The web in a box - a next generation web framework for the Perl programming language

IRC log for #mojo, 2015-04-22

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

All times shown according to UTC.

Time Nick Message
00:00 bowtie joined #mojo
00:03 basic6_ joined #mojo
00:12 bowtie joined #mojo
00:36 hshong joined #mojo
00:50 bowtie joined #mojo
00:54 punter joined #mojo
01:03 bowtie joined #mojo
01:05 asarch joined #mojo
01:26 bowtie joined #mojo
01:29 klapperl joined #mojo
01:50 inokenty-w joined #mojo
02:04 noganex joined #mojo
02:06 bowtie joined #mojo
02:15 bowtie joined #mojo
02:44 Zoffix ,
02:45 jberger ;
02:45 Zoffix :)
02:46 Zoffix The comma marks the time when I pushed a module that uses Mojo::DOM to CPAN :) I'm testing your new bot :)
02:46 Grinnz heh
02:47 Zoffix GumbyPAN got it :P
02:47 bowtie joined #mojo
02:48 Grinnz that module sounds like spyware, just sayin.
02:48 Grinnz i know the name comes from the site :P
02:49 Zoffix haha
02:50 jberger actually I just pushed a module to CPAN as well, my ; isn't far removed from it
02:58 bowtie joined #mojo
03:14 bowtie joined #mojo
03:24 freshmojo joined #mojo
03:25 basic6 joined #mojo
03:25 bowtie joined #mojo
03:34 cpan_mojo Mojo-JWT-0.02 by JBERGER https://metacpan.org/release/JBERGER/Mojo-JWT-0.02
03:35 bowtie joined #mojo
03:45 bowtie joined #mojo
03:49 Snelius it's may be interested http://cpanmetadb.plackperl.org/versions/
03:52 Zoffix A lovely colorblind-unfriendly chart \o/
03:52 kaare joined #mojo
03:52 Zoffix Ah, nm, it shows more info when you hover slices.
03:59 bowtie joined #mojo
04:04 cereal_ joined #mojo
04:09 bowtie joined #mojo
04:33 Grinnz 5.10.1 as the biggest chunk... unfortunately not surprising
04:34 bowtie joined #mojo
04:44 bowtie joined #mojo
04:58 bowtie joined #mojo
05:08 bowtie joined #mojo
05:08 * batman doesn't understand the fuzz about the cpan_mojo delay
05:10 Zoffix Fastar! Bettar!
05:10 preaction Hardar! Strongar!
05:18 bowtie joined #mojo
05:29 bowtie joined #mojo
05:31 punter joined #mojo
05:40 bowtie joined #mojo
05:46 freshmojo joined #mojo
05:57 davido_ joined #mojo
06:10 bowtie joined #mojo
06:14 kaare joined #mojo
06:16 kaare_ joined #mojo
06:21 dod joined #mojo
06:25 dod joined #mojo
06:26 bramirez joined #mojo
06:29 bowtie joined #mojo
06:33 bramirez joined #mojo
06:41 bowtie joined #mojo
07:04 bowtie joined #mojo
07:05 gatitskiy joined #mojo
07:09 trone joined #mojo
07:11 eseyman joined #mojo
07:26 bowtie joined #mojo
07:26 dod joined #mojo
07:35 cpan_mojo Mojo-SMTP-Client-0.10 by OLEG https://metacpan.org/release/OLEG/Mojo-SMTP-Client-0.10
07:35 cpan_mojo Mojolicious-Command-migration-0.1 by LIKHATSKI https://metacpan.org/release/LIKHATSKI/Mojolicious-Command-migration-0.1
07:44 bowtie joined #mojo
07:50 bramirez joined #mojo
08:01 ryozi joined #mojo
08:01 bowtie joined #mojo
08:06 bramirez joined #mojo
08:20 Vandal joined #mojo
08:29 fhelmber_ joined #mojo
08:33 jbob joined #mojo
08:34 oetiker joined #mojo
08:46 bowtie joined #mojo
08:56 bowtie joined #mojo
09:04 memowe joined #mojo
09:05 punter joined #mojo
09:11 likhatskiy joined #mojo
09:12 bowtie joined #mojo
09:24 bowtie joined #mojo
09:25 riche what is the Mojolicious minimum perl requirement? I am trying to figure my limits on unicode handling
09:35 bowtie joined #mojo
09:39 Lee riche: 5.10.1
09:40 riche thanks, Lee.
09:45 bowtie joined #mojo
09:54 bowtie joined #mojo
10:07 bowtie joined #mojo
10:13 jnbek joined #mojo
10:16 bowtie joined #mojo
10:26 bowtie joined #mojo
10:31 riche sri:  Grinnz mentioned Mojo::UserAgent::Transactor yesterday, now I think just ln 173-180 needs re-evaluation in light of what I'm going through in the W3 spec
10:31 riche so what I did yesterday really was BS
10:32 marcus https://aphyr.com/posts/322-call-me-maybe-mongodb-stale-reads at this point I am puzzled as to why anyone would choose to work with mongodb
10:34 riche marcus: I have been saying that for a long time, but I guess my abstract communications about it just aren't effective
10:36 bowtie joined #mojo
10:37 riche sri: actually to be more clear, 177-180.  But the charset setting also requires re-evaluation based on Rule 2 (form vs. document encoding)
10:44 chansen joined #mojo
10:46 bowtie joined #mojo
10:57 bowtie joined #mojo
11:07 bowtie joined #mojo
11:09 mattastrophe joined #mojo
11:13 kaare_ joined #mojo
11:20 bowtie joined #mojo
11:30 bowtie joined #mojo
11:31 bowtie joined #mojo
11:33 bowtie joined #mojo
11:45 neilhwatson joined #mojo
11:46 bowtie joined #mojo
11:58 mattastrophe joined #mojo
11:59 riche joined #mojo
12:01 bd marcus: also read the other "call me maybe"s
12:04 bd jberger: heh, i think i use the elevator mp3 as my ringtone
12:16 marcus bd: I've read a lot of them.
12:17 marcus jepsen seems pretty cool.
12:17 jberger riche: just remember that more recent perls have had major Unicode improvements
12:18 riche of course
12:18 riche this is why I needed to know the threshold
12:19 riche you know even browsers can't get their flippin encodings straight
12:21 jberger batman: it's really just by comparison with the old feed, which was remarkably fast
12:22 jberger it's not actually important
12:22 batman my point is: why do anyone care?
12:22 batman the time is only relative to the person who should care the least (the person releasing it)
12:22 jberger people are proud of their releases
12:22 batman for everyone else it happens exactly when it does. not late. not early.
12:23 jberger right it doesn't actually matter
12:23 batman s/actually// :)
12:26 * batman have to care less
12:26 bramirez joined #mojo
12:31 guru joined #mojo
12:32 marcus caring is creepy
12:33 marcus https://www.youtube.com/watch?v=PUMkxH03V5c
12:35 bd there was a talk by the soureforge guys, where the "download latest version" button was very important to the person releaseing. and it took for them too long, so lots of people wrote to sf.net your "i mada a new release" button is not working.
12:54 riche wtf-ville:  https://gist.github.com/rpcme/d43643896f28f0ab1d9f
12:57 riche anything that is 2+ browser in that list I am going to make doc enc / form enc test for this form url encoding issue
12:58 riche yes, ep included in ep matrix fun
13:01 jberger every time I see IANA I of course think "I am not a", so "I am not a MIME Name" is how I read that :-)
13:02 riche after a few beers it could be synonymous
13:03 riche i wouldn't have to do the ep's if "we didn't have to see" what these browsers really did but.. i guess we do
13:04 riche but I already know the current impl is screwed for form encoding rules, it's just a matter of how much it has to change
13:05 riche and form=>{"foo"=>"bar","bar"=>"baz"} isn't enough, because W3 encoding rules take into account element types
13:06 jberger another mongo laffer is out: https://aphyr.com/posts/322-call-me-maybe-mongodb-stale-reads
13:07 riche jberger: marcus mentioned earlier
13:07 mattastrophe joined #mojo
13:07 jberger oh, sorry, I missed that
13:07 jberger marcus++
13:07 riche first post, irc style
13:08 jberger there were too many other missed items that caught my eye, I didn't get back to that one
13:16 jberger WEBSCALE!
13:16 purl webscale is an anagram of claw bees
13:25 * jberger is liking the new gmail inbox
13:25 sri ouch, looks like the wiredtiger engine is pretty much unusable atm
13:26 jberger I have three inbox invites, does anyone want one?
13:32 amon joined #mojo
13:32 mattastrophe joined #mojo
13:40 genio jberger: really?  You like it so far?  It makes everything so large it's hard to see anything
13:41 genio I've continued to use regular gmail because it's too fluffy
13:44 genio and it doesn't use the entire width of the screen
13:51 jberger maybe I wouldn't use it for work, but for home use it seems to be handy
13:52 genio I guess I should reserve further judgement until I actually try to get past the UI changes and see if it's more useful to me than the regular gmail app
13:53 * genio might be being a stick in the mud solely due to a UI I'm not immediately fond of
13:55 skingsbu joined #mojo
13:55 gryphon joined #mojo
14:04 memowe joined #mojo
14:05 Insane joined #mojo
14:05 Insane hi all, what if I have a Mojo app, which requires some perl modules to be installed from cpan. What is the best and easiest way to automatically install those modules on application first start?
14:08 Grinnz automatically installing modules is usually a bad idea, but if it's not a CPAN project i tend to include a cpanfile so i can just run cpanm --installdeps .
14:09 Insane it is not a CPAN project, it is internal proect with dployment from local git server, and there are no cpanm on out servers
14:09 Insane only default CPAN
14:10 Grinnz default CPAN doesn't know cpanfiles, sorry
14:10 Onigiri genio: it's rather meh compared to the other one.
14:11 gabiruh joined #mojo
14:11 Grinnz you could make a script that runs cpan -i Some::Module Some::Other::Module etc
14:11 Grinnz but like automatic installation, that will probably install them soemwhere that's not useful
14:14 jberger make a Build.PL or Makefile.PL and just use it directly
14:16 gabiruh joined #mojo
14:16 Grinnz yeah essentially the best way to "automatically install dependencies" is to make it a CPAN-like project
14:16 ribasushi Insane: note that it can very well be Your::InHouse::App::Dependencies
14:16 Grinnz because then you can use a CPAN installer, and it already knows how to install them
14:16 ribasushi which has nothing in lib/
14:17 ribasushi which will make `cpan .` just work
14:17 ribasushi (and so would `cpanm .` or `cpanm . --installdeps`)
14:18 * tempire hugs sri
14:18 gabiruh joined #mojo
14:33 tomboh joined #mojo
14:35 sri \o/
14:36 sri this is a pretty damning issue for mongodb https://jira.mongodb.org/browse/SERVER-17975
14:50 disputin joined #mojo
14:53 sri little sad to read everywhere that only foundationdb got distributed databases right
14:54 ribasushi sri: because getting it right requires an investment pretty much nobody is willing to make
14:54 Zoffix I wonder if Hari Seldon made it...
14:58 asarch joined #mojo
14:59 ribasushi https://www.youtube.com/watch?v=4fFDFbi3toc
14:59 ribasushi sri: ^^ almost the entire talk is about what it took them to get there, highly recommended
14:59 sri hahaha... so the oracle eula actually forbids publishing research like the jepsen tests
15:00 sri ribasushi: and now apple is hiding it from the world :(
15:00 ribasushi sri: because it is *that* good
15:01 ribasushi sri: but given the current "research culture" nobody will invest into something like this without a guarantee of it being closed-sourced for good
15:02 Grinnz_ that's business interests for you :)
15:03 Grinnz_ screw advancing humanity, as long as they can advance their profits
15:15 jberger Grinnz_: interesting that that was the original argument for patents
15:15 jberger encourage disclosure in exchange for several years of protection
15:15 Grinnz_ yeah... patent and copyright law had good intentions, but they are abused far too often
15:23 sri hmm, rails 5 is getting websocket support... and it seems like a total mess so far
15:23 sri http://railsbird.tumblr.com/post/117076597115/existential-crisis-at-railsconf
15:24 Grinnz_ funny how designing from the ground up for this stuff works out so much better
15:25 sri from what i hear, this one got a lot of rails folks scared http://www.phoenixframework.org/
15:27 sri not too exciting for mojolicious users ;p
15:27 sri but maybe there's some stuff we can steal
15:28 Grinnz_ i keep hearing about that elixir stuff
15:28 sri http://www.phoenixframework.org/v0.11.0/docs/channels
15:54 PotatoGim^Home joined #mojo
16:01 jberger ribasushi: that talk is awesome
16:04 stryx` joined #mojo
16:08 sri rails makes websockets really complicated :o
16:09 riche i think maybe too tired ... what is wrong with this? https://gist.github.com/rpcme/2e621ee64202da7b587c
16:09 riche i can't get append to work it is so weird
16:13 sri you want append_content
16:13 riche but I am using append -- and basically doing the same as the example in the doc
16:13 riche maybe I should want that ... but nothing in Mojo::DOM explains why?
16:14 sri how can the root node have siblings?
16:14 sri it's a logical error
16:17 mattastrophe joined #mojo
16:20 riche i think the first line in the append doc should go then ==> $dom = $dom->append('<p>I ♥ Mojolicious!</p>');
16:20 riche it's probably a copy paste error from append_content
16:21 riche i got it working with append finally ... just confusing
16:21 sri ?
16:22 sri you make no sense
16:24 mattastrophe joined #mojo
16:24 riche what I mean is remove that line from the append method documentation in Mojo::DOM
16:24 riche because it does-not-work
16:25 sri of course it does
16:25 riche whatever
16:25 purl whatever is, like, left over I just deal with myself, or something
16:25 sri i was about to mention how easy it is to make work the way you expect https://gist.github.com/anonymous/09ab77c59b39e5d2ecf7
16:26 Oleg joined #mojo
16:26 riche it is ok
16:26 riche i already use ->at and just created the dom obj with a div
16:26 riche but if you don'
16:26 sri but this misconception makes me sceptical if too much dwim would be harmful here
16:26 riche t do that that line will never work
16:26 riche not never work
16:26 riche just not work without knowing what should go before it
16:27 sri like i said, it works perfectly, and you've just got a logic error there
16:27 sri because a root node can of course not have siblings
16:28 riche yes so you have to do ->at ... fine I get that now
16:31 jberger what ruby/rails really needs now is for someone to throw a coffee cup
16:31 jberger in 15 years they can have ... a promise of a replacement
16:32 sri you'd think if they do their own js library (really coffeescript), they'd also add a poll fallback... but no, they get the worst of both worlds
16:33 kaare joined #mojo
16:34 jberger that existential crisis post reads a lot like I imagine the pre-2000 perl world felt
16:35 sri i think there has been a crisis ever since node.js
16:36 jberger again, I really think that all programming languages follow a popularity curve of some initial ramp-up (a few years-ish) a boom and eventual drop with a tail
16:36 jberger and the only difference between languages is how high is the crest and how long is the tail (before it hits zero)
16:36 jberger the more important one perhaps is, how long is the tail
16:36 jberger perl seems to have survived and will prove to have a long tail
16:37 jberger will ruby/rails? I'm curious
16:37 jberger I'm also curious about node, though javascript isn't going anywhere
16:37 sri except for javascript, rules don't apply there
16:37 jberger javascript gets a weird pass because of browsers, there is no choice
16:38 sri i guess sql too
16:38 jberger that wasn't obvious a few years ago
16:38 jberger now I think you might be right
16:38 jberger SQL's tail will be very long indeed
16:38 lb jberger: no language that ever entered production hits zero
16:39 jberger lb, no, but some do eventually become as obsolete as can be considered irrelevant, at least as far as new projects are concerned (which is a reasonable metric)
16:39 mattastrophe joined #mojo
16:39 jberger very little new work being done in fortran these days
16:39 lb that's very non-mainframe-centric ;)
16:40 jberger there are some old fortran systems that will live forever
16:40 jberger but very few new projects
16:40 lb and 4gl
16:40 riche I still have to do COBOL and JES jobs
16:40 jberger riche: yes, but mostly maintenance right?
16:41 lb COBOL will never ever ever go away
16:41 lb no, plenty of new projects in cobol
16:41 jberger and even if not maintentance, still for the same companies with existing systems
16:41 riche but my favorite obscure language is SNOBOL
16:41 riche my last program in SNOBOL was about 10 years ago, however.
16:42 riche jberger: no way... there is new COBOL everyday
16:42 * lb worked on a project converting CA-Ideal into EGL
16:42 jberger interesting
16:42 riche i really hate CA products for zOS
16:42 riche klunky, over-engineered
16:42 lb EGL is a crazy-bad IBM language compiling into java or cobol
16:42 riche built for large consulting teams
16:43 lb CAs products are probably mostly something they bought and can start over-charging for
16:43 riche many ppl are getting off it
16:43 lb CA is more evil than oracle and microsoft combined
16:43 riche actually I am pretty certain you can generate .NET as well
16:43 jberger well perhaps COBOL is like javascript, platforms require it (mainframes vs browsers)
16:44 mst the last cobol shop I was aware of figured out how to get DBIx::Class to talk to their ISAM db and switched
16:44 riche many ppl are just generating the final .NET and dropping CA, they are so fricken expensing
16:44 * lb still wonders which of Oracle and CA will buy microsoft for its legacy customers
16:44 jberger yikes, continuing a project from generated source code, sounds awful!
16:45 jberger lb: wow, what a prospect
16:45 purl somebody said a prospect was their main office.  I was in a tiny sattelite office in downtown Philadelphia
16:45 jberger purl: forget a prospect
16:45 purl jberger: I forgot prospect
16:45 lb jberger: it was not a kind idea when i got it, microsoft has improved since :)
16:45 riche jberger: yeah, spend some time refactoring and save about $3M over two years... not too shabby
16:47 riche mst: there are a LOT of mainframe shops still. and yes they still write cobol, and cobol programmers make a premium
16:48 mst riche: well, duh, that's just a datapoint based on people -I- know
16:48 riche :)
16:48 lb most of them have no imagination, and never program at home
16:48 mst I'm not sure 'having imagination' and 'writing COBOL for eight hours a day' are really compatible
16:48 lb that's just a datapoint based on people -I- know, of course
16:48 lb and code i've read :(
16:49 lb let's just say they like copy paste
16:49 riche well, i wouldn't be that harsh, lb, but yeah many of them are career programmers.  not software designers.  there's a difference
16:50 lb speaking of, i'm reading php code this week
16:50 riche fuck that
16:50 riche I would rather do voluntary, hourly, colonoscopy
16:51 lb that's cheating, you can grow to like most anything
16:51 riche okay this is getting distracting.  need to go finish this thing.
16:51 lb :)
16:52 riche srt: i created a mojo app to autogenerate all the forms for doc encoding and form encoding.  now i can finally see what things are doing
16:52 riche so there won't be any question
16:55 punter joined #mojo
17:07 sri marcus, tempire, batman, crab: would be nice if you could participate more in github issues
17:08 tempire will do
17:08 jberger riche++
17:11 riche yeah I was going to do it manually until i saw there were 1,875 different ways it could be configured, and that's not taking into account posting different strings
17:12 riche oh and that's only for encodings that are supported in at least 2 of 3 major browsers
17:12 mst on mongodb: https://aphyr.com/data/posts/322/docs-consistency.jpg
17:20 sri riche: sounds like you might be going a little too far now :)
17:22 sri special handling for charsets other than UTF-8 is out of scope imo
17:25 riche korean document encoding with japanese form encoding wheee
17:26 lb mst: what's wrong with the warm fuzzy feeling that you might have stored something somewhere
17:26 jberger lb: hahahahaha
17:26 purl LOLCON 5 reached.
17:26 mst lb++
17:33 riche well I can say for sure this works well:  app->types->type($docenc => "text/html; charset=" . lc $docenc);
17:33 riche thanks for having that in the doc, wouldn't have figured it out
17:35 mattastrophe joined #mojo
17:55 berov joined #mojo
18:00 Oleg is it possible to disable using of Mojo::Asset::Memory for uploads and always use Mojo::Asset::File?
18:04 Oleg MOJO_MAX_MEMORY_SIZE=0
18:07 Oleg oh, this works: $ENV{MOJO_MAX_MEMORY_SIZE} = -1
18:09 mattastrophe joined #mojo
18:12 riche sri: it seems like mojo does not honor form data encoding?
18:12 riche meaning the accept-charset attribute?
18:13 asarch Is there any way to get the list of all $r->get in the app (in order to build a dynamic CSS menu)?
18:23 sri Oleg: why?
18:24 Grinnz_ if you always need a filehandle for something, perhaps
18:24 Grinnz_ (i have a workaround that does this for a virus scan that takes a filehandle...)
18:24 Oleg sri: I already reconsidered to do this :D
18:24 guru joined #mojo
18:25 sri it auto-upgrades from memory to file if the upload is big enough
18:25 Grinnz_ well, not "does this", my workaround is to create a filehandle to a scalar containing the contents if it's a Memory asset
18:25 sri so, there really shouldn't be anything to worry about
18:26 Grinnz_ i dont suppose there's a way to force it to upgrade to a file?
18:26 Grinnz_ that would be easier...
18:27 sri hmm
18:27 punter joined #mojo
18:27 Grinnz_ asarch: https://metacpan.org/source/SRI/Mojolicious-6.08/lib/Mojolicious/templates/development.html.ep#L231 is how the exception page does it... unfortunately the logic is template embedded
18:28 asarch Thank you Grinnz_
18:28 asarch Thank you very much :-)
18:30 Oleg In my case I didn't want to use move_to to get the file, because I need to unzip it and don't need the file itself. So, wanted to unzip $upload->asset->path directly
18:30 Oleg But now I am ok with move_to()
18:31 sri Grinnz_: hardly seems worth it https://gist.github.com/anonymous/ff2ae6d5c4928025581a
18:32 Grinnz_ sri: i could just do that manually too i guess, but it does make the functionality clearer imo
18:32 batman sri: Sorry, but I have now idea what to say about the form encoding.
18:33 sri batman: well, if you don't understand the description, please say so
18:33 sri it's what i said too actually
18:33 batman I kind of agree that something that works so well should not be changed, but I also don't mind if it gets changed
18:33 Grinnz_ sri: possible alternative: ability to get a handle to scalar contents of a memory asset?
18:34 Grinnz_ i seem to remember that not working on all perls, but i dont know if that was pre-5.10.1 or not
18:34 sri Grinnz_: does not seem very elegant
18:34 batman sri: I think I understand, but I have no idea which option is the best :(
18:34 Grinnz_ sri: perhaps...
18:34 purl then again, perhaps not
18:35 sri batman: then you are neutral on the issue, say so!
18:35 Zoffix joined #mojo
18:36 sri well, actually you're worried about making a mistake, so i'd count that as slightly -1
18:36 sri at least until it has been proven that it is the right thing to do
18:37 batman sri: Ok. I will when I get home.
18:37 * sri will state his requirements more publically too then
18:39 batman Requirements? Regarding commenting?
18:41 * Grinnz_ is chasing a very peculiar issue which is probably mysql's fault
18:42 Grinnz_ (as happens basically every week, naturally)
18:43 sri batman: https://github.com/kraih/mojo/pull/786#issuecomment-95297262
18:44 Kripton joined #mojo
18:45 batman sri: I'll still comment when I get home. It kind of sucks that there's no right and wrong here :(
18:45 batman At least that's how I understand it.
18:45 riche i closed the pull request
18:46 riche 1) what i started doing was wrong
18:46 riche 2) it seems I can't ever communicate anything right here, so I am just going to hack through it on my own
18:46 riche i already lost more than a day on this stupid issue
18:47 sri and putting some more pressure on the rest of the core team to vote https://github.com/kraih/mojo/pull/785#issuecomment-95298765
18:47 sri no vote means you agree with the majority
18:48 sri riche: that's unfortunate
18:48 Grinnz_ sri: i've seen that too (the charset from the REST APIs i've used)
18:49 riche i just give up, i have a funding meeting next week and I must get this poc done
18:49 sri also wonder if Mojo::UserAgent should be sending application/json;charset=UTF-8
18:49 Grinnz_ you mean for the json generator?
18:49 sri yes
18:49 Grinnz_ hmm, i know i'm using that for something
18:50 sri riche: i was actually hoping more people would get involved if i post a preliminary vote...
18:50 sri closing it right away kinda sucks
18:50 riche the code i changed was dead wrong anyways
18:50 sri oh
18:51 Grinnz_ seemed correct to me
18:51 riche there are other conditions to take into account, including acknowledging document and form encoding
18:51 sri yea, tests seemed to indicate you were right
18:51 Grinnz_ that's a whole other scope really
18:51 Grinnz_ nothing to do with URL encoding
18:51 sri indeed
18:52 riche the problem I have is not url encoding!  it is form data which is treated as parameter
18:52 riche the payload with the colon is form data
18:52 riche that is why i said earlier today the lines that were the culprit
18:53 Grinnz_ form data is parameters
18:53 riche but the encoding rules are different for form data
18:53 riche non-get
18:53 riche non-query-strings
18:53 riche anyways
18:56 Grinnz_ sri: i only use one API with the json generator atm, and its documentation specifies "application/json", but i'm not sure if it would care
18:56 sri for the record, this description of escaping rules is pretty clear https://url.spec.whatwg.org/#urlencoded-serializing
18:57 sri whatwg++ # mostly pragmatic
18:57 riche that is basically Rule 4 from W2
18:57 sri : is 0x3A btw.
18:57 riche W3
18:57 sri w3c steals everything from whatwg
18:57 riche yes and chrome sends that
18:57 sri if there's two versions of a spec, whatwg is always right
18:57 riche but of course, this missed the first three rules
19:02 sri we are actually a bit more generous than the spec
19:02 sri * - . 0-9 A-Z _ a-z
19:02 odin22 joined #mojo
19:02 sri that's all it allows
19:04 disputin joined #mojo
19:05 Grinnz_ sri: looks like what chrome is sending on POST
19:07 riche okay i reopened it after looking at the code again.  at least let people comment on it.  but I am more nervous than yesterday because it seems the form data rules are differerent than URI and i didn't get that far yet
19:08 sri riche: you're right though, your pull request is 100% wrong
19:09 riche yes the _form method needs to directly call url_encode for the regex override
19:09 sri also the chars you've chosen
19:09 riche otherwise it is really messy and hoping for the best
19:09 riche yes
19:10 riche what can i say? i wanted to change as little as possible
19:10 riche but then it turned into this "thing"
19:12 riche okay I will change it now to match the rules on the whatwg
19:12 riche see what happens
19:16 Grinnz_ sri: i tested and the JSON api i am using does not care about the charset
19:16 riche interesting ... only test 78 failed extra which has the tilde ... that's a good sign
19:17 Grinnz_ sri: not much of a sample size unfortunately
19:24 riche created test for 0x20
19:24 riche Append U+002B to output.
19:24 riche and failed
19:24 riche so... a little more work
19:26 Zoffix left #mojo
19:27 batman sorry... i updated my comment on #786
19:40 sri allright, the problem becomes rather simple if we apply html rules for key value pairs, and rfc 3986 for raw query strings
19:40 Grinnz_ riche: 0x20 is converted to + in a separate step in Mojo::Parameters, after the url_escape currently
19:40 riche yes I saw that
19:40 Grinnz_ riche: as it should be.
19:41 riche deal is there was no test for $params->param('foo' => 'bar baz')
19:41 riche so i just wanted to make sure
19:41 Grinnz_ yes that is a good test to add
19:42 good_news_everyon joined #mojo
19:42 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/vfCQZ
19:42 good_news_everyon mojo/master b46e359 Sebastian Riedel: improve HTML Living Standard compliance of Mojo::Parameters (closes #786)
19:42 good_news_everyon left #mojo
19:44 sri key/value pars are urlencoded, raw string is the query part of a URL
19:45 Grinnz_ what is the raw string used for?
19:45 sri unusual query strings
19:46 sri Mojo::UserAgent needs to pass them along verbatim
19:46 Grinnz_ ah
19:46 sri like /ping?http://mojolicio.us/perldoc
19:46 Grinnz_ that's unusual alright...
19:46 sri i've seen it more than once in this channel
19:49 good_news_everyon joined #mojo
19:49 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/vfC5r
19:49 good_news_everyon mojo/master 7de701b Sebastian Riedel: mention the specs
19:49 good_news_everyon left #mojo
19:50 sri for the future
19:50 sri this will come up again, i'm sure
19:50 riche sri, i don't get what just happened
19:50 riche remember this is my first pull request so i don't know
19:51 sri i had a minute to look at the specs, and dictated a solution
19:51 riche okay so what i changed in my fork just now is irrelevant ... ok got it
19:51 sri (and grinnz browser research, which helped a lot)
19:53 punter joined #mojo
19:55 Zoffix joined #mojo
19:56 sri riche: you left the tilde in
19:57 sri which is RFC 3986, but not HTML Living Standard
19:57 Zoffix left #mojo
19:57 riche yah i was like 5 sec away from committing what you just did
19:57 riche anyways
19:57 riche thanks
19:57 sri well, then i would have applied it ;p
19:58 sri didn't want to ask anymore, since you seemed a little agitated
20:00 good_news_everyon joined #mojo
20:00 good_news_everyon [mojo] kraih pushed 3 new commits to master: http://git.io/vfCNj
20:00 good_news_everyon mojo/master d4c51e6 acp: Fix IE bug with decoding json response without encoding.
20:00 good_news_everyon mojo/master 2be5af4 acp: application/json MIME type updated in POD.
20:00 good_news_everyon mojo/master 050606c Sebastian Riedel: Merge pull request #785 from kak-tus/ie_json_res_encode_fix_2...
20:00 good_news_everyon left #mojo
20:07 good_news_everyon joined #mojo
20:07 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/vfCpM
20:07 good_news_everyon mojo/master b33df46 Sebastian Riedel: more names
20:07 good_news_everyon left #mojo
20:12 mst he quit
20:20 batman sri: i will be more active on github in the future. i was thinking i would just be a disturbance when i didn't know what i was talking about.
20:21 batman so... let me know when i start being a disturbance :)
20:21 jberger batman: that was the hard thing on this one, nobody knew anything
20:21 mst "as if a million users cried out in horror, and were suddenly silenced"
20:22 jberger hahahaha
20:22 purl LOLCON 4 reached.
20:22 batman jberger: yeah... not easy
20:27 jberger gods dammit! its snowing again!
20:30 sri it's summer... get your shit together chicago!
20:31 sri ... it's quite col around here too though
20:31 sri d
20:36 jberger been quite a bad year for the convertible
20:36 jberger a couple of good days, and a pricy repair for some hydraulic hoses
20:36 jberger not a great trade off so far :(
20:42 Zoffix joined #mojo
20:51 * Zoffix wonders whether there's anything akin to   $c->params(qw/foo bar baz ber/) instead of having to call $c->param for each one....
20:53 Grinnz_ Zoffix: removed in 6.0
20:53 Grinnz_ Zoffix: you could access it as a hash though
20:53 Zoffix Somewhere in $c->req-> ?
20:54 Grinnz_ @{$c->req->params->to_hash}{qw/foo bar baz ber/}
20:54 Zoffix >_<
20:55 Grinnz_ or however you'd do that with the silly postderef syntax.
20:55 Zoffix It's not silly, it's awesome :)
21:01 Zoffix Speaking of which.... Time to start using them in my app
21:01 jberger $c->req->params->to_hash->@{qw/foo bar baz ber/}
21:01 jberger how is that silly?
21:01 Zoffix Yeah, looks much nicers
21:03 memowe joined #mojo
21:06 Grinnz_ ok, that does look nice i admit
21:07 Grinnz_ no asterisk makes it better
21:08 trone joined #mojo
21:08 sri purl: set Grinnz on fire
21:08 * purl sets Grinnz on fire
21:09 * Grinnz_ sets asterisks on fire
21:09 Zoffix purl, set purl on fire
21:09 purl Zoffix: sorry...
21:09 Grinnz_ lol
21:09 * sri waits for someone to say "purl: set sri on fire"
21:09 * Zoffix already knows what happens :)
21:09 sri :p
21:16 jb360 joined #mojo
21:19 mattastrophe joined #mojo
21:23 hernan604 purl: set sri on fire
21:23 * purl sets hernan604 on fire
21:23 hernan604 :p
21:23 hernan604 Dont press that red button!!
21:31 Zoffix purl, set sr​i on fire
21:31 purl Zoffix: what?
21:31 Zoffix awww.... my amazing clever hack failed :(
21:34 Zoffix purl, set sr​i on fire
21:34 * purl sets sri on fire
21:34 Zoffix \o/
21:49 jberger purl: set Zoffix on fire
21:49 purl jberger: i'm not following you...
21:50 sri -.-
21:50 Zoffix :O
21:52 jberger burn it to the ground
21:52 * purl sets jberger on fire
21:55 Grinnz_ lol
22:10 dave purl, what is the meaning of fire?
22:10 purl i haven't a clue, dave
22:33 Grinnz_ SELECT thing FROM table WHERE id BETWEEN 100 and 150 GROUP BY id HAVING id=120; # null           SELECT thing FROM table WHERE id=120 GROUP BY id; # correct result
22:33 Grinnz_ i fucking hate mysql
22:36 Grinnz_ taking the group by out of the first query also solves it, but this select happens to also include some aggregate functions
22:54 Kripton joined #mojo
23:35 jabberwok http://deps.cpantesters.org/depended-on-by.pl?module=Mojolicious     in case anyone wonders what CPAN modules depend on Mojolicious.\
23:40 jberger jabberwok: hehe, I already tried it :-)
23:42 Zoffix That's quite a list :)

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