #mojo
17:31 sri hmm, been playing with sending log messages to Test::More::diag() when running in the test harness... but the result is rather disappointing
17:36 sri yko: while i'm quite sure the param parser is fine, there is something else you could do
17:36 yko there always a lot
17:36 * marty hands sri the whip
17:37 yko but please continue
17:37 sri the query string is by definition an arbitrary string, not necessarily a chain of key/value pairs
17:37 sri currently we don't have good support for the arbitrary string form
17:37 sri one could for example store the raw string by default, and only parse on demand
17:38 sri keeping the original query string as is
17:38 yko yeah, first they put you into deep dark pit, then they deny all your arguments and telling you are wrong. and then they tell you what to do and swinging with the whip...
17:38 yko sri: makes sense
17:38 purl somebody said makes sense was http://xrl.us/bjq3kq
17:38 marty and then finally, the alligator pit.  :p
17:38 sri wait...
17:38 sri yko is out of the alligator pit?
17:39 yko i've got keys
17:39 sri :o
17:39 yko and made few backdoors
17:39 yko i had a lot of time :p
17:39 marty <in my best Gollum voice>  yko has been sneeeeking.  :)
17:41 sri there might even be performance benefits with that solution
17:45 sri of course there's some work behind it, spec checking, making sure things don't get slower...
17:46 sri yko: in case you're still interested
17:46 yko i had problem. i solved it in another way :p
17:46 yko but well, i'm interested
17:49 * sri cracks the whip
17:49 sri :)
17:52 miyagawa sri: PSGI server seems to return $status=undef
17:52 miyagawa when an inline template is not found in Mojolicious::Lite
17:52 miyagawa is it a bug  or is my code wrong
17:52 miyagawa :)
17:53 sri if nothing is rendered it assumes delayed rendering by default, dunno how that plays out for the psgi binding
17:54 miyagawa it just breaks the PSGI protocol
17:54 miyagawa headers and body is set, just status is undef
17:54 miyagawa should be 404 i think..
17:54 sri i see
17:54 sri should be the same for cgi i guess
17:55 sri fastcgi too i guess
17:56 sri gonna default to 404
17:58 * miyagawa nods
18:02 GitHub155 joined #mojo
18:02 GitHub155 mojo: master Sebastian Riedel * 100caab (4 files in 2 dirs): fixed all blocking server bindings to default to a 404 status code - http://bit.ly/krN7ve
18:02 GitHub155 left #mojo
18:04 sri that should do the trick
18:13 fhelmber_ left #mojo
18:36 Akron joined #mojo
19:07 sh4 left #mojo
19:23 snearch joined #mojo
19:29 karamorf joined #mojo
19:30 karamorf I'm dealing with some legacy code and that code is going to print the html to STDOUT for one of my requests. Is there a nice way to deal with that or should I capture STDOUT and then send it as text to the render method?
19:40 metaperl I would definitely  capture STDOUT and then send it as text to the render method.
19:40 perlrocks Twitter: "11.20.25 Mojolicious Multiple Vulnerabilities - http://tinyurl.com/3mw5trg, #itsecurity" --itsectoday http://twitter.com/itsectod​ay/status/68762582575628288
19:41 metaperl ouch
19:44 Foxcool left #mojo
19:47 sri ancient history
19:49 sri but it's funny how we are now on everybodys radar
19:52 snearch left #mojo
19:55 GitHub115 joined #mojo
19:55 GitHub115 mojo: master Sebastian Riedel * c7a687f (4 files in 4 dirs): updated jQuery to version 1.6.1 - http://bit.ly/mNaOX6
19:55 GitHub115 left #mojo
19:58 sri http://www.sproutcore.com/ # damn that's cool, click on the carousel
20:02 non left #mojo
20:03 metaperl the whole sproutcore page is greyed out when I visit it
20:03 jfuller joined #mojo
20:03 metaperl ok refresh and it's up
20:04 metaperl sri wher eis the carousel?
20:20 dekimsey left #mojo
20:21 Foxcool joined #mojo
20:22 amoore left #mojo
20:22 xaka left #mojo
20:23 karamorf If I use open to capture STDOUT, how do I close that? guess its more of a Perl question instead of Mojo ...
20:25 metaperl karamorf - these are Perl subs that print to STDOUT?
20:25 metaperl maybe use IO::Stringy so that STDOUT goes to a variable
20:27 karamorf actually it works fine if I use IO::Capture::Stdout ... I was just curious about using open like that
20:27 karamorf oh, to asnwer your question: yes these are just some subs that print to stdout.
20:28 metaperl ok, cool
20:37 perlrocks Twitter: "Imma make a killing off of my Mojolicious Cigarette and Co lmao" --jojomurda http://twitter.com/jojomurd​a/status/68776749894348800
20:59 perlite_ joined #mojo
21:00 Foxcool left #mojo
21:02 perlite left #mojo
21:02 perlite_ is now known as perlite
21:26 Akron left #mojo
21:30 metaperl left #mojo
21:39 tempire lulz
21:39 tempire http://search.cpan.org/~tels/Dev-​Bollocks-0.06/lib/Dev/Bollocks.pm
21:45 kaare left #mojo
22:15 metaperl joined #mojo
22:31 marty left #mojo
22:46 perlrocks Twitter: "Lennart @ @ mart t1mdp Dito here. The difference between perl and perl-MoJo for normal people is huge huh." (nl) --PeterMeus http://twitter.com/PeterMeu​s/status/68809332082483200
22:53 karamorf left #mojo
23:14 alnewkirk left #mojo
23:22 alnewkirk joined #mojo
23:32 Akron joined #mojo
23:37 j3nnn1 joined #mojo

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