Camelia, the Perl 6 bug

IRC log for #mojo, 2011-08-14

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

All times shown according to UTC.

Time Nick Message
00:10 sri tempire: screenflow 3 is out
00:11 tempire Almost forgot about that...got a free upgrade coupon
00:12 tempire some nice new features, too
00:51 Skylos__ joined #mojo
01:22 xaka joined #mojo
02:05 Akron joined #mojo
02:06 Akron sri: Maybe ->pcdata and ->all_pcdata for raw ->text and ->all_text methods in Mojo::DOM?
02:18 sri Akron: better, but pcdata includes nested elements per definition
02:19 sri but we are getting closer :)
02:19 ispy_ joined #mojo
02:37 Akron Where do you have this in spec?
02:42 Akron I refer to the def in http://www.w3.org/TR/xml/#sec-mixed-content where #pcdata refers only to the character data in mixed content, where mixed content can optionally have childs. So - I would say this is pretty close to what these methods should return. Raw would - in my opinion - indicate unresolved entities.
02:43 Akron childs -> children
03:37 kaare joined #mojo
04:05 SmokeMachine joined #mojo
04:19 SmokeMachine joined #mojo
05:03 SmokeMachine_ joined #mojo
06:05 batman joined #mojo
06:05 batman sri: great to hear about <pre> :)
06:06 batman i don't see why <p> support is required... and <br>? i'm lost :)
06:45 sromanov joined #mojo
06:52 sugar joined #mojo
06:57 marcus lost on a sea of pea soup
07:05 marcus tempire: jses flash for sound tho
07:07 batman marcus: https://github.com/jhthorsen​/mojolicious-plugin-urlwith # what do you think?
07:07 batman i think i'm satisfied now... it even got tests ;)
07:08 marcus batman: I like
07:08 batman you say the word, and i'll ship it...
07:08 marcus what's the magic word?
07:08 purl hmmm... the magic word is pizza
07:08 marcus pizza!
07:08 batman it's "ship it" :)
07:08 batman sorry.. that's two words :(
07:09 batman shipit...?
07:09 purl shipit is pretty cool or nice in that space
07:09 marcus muhaha
07:09 marcus shipit
07:09 marcus damnit, now I've been listening for nyan cat for like 10 minutes
07:09 marcus I can feel my mind melting
07:09 batman no good. stop the dance.
07:10 marcus no good. stop the dancer.
07:11 batman a bit risky to take "shipit" from a man who has been listening to nyan cat for 10 minutes...
07:12 batman hope you still like the api in a day or two :P
07:12 batman anyway... it's successfully posted to pause now.
07:12 perlrocks Twitter: "Mojolicious-Plugin-UrlWith 0.01 by JHTHORSEN - http://t.co/mEkcAxL" --cpan_new http://twitter.com/cpan_new​/status/102638783664631808
07:12 batman crap. forgot to add deps to Makefile :(
07:13 * marcus sees a 0.02 in your near future
07:13 batman :D
07:13 batman which version of mojo should i require?
07:14 marcus did you also remember to do a disttest?
07:14 batman 1.00 ? :P
07:14 batman what's that?
07:14 purl rumour has it that is what I'm trying to figure out.  The session that is listening on the port
07:14 marcus that's probably fine
07:14 marcus batman: it's when it makes a dist directory and runs the test from that
07:14 marcus to catch manifest errors and such
07:14 marcus $ make disttest
07:14 batman no, i didn't...
07:15 marcus batman: I always do that before I ship a new version
07:15 marcus of something
07:16 batman i've added requires Mojolicious => 1.00; and Test::More => 0.90;
07:16 batman i think that's it
07:18 batman hm... i can't login to pause :/
07:19 perlrocks Twitter: "Mojolicious-Plugin-UrlWith 0.02 by JHTHORSEN - http://t.co/Dxxn50U" --cpan_new http://twitter.com/cpan_new​/status/102640412786827264
07:20 batman wow. it's already on metacpan :)
07:20 batman that's cool
07:23 batman so now i think i'm only missing a way to add #fragment using link_to() or link_with() ...
07:23 batman but i guess that will be saved to another day
07:49 tempire someone from nowjs said they're about to release python/ruby/php support for nowjs
07:51 tempire I asked about Perl, they said to email steve.
07:51 tempire so I asked Steve.
07:51 tempire go steve
08:05 Foxcool joined #mojo
08:38 exc joined #mojo
08:42 baton8 joined #mojo
10:12 Alias joined #mojo
10:41 marcus tempire: steve says yes?
10:43 sri always do what steve says
10:51 marcus hmm
10:51 marcus I haz metacpan running on my machine. Neat.
11:01 marcus Would be nice if it showed the Changes file a bit more accessible
11:27 Akron joined #mojo
11:36 Akron sri: What definition of pcdata do you refer to?
11:40 preflex_ joined #mojo
11:59 sri Akron: http://en.wikipedia.org/wiki/PCDATA # "Although its name and its appearance in DTD suggest so, #PCDATA itself is not a semantic term for character data;"
12:17 Akron sri: Well - isn't a pragmatic term more useful? ;) In XML spec it's part of the definition of dtds - and there it has the meaning as described above. The Wikipedia-Entry iis funny ... "#PCDATA is the keyword to specify mixed content" ... ehm ... I would say that's wrong.
12:22 Akron That #pcdata can not be part of a sequence or anywhere else in an optional sequence (in DTDs - not elsewhere) has his historical reasons rooted from SGML - with the idea of allowing less complex parser builders for SGML grammar by introducing a sort of determinism.
12:22 Akron I forgot the concrete reason.
12:24 exc joined #mojo
12:37 sri Akron: yea, it's confusing
12:37 sri i can't really find a reason for why wikipedia says that
12:43 Akron I would say, Wikipedia says it, because it's introduced in the mixed content section and they were misinterpreting this as being the keyword for it, which is wrong. The reason, why something like (a|#PCDATA|b) is not allowed in DTDs is a) of historical nature and b) of theoretical nature, as the SGML standard has a strict interpretation of unambigous unit order. There are lots of papers by...
12:43 Akron ...Anne Brüggemann-Klein about that topic. But ... her work makes it even more confusing. I remember reading a part in the Annex of Goldfarbs book where it was pretty clear from a software engineering point of view from that time ... I don't remember the words exactly, but I remember that I understood it then. ;)
13:10 sri next question would be if we actually need ->pcdata and ->all_pcdata anymore :)
13:11 sri if someone would want to work with the actual raw data he would surely just use the dom tree
13:32 preflex joined #mojo
13:33 GitHub149 joined #mojo
13:33 GitHub149 [mojo] kraih pushed 1 new commit to master: http://bit.ly/qqTUtY
13:33 GitHub149 [mojo/master] fixed small preformatted text bug in Mojo::DOM - Sebastian Riedel
13:33 GitHub149 left #mojo
13:36 Akron sri: Iterating through the domtree is not that much fun for this, I would say ... I would vote for both methods. And I wouldd say, text and all_text should use this data and strip whitespace afterwards.
13:38 Akron (Ah - sorry. Stupid idea as you only delete ws in text and not in cdata resp. raw.)
13:38 sri a) stripping whitespace afterwards is really hard b) i don't see the point
13:38 sri right, we also don'tr trim <pre>
13:39 sri actually trimming afterwards for recursive extraction is even impossible
13:41 Akron Yeah - I see. The pre-rule is now in Mojo::DOM - not in Mojo::DOM::HTML?
13:41 sri Mojo::DOM::HTML doesn't do text extraction
13:42 sri extraction and manipulation is all Mojo::DOM
13:42 sri parsing and rebuilding is Mojo::DOM::HTML
13:43 Akron Yes, but I thought all the HTML specific things should go in there ... Why not make the trim parameter of _text reachable from text and all_text, introducing ->text and all_text in Mojo::DOM::HTML with inheriting Mojo::DOM::_text with pre-specific handling?
13:44 sri that sounds like abstraction for the sake of abstraction
13:44 sri i don't see it make the more maintainable or even smaller
13:44 sri quite the contrary
13:45 kaare joined #mojo
13:46 sri hey, if it results in less code to maintain knock yourself out :)
13:46 Akron In that way I could say ->text(0), meaning, I don't want to trim, making ->pcdata not necessary. And all specifics would be in ::HTML. Maybe not smaller, but I would have the functionalities of ->pcdata with nearly no cost and you wouldn't need the pre-rule in the non-specific function.
13:46 Akron :)
13:47 Akron Okay ... I'll give it a try.
13:48 Foxcool joined #mojo
13:48 sri but i'm skeptical
13:49 Akron That's fine. ;)
13:49 sri before adding pcdata i would also like to see a real world use case
13:50 sri just because we can doesn't mean we should ;)
13:50 Akron That's for the testing part of the patch.
14:02 Akron sri: Ah ... okay, I missed the relation between dom and dom::html - putting the ->text with corrrect pre interpretation in there doesn't help, I agree.
14:24 exc is it just me or does perldoc.perl.org have connection issues these days?
14:28 sri http://www.downforeveryoneor​justme.com/perldoc.perl.org
14:28 sri :D
14:28 exc LOL, thanks
14:32 alnewkirk LOL, now they're pushing hosting ads ... wow, nice
14:46 Firkin joined #mojo
15:01 sri Akron: i've thought about that solution too, but as long as we don't have real world use cases there is no point making the code more complicated
15:02 Akron I attached a real world use case from PoCo.
15:03 sri Akron: i don't see the point
15:03 Akron The formatted-string there acts nearly the same as <pre />, although cdata is recommended of course.
15:05 sri Akron: what is the real world use case there?
15:05 Akron So there are environments in real world acting nearly the same way as in html.
15:06 Akron PoCo is real world. There is even an XML schema attribute to allow formatted things to preserve whitespace: http://www.w3schools.com/schema/schema_facets.asp
15:06 Akron Addresses are the common example, but I also know examples from the TEI spec, for example to markup prose.
15:06 sri yes, but where is the real world use here?
15:07 sri this is all purely a theoretical problem
15:07 Akron Maybe I am getting your question wrong?
15:07 sri we don't do a lot of things strictly according to spec, but pragmatic
15:08 sri when i say "real world", i mean stuff people actually do regularly out there
15:08 sri (with Mojo::DOM)
15:09 sri as far as i can see this is so uncommon that using the raw dom tree is perfectly justifiable
15:09 Akron Well ... i already parsed TEI data with Mojo::DOM. ;)
15:09 Akron My point is: not dealing with strict rules is fine as long as it enables fine parsing.
15:10 sri what is fine parsing?
15:10 Akron But it's wrong to disable parsing in some cases and lose information, when this is not wanted.
15:10 * sri doesn't follow
15:11 Akron Fine parsing -> simply works. Even with slightly messed up markup.
15:11 * alnewkirk suspects he just made that up
15:11 sri google doesn't know that term
15:12 Akron No - sorry - alnewkirk is right.
15:12 alnewkirk :} never get tired of hearing that
15:12 Akron :) With fine I meant: It works in a non-strict way as I may expect it.
15:14 sri all information is preserved in the dom tree, if the default behavior of ->text doesn't work for 0.0000000001% of our userbase and they have to use a lower level api i'm fine with that
15:14 Akron I am pretty confused, that you don't think PoCo is a reasl world use case. So XSLT with whitespace-preserving isn't either?
15:16 sri is it common enough to justify making the api ugly?
15:17 sri i mean, how often do you extract formatting sensitive information from xslt files with Mojo::DOM?
15:18 Akron Well ... more often than the other 99,9% of the userbase for sure ... ;)
15:18 Akron Okay - this is maybe more real world: http://www.webdeveloper.com/f​orum/showthread.php?t=204310
15:18 sri i'm a little more reluctant here because supporting this case opens the door for more formatting features like <br> and <p> in a html context
15:19 Akron Funny enough, they engage using xml:space ... which is really not that nice. Although it would be easy to support that as well as pre.
15:19 sri see
15:19 Akron I understand your concerns, although I don't think it's that ugly.
15:19 sri that's what i mean exactly
15:20 Akron Yes, but with text(0) you could handle all these case manually.
15:20 sri maybe Mojo::DOM needs to be refocused on HTML5
15:20 Akron :(
15:20 Akron I'm doing a lot more XML with it than HTML for sure.
15:20 sri xml has so manu things i really don't want to support
15:20 sri *many
15:21 alnewkirk package Mojox::DOM::XSLT; use base 'Mojo::DOM'; ... 1;
15:21 sri this is the discussion i never wanted :/
15:22 Akron I know you hate namespaces. ;)
15:22 sri why did you change the argument order in the patch?
15:23 Akron Well - I would say the opposite: Don't support 'pre' in Mojo::DOM and say, when people want pre to be whitespace preserved, they have to use text(0).
15:23 sri actually, your patch seems broken
15:24 Akron Just to make clear that the first argument is the next that's passed. Broken?
15:24 sri pre can be a child element in all_text
15:24 sri doesn't work
15:24 purl Look buddy, doesn't work is a strong statement. Does it sit on the couch all day? Is it making faces at you? Does it want more money? Is it sleeping with your girlfriend? Please be specific!
15:25 Akron Which percent of the userbase wants an all_text with nested pre? ;)
15:25 Akron sri: Which test doesn't work?
15:25 sri i would assume a huge majority
15:26 Akron Really? What's the real world use case for that? Sound's funny.
15:26 sri every single website using <pre>?
15:26 Akron It passes all tests I had. What is broken?
15:27 sri mojo get mojolicio.us 'pre' all
15:28 Akron No - the usecase for retrieving an all_text with nested preserved whitespaces?
15:28 sri disabling text trimming for xml might be sensible, but then we suddenly have the xhtml problem
15:28 Akron Hm ...
15:28 Akron Ah - okay, right.
15:30 sri mojo get mojolicio.us '#introduction' all
15:30 sri all code examples are broken
15:30 sri without pre handling
15:31 sri by definition changing anything in a <pre> block breaks it
15:32 Akron Okay ... pre may be fine to stay and the disabling of xml for this may be malicious.
15:33 Akron Well - you could enable it for xhtml if the namespace is xhtml but ... I think we both wouldn't like that.
15:33 exc joined #mojo
15:51 GitHub142 joined #mojo
15:51 GitHub142 [mojo] kraih pushed 1 new commit to master: http://bit.ly/oRX1SM
15:51 GitHub142 [mojo/master] improved Mojo::DOM to allow smart whitespace trimming to be disabled - Sebastian Riedel
15:51 GitHub142 left #mojo
15:51 sri Akron: i'm still not sure about this
15:54 Akron sri: I'll try to find other real world examples. For the moment I like the solution. You didn't like my rw example from PoCo?
15:54 Akron Thanks for introducing the functionality! sri++
15:54 Akron (... for the moment.) ;)
15:54 sri just didn't want to add xml into the mix, problem is tricky enough without it already ;p
15:56 sri Akron: also your example didn't test both entries, which seemed a bit odd
15:56 Skylos joined #mojo
15:57 sri although resisting a simpsons example is really hard
15:58 Akron The second entry was the interesting one. I checked the first with one test on the address though, just to check cdata treatment.
15:59 Akron Funny that the PoCo guys had the simpsons address already in their example. ;)
16:00 Akron The name in PoCo can also be formatted - however I wasn't able to think of a good use of ws in names.
16:03 sri Akron: you actually did not test the cdata but some normal field
16:04 sri that's the odd part
16:05 Akron https://github.com/Akron/m​ojo/compare/pcdata#L1R1928
16:05 Akron That's the cdata-raw-treatment-check for the first entry.
16:05 sri oh, missed that, sorry
16:06 Akron :)
16:07 Akron Yeah - mixing css selectors and autoload accessors is only allowed in test suites. ;)
16:09 Akron Hungry ... se you later!
16:09 Akron se -> see
16:11 Skylos I love tests
16:13 GitHub7 joined #mojo
16:13 GitHub7 [mojo] kraih pushed 1 new commit to master: http://bit.ly/rcAvTc
16:13 GitHub7 [mojo/master] added more Mojo::DOM tests - Sebastian Riedel
16:13 GitHub7 left #mojo
16:13 sri Akron: and added :)
16:14 sri the cdata case is not uninteresting
16:14 Akron sri: Yeah! \o/ - Little sidenote: This is the first time you used uppercase usernames in the Change-file!
16:14 sri oh shit
16:15 Akron Well - it's recommended. And it's good because you treat it in building - not in extraction.
16:15 GitHub179 joined #mojo
16:15 GitHub179 [mojo] kraih pushed 1 new commit to master: http://bit.ly/o6q8XI
16:15 GitHub179 [mojo/master] fixed changes ;p - Sebastian Riedel
16:15 GitHub179 left #mojo
16:16 Skylos is that what tis supposed to do... consume newlines when not in a CDATA?
16:16 Akron The mean thing is: 2 of the 3 real world space preserving xhtml elements are already treated as raw when builded (style and script). pre is the only other case, so I have no good arguments against it's special treatment.
16:17 sri script and style are even more special though
16:18 Akron Yepp. Now less code and more food.
16:18 sri still need to come to terms with ->text(0)
16:18 Akron Skylos: It treats the pcdata as xml:space="preserve" would be enabled.
16:19 Akron sri: You will support xml:space tomorrow, right?
16:19 * sri sets Akron on fire
16:19 Akron Damn ... I have to eat ice as an appetizer ....
16:20 Akron Talking about xml:base for a moment ....
16:54 xaka joined #mojo
16:57 ispy_ joined #mojo
17:03 sri hmm, there are more problems with ->text(0)
17:03 Skylos its great that the problems can beillustrated with failing test cases!
17:03 sri it does not work well for Test::Mojo
17:03 Skylos its like an api for bug fixing!
17:03 sri maybe ->pcdata is the better solution
17:04 sri then we could have ->pcdata_is(…) and the like
17:07 yko sri: emm.. what's happened with Mojo::Path::append ?
17:10 sri yko: broken
17:10 sri as in "what the hell, how did this ever get added broken"
17:10 yko -_-
17:11 sri don't say you were using it?
17:11 sri what for?
17:11 purl for fun.
17:11 yko mojo::cpan::uploader was depend on it
17:11 sri :O
17:11 yko well, i have to replace it
17:11 sri it encoded stuff you appended
17:11 sri i was certain nobody could have used it :S
17:12 yko or maybe remove mojox::cpan::uploader at all. it's just a little better than CPAN::Uploader. was a nice toy
17:13 yko well, it was used only in tests
17:14 sri well, you can bring back ->append as a working version if you like
17:14 yko actually tests are the thing that being broken most often
17:14 yko via updates i mean
17:15 sri well, when i removed it absolutely nothing in mojolicious broke, it didn't have a single test...
17:19 sri looking through git, it was there since the first commit, must have been simply forgotten
17:20 sri poor method, never used for anything
17:20 sri broken from the start
17:20 j3nnn1 joined #mojo
17:24 sri but yea, if you think you can find a use for it… just bring it back with tests
17:24 yko nevermind, i already removed it
17:25 yko just was thinkig maybe it was removed by occasion
17:26 sri i was so sure nobody could be using it, sorry :(
17:34 abra joined #mojo
17:34 abra joined #mojo
17:48 sri interesting, Akron is wrong
17:48 sri pcdata does not only refer to text content
17:57 ispy_ joined #mojo
18:03 Akron joined #mojo
18:03 sri Akron: i've found quite a few more refereces for pcdata representing more than just text content
18:04 sri even the xml spec refers to it as a keyword for mixed content
18:04 Akron sri: Hi again. Examples?
18:05 sri i've just been googling for parsed character data
18:07 Akron Where does the spec refer to "PCDATA as a keyword for mixed content"? And where are elements allowed to be nested in mixed data with only allowed #pcdata?
18:07 perlrocks Twitter: "Perl programmers must be psychic.I recently thought about this functionality and it's already up on #CPAN. http://t.co/dFkO08P #perl #mojo" --arpadszasz http://twitter.com/arpadszas​z/status/102803661457063936
18:08 sri http://www.w3.org/TR/xml/#sec-mixed-content
18:09 Akron I can't see it. It's like what I said here: http://irclog.perlgeek.de/​mojo/2011-08-14#i_4274430
18:10 sri there are many contradicting explanations on the web, in this case i think i trust wikipedia
18:10 Akron Mixed content can contain nested elements and has - by definition - #pcdata as part of its content model.
18:10 sri and i've just looked through all the specs, nowhere is it actually really explained
18:11 Akron I can give you a quote from Goldfarb as soon as I can get the boook.
18:13 Akron sri: But where are references where #pcdata can contain nested elements?
18:13 sri wikipedia
18:13 purl wikipedia is actually useful or the slashdot's book report or "mostly harmless" or UNEMPLOYED PH.D. DEATHMATCH!!!! or often authoritative or often inaccurate or the "faith-based" encyclopedia or full of very sill^Wuseful information or http://www.penny-arcade.com/comic/2005/12/16 or http://xrl.us/bfy9mj or http://xkcd.com/c214.html or http://xrl.us/bfy9m8 or held together with string, baling wire and stupid people
18:13 sri the xml spec does not explain it either
18:14 sri the point is, there is no explanation!
18:14 Akron Which lemma in Wikipedia?
18:14 sri http://en.wikipedia.org/wiki/PCDATA
18:14 sri quite sure i linked it before
18:14 sri "Parsed Character Data (PCDATA) is a data definition that originated in Standard Generalized Markup Language (SGML), and is used also in Extensible Markup Language (XML) Document Type Definition (DTD) to designate mixed content XML elements"
18:16 Akron Yes - but I said the wording is wrong and this wording is nowhere in the XML spec. The point is, that there is no definition of pure text content in the spec, only pure element content and mixed content. Pure text content is part of mixed content, where no elements are defined as children.
18:17 sri well, i have given you a reputable source, now it's your turn
18:17 sri if in doubt wikipedia wins
18:17 Akron I'll check Goldfarb.
18:17 sri how about you try and fix the wikipedia article and we see what happens? :)
18:18 Akron I am not part of Wikipedia and happy with it.
18:20 Akron Oh - Goldfarb is in the library of our university. I can fetch it today.
18:21 sri i expect a photo of the paragraph! :)
18:23 Akron Hmpf. You are sooo distrustful Mr supervillain. ;)
18:23 sri it's my job to be! ;p
18:23 alnewkirk sri, can you design a t-shirt with the Mojolicious cloud and rainbox, .. i'd like to buy some
18:24 sri cloud doesn't look so nice on t-shirts
18:24 sri would need a working concept first
18:25 * Akron is on his way to the library.
18:26 sri been thinking about a cloud skull or so… but didn't have much time for design recently
18:28 baton8 joined #mojo
18:55 Firkin left #mojo
19:03 tabbi joined #mojo
19:03 tabbi left #mojo
19:05 mspo huh, I just figured out that I wasn't using my controller at all ;)
19:05 mspo maybe it's a sign that I should migrate some logic out of the helper
19:10 Akron Goldfarb is even more confusing on that topic ... the term is really vagua. Problem is, #pcdata is a parsing instruction as well as part of the content model. As an instruction it of course allows markup, but as a token it doesn't.
19:12 Akron The quote that supports my suggestion for ->pcdata the most is Goldfarb (1990, 412, §30): "All data characters occurring between successive tags are considered to satisfy a single "'PCDATA" token, even if some were declared to be character data by marked section declaration."
19:14 Akron sri: I can send you copies of the pages this afternoon.
19:15 sri Akron: not necessary unless you find something more specific
19:16 Akron sri: In SGML the correct term of what we want is RCDATA, which is a processing instruction allowing only for character data and resolved entities. But this wasn't introduced to XML as it wasn't necessary - PCDATA only was needed for SGML as there are some cases where characters where both markup and character data.
19:19 Akron sri: That's not the case with XML so I would say #pcdata tokens can only contain character data and resolved entities. However - it's absolutely clear that it can't contain nested elements and I'm curious about this interpretation.
19:20 Akron sri: There is a simple explanation online I quite like: http://www.stylusstudio.com/​xmldev/199907/post30310.html
19:31 JonChisciotte joined #mojo
19:32 sri Akron: ""parsed character data (#PCDATA)" - Parsed character data is text
19:32 sri > that will be examined by the parser for entities and markup."?
19:33 sri well, if there's something you've convinced me of it's to never refer to anything as pcdata :D
19:34 Akron sri: Yes. That's the instruction term - not the token term. This data has to be parsed. That's what the instruction says.
19:34 JonChisciotte Hi, I have a mojolicious app and I need to fork a process to execute an independent perl script... I'm using perl's fork or Forks::Super but they always create zombies when processes end. I'm using morbo to serve the route that do the fork. Could you give me suggestion to what doc to read, or other techniques to make what I need? Thank you
19:34 Akron sri: Okay - this maybe the best interpretation. :)
19:35 sri JonChisciotte: my suggestion is "don't"
19:35 sri forking from a web app is a recipe for disaster
19:36 JonChisciotte sri: ok, but I need to start a long process task by the interface... how to do it?
19:36 sri external job queue and worker
19:36 JonChisciotte sri: should I create another app listening to another port and send commands async?
19:37 stephan48 ye
19:37 JonChisciotte sri: I took a look to TheSchwartz but it seems it requires 6% of cpu even if it is doing nothing...
19:45 jms joined #mojo
19:53 sri yko: push $up->url->path->parts, 'subdir';
19:53 Akron sri: Maybe not as good as Goldfarb but in my opinion better than Wikipedia: St. Laurent and Fitzgerald (2005, 3rd edition of "XML Pocket reference", p. 35): "The simplest kinds have content models for parsed character data (#PCDATA), which means that these elements must contain only text -- no element children.""
19:53 sri that's what i would do
19:54 sri (yes, i use 5.14 features!)
19:54 sri :)
19:56 sri parts is an array of unescaped strings, perfect for manipulation
19:59 jms hi there
19:59 purl hey, jms.
19:59 * sri waves
20:00 jms I have few questions about mojolicious
20:00 jms the server, morbo
20:00 jms is it scalable?
20:01 jms meaning, can be used in a cluster?
20:01 sri morbo is a development server
20:01 jms is not considered as potential web server
20:01 jms ?
20:01 sri while it is rather speedy, it is not for production use
20:02 jms ok
20:02 jms and there's any plan to get to production?
20:03 jms because could be interesting to add spdy functionality as google does
20:03 jms it can boost up mojo
20:06 sri hypnotoad is our default production server
20:07 jms ok
20:07 JonChisciotte all the glory to the hypnotoad!
20:07 GitHub51 joined #mojo
20:07 GitHub51 [mojo] kraih pushed 1 new commit to master: http://bit.ly/qSBPvg
20:07 GitHub51 [mojo/master] deprecated Mojo::Path->append in favor of using Mojo::Path->parts directly - Sebastian Riedel
20:07 GitHub51 left #mojo
20:07 ispy_ joined #mojo
20:08 sri yko: ok, it's properly deprecated now
20:08 yko thanks much
20:08 sri the old broken behavior makes it tricky
20:08 sri but this is better than nothing i guess
20:09 jms I'll have a look on it, thanks for the info
20:10 jms btw: cook that you use futurama names, love the tv series
20:10 Kovensky joined #mojo
20:14 sri yko: will recheck and release later
20:33 exc joined #mojo
20:40 GitHub73 joined #mojo
20:40 GitHub73 [mojo] kraih pushed 1 new commit to master: http://bit.ly/qEQFE7
20:40 GitHub73 [mojo/master] trim arguments are experimental - Sebastian Riedel
20:40 GitHub73 left #mojo
20:43 sugar_ joined #mojo
20:49 perlrocks Twitter: "Mojolicious 1.77 by SRI - http://t.co/IXLEZtM" (ms) --cpan_new http://twitter.com/cpan_new​/status/102844392599060483
20:54 vel joined #mojo
21:07 vel joined #mojo
21:17 JonChisciotte excuse me, following the synopsis of Mojo::IOLoop I'm listening to a  port and connecting to it with the browser I can see it works. I'm trying to connect with $loop->connect but I cannot see any results, http://pastebin.com/ehUezd8Z What I'm doing wrong?
21:19 sri JonChisciotte: it helps if you start the loop ;)
21:22 JonChisciotte sri: uh... but maybe I'm starting to understand...
21:24 trone joined #mojo
21:24 JonChisciotte for each connect I make a should use an on_connect to send data and then stop connection
21:25 JonChisciotte on the other side I'll read the chunk and I'll take over the world depending the value I've sent.. It seems a good and easy alternative to fork
21:42 alnewkirk joined #mojo
21:44 alnewkirk sri, what about a unicorn with the cloud and rainbow branded on the side?
21:49 sri alnewkirk: poking a hole into the catalyst logo?
21:50 alnewkirk Lol, sure, why not
21:51 alnewkirk sri, a raptor eating the catalyst logo would be better but I dont know the extent of your design skills
21:56 sri design skills are not the problem, there are a lot of restrictions on high quality t-shirt prints
21:58 sri you can forget about all those fancy images you have in your head
22:06 alnewkirk true, but I've seen some pretty fancy vector art
22:37 sri i wouldn't hold my breath though, ever since the raptor experiment i'm not very motivated to design more t-shirts
22:46 alnewkirk sri, what do you mean ... raptor experiment? what happened to phenona?
22:50 sri my raptor shirts don't sell, there are only a handful out there
22:51 sri at least not via spreadshirt
22:51 sri you don't know what happened to phenona? seriously?
22:52 sri then you deserve having to google for it!!! ;p
23:05 alnewkirk sri, rude .... lol jk
23:05 alnewkirk i heard a 13 sold a paas to activestate but i didnt know that was you
23:06 alnewkirk I'm going to buy a raptor shirt tomorrow
23:07 sri i have nothing to do with phenona
23:08 alnewkirk oh, sorry
23:13 Akron joined #mojo
23:44 Kulag joined #mojo

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