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

IRC log for #mojo, 2014-09-20

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

All times shown according to UTC.

Time Nick Message
00:28 d4rkie joined #mojo
00:35 agorman joined #mojo
00:47 preaction joined #mojo
02:24 KCL_ joined #mojo
02:29 ignacio_ joined #mojo
02:32 tbushell8512 joined #mojo
02:42 noganex joined #mojo
02:49 tbushell8512 left #mojo
02:58 d4rkie joined #mojo
03:17 basic6_ joined #mojo
03:22 sujithm joined #mojo
03:22 lipizzan joined #mojo
03:27 firnsy joined #mojo
05:22 sh4 joined #mojo
05:33 irq joined #mojo
06:13 phillipadsmith joined #mojo
06:13 jalie joined #mojo
06:17 chansen joined #mojo
06:18 ceej_ joined #mojo
06:33 jamesaxl joined #mojo
06:37 dod joined #mojo
06:40 Vandal joined #mojo
07:12 sujithm joined #mojo
07:16 marcus sri: that does look nice indeed.
07:23 mr-fooba_ joined #mojo
07:49 mr-foobar joined #mojo
07:56 mr-foobar joined #mojo
08:01 sujithm_ joined #mojo
08:47 doublelel joined #mojo
09:02 batman holy crap! that looks very nice :)
09:02 batman helsinki++
09:21 basiliscos joined #mojo
09:34 _eugen joined #mojo
09:47 sujithm joined #mojo
09:55 sujithm_ joined #mojo
09:56 mr-foobar joined #mojo
10:02 denis_boyun joined #mojo
10:13 irq_ joined #mojo
10:17 woz joined #mojo
10:19 ladnaV joined #mojo
10:59 woz joined #mojo
11:05 woz joined #mojo
11:13 KCL joined #mojo
12:30 Akron joined #mojo
12:39 neilhwatson joined #mojo
12:49 Akron thomasoniii++
12:49 Akron I like running Mojo stuff on my Raspi - it works very well.
13:12 sujithm joined #mojo
13:16 d4rkie joined #mojo
14:28 woz joined #mojo
14:51 cpan_mojo Mojolicious-Plugin-JSUrlFor 0.14 by Viktor Tuskyi - http://metacpan.org/release/KOORCHIK/Mojolicious-Plugin-JSUrlFor-0.14
14:54 woz joined #mojo
14:59 basiliscos joined #mojo
15:25 woz joined #mojo
16:13 dotandimet joined #mojo
16:30 irq joined #mojo
16:43 woz joined #mojo
16:43 sujithm joined #mojo
16:44 * sri sooooo wants to use signatures in the mojo docs
16:48 sri https://gist.github.com/anonymous/3a012d544df830e92de8
16:49 sri the frontpage example looks so much more modern
16:49 dotandimet joined #mojo
17:18 jamesaxl joined #mojo
17:19 dotandimet joined #mojo
17:22 marcus sri: looks good!
17:41 dotandimet joined #mojo
17:45 woz joined #mojo
18:17 Akron joined #mojo
18:18 dexteruk_ joined #mojo
18:34 dexteruk joined #mojo
18:46 woz joined #mojo
19:20 sri yea, we really need to find a way to have signatures in mojolicious
19:21 sri it's a shame there are so few xs hackers that can make it happen
19:33 woz joined #mojo
19:33 preaction joined #mojo
19:43 irq joined #mojo
19:47 sri i wonder what would happen if the docs recommended 5.20 instead of 5.18 and a few examples contained "use experimental 'signatures';"
19:48 sri guess it would cause too much trouble
20:06 marcus I guess we should just wait for current distros to ship with 5.20.. next year or so?
20:07 marcus yosimite will ship with 5.18
20:08 sri hopefully we'll have a polyfill by then and can just *use* signatures everywhere
20:09 mst have you experimented with 'use signatures;' ?
20:09 mst that's like 80% of a polyfill
20:10 sri briefly, biggest annoyance was actually the :proto attribute
20:11 sri i imagine lack of arity checks would make upgrading perl a little tricky, since nobody will get it right if they are not forced to in the first place
20:13 sri oh wow, i didn't realize it actually passes tests on 5.8.9
20:13 sri that's crazy
20:13 sri that's a big plus
20:15 sri it might actually be worth it :o
20:16 sri we would have to document that we are using the signatures subset shared by the core feature and CPAN module though
20:17 * sri pokes marcus, batman, tempire, jberger, crab
20:17 marcus sri: how would your example above look?
20:17 mst exactly the same
20:17 sri yea
20:17 mst it's only the arity and the defaults stuff that's not in signatures.pm
20:17 sri Mojo::Base would do the magic
20:18 mst and that could be fixed, it's just nobody's done so yet
20:18 sri the prototype attribute can be worked around with Scalar::Util::set_prototype i guess
20:20 marcus That sounds pretty sweet to me.
20:23 marcus sri: but what about the signature depedency in Makefile.PL? It'll be there even for 5.20?
20:23 sri nope, i'd make it optional
20:23 marcus *signatures
20:23 marcus ok
20:23 sri no deps on 5.20.0
20:27 sri i'll cook up a branch
20:27 sri this will be fun
20:30 marcus things to do on a saturday night during the slow revolution.
20:39 sri mst++ # asking the right questions
20:39 sri haha, 28 modules installed on 5.10.1
20:40 sri not that i really care ;p
20:56 berov joined #mojo
21:04 sri argh, i know i'm sooo close, but i can't seem to get it to work
21:08 cpan_mojo Mojolicious-Plugin-BootstrapHelpers 0.0174 by CSSON - http://metacpan.org/release/CSSON/Mojolicious-Plugin-BootstrapHelpers-0.0174
21:26 sri https://gist.github.com/anonymous/84440b1fb221b7b09835#file-signatures-diff-L48
21:26 sri failing pretty hard there
21:29 sri and when i say failing pretty hard i mean https://gist.github.com/anonymous/fd1af304461f863d5983
21:30 marcus wow, scary
21:31 sri http://m.c.lnkd.licdn.com/mpr/mpr/p/2/005/069/03a/0734a49.jpg
21:31 marcus I'm drinking cognac and watching that meta talk from strangeloop
21:36 jesteves joined #mojo
21:38 sri yea, i have no clue how to import signatures into the uplevel scope
21:39 sri mst: got an idea?
21:40 * sri has no clue how these endofscope hooks work
21:43 jesteves Hello all.  Maybe a basic question, but can't find my way around it.  What is the best way to access "raw" body of a request?
21:44 sri http://mojolicio.us/perldoc/Mojolicious/Controller#req
21:46 jesteves I'm using $request->content->build_body, both client and server side, cause I need to sign and compare signature of the body, but...
21:49 sri i have a feeling the endofscope hooks might make adding signatures support to Mojo::Base impossible :S
21:49 jesteves when the content is multipart, I'm getting headers on the parts (Content-disposition; Content-type) in different order, and then my signature check fails
21:51 Leucomelas joined #mojo
21:51 sri hook after_build_tx => sub { shift->req->content->auto_upgrade(0) }; or so
22:03 sri oh shit
22:03 sri https://rt.cpan.org/Public/Bug/Display.html?id=80450
22:03 sri it's the fricking one-liner
22:04 sri one thing is clear now though
22:05 sri we could totally activate signatures on 5.20.0+
22:07 sri then document the fact (use 5.20.0+ if you want signatures), recommend 5.20, and use them in the docs ;p
22:10 nicomen ame issue that perl-mop has with -e ?
22:10 nicomen *same
22:12 jesteves sri: thanks: disabling auto-upgrading works for me
22:13 good_news_everyon joined #mojo
22:13 good_news_everyon [mojo] kraih created signatures (+1 new commit): http://git.io/noqBWg
22:13 good_news_everyon mojo/signatures f0ff29c Sebastian Riedel: added basic support for Perl 5.20 signatures
22:13 good_news_everyon left #mojo
22:13 sri well, that's how far i got
22:14 sri the signatures module is bad, really bad
22:17 jesteves left #mojo
22:21 sri honestly, considering it's just a polyfill, i wouldn't even mind a source filter... lol
22:25 sri might even be better, since more people can hack on it
23:01 jberger this might be more crazy, but aren't the syntaxes between the cpan module fairly consistent
23:01 jberger could older perls use Method::Signatures or Function::Parameters or Kavorka?
23:05 jberger ye gods! look at these caveats!: https://metacpan.org/pod/Kavorka#CAVEATS
23:08 jberger hmmm, a couple times I have thought about using Moops to make an code-sexy example, maybe that should be a weekend project
23:12 woz joined #mojo
23:13 KCL_ joined #mojo
23:14 sri jberger: recommendations?
23:15 purl recommendations are good.  Even has little local indie bands.
23:15 jberger of the three I would guess Function::Parameters
23:15 sri jberger: as far as i can see Kavorka doesn't even extend the sub keyword!
23:16 jberger although everything that Toby Inkster does it top notch IMO
23:16 sri jberger: also doesn't extend the sub keyword
23:16 sri jberger: what am i missing?
23:17 jberger I don't know if any of them do or even if they can
23:17 sri :S
23:17 jberger they all trigger on an installed keyword
23:18 sri so, your proposal doesn't work at all
23:18 jberger it wasn't a fully formed idea to be sure :-)
23:18 * jberger likes to spitball, sri knows this
23:19 * sri gets easily frustrated, jberger knows this :)
23:19 jberger truthfully, I know its hard to wait, but I think we should wait until a non-experimental signatures is released
23:20 * sri runs aimlessly yelling through the channel
23:20 D4RK-PH0ENiX joined #mojo
23:20 jberger that or fully throw our support behind one of the CPAN modules
23:21 sri all my apps now contain "use experimental 'signatures'"
23:22 sri i mean, seriously look at that app using signatures above
23:22 sri don't you want that on the frontpage?
23:23 jberger of course, but putting experimental code into mojo is asking for trouble
23:23 jberger mojo is cutting edge, yes
23:23 sri if nobody starts using signatures, they will never go stable
23:23 jberger but its also not reckless or no one will base critical apps on it
23:24 jberger sri: yes it will
23:24 sri when was the last time an experimental feature went stable?
23:24 jberger many of them did in the last revamp
23:24 jberger experimental use to just be a designation in the pod, now its codified
23:24 sri and what do you think was the shortest timeframe?
23:25 jberger no one care about it until the warnings came
23:25 sri there are no criteria for stabilizing an experimental feature
23:25 jberger PDL used lvalue subroutines for 15 years before they were just recently made non-experimental
23:25 jberger I think the criteria is best explained in ~~
23:25 sri do you think lvalue subs would have become stable if you didn't use them?
23:26 sri you are arguing for me!!!
23:26 jberger they know its going to change, so its experimental
23:26 sri you've been using experimental lvalue subs for 15 years
23:26 jberger ok pause
23:26 sri you're just making assumptions, nothing you say is written down
23:26 jberger I'm only marking time since using an experimental feature emitted a warning
23:27 jberger before then it was just guidelines
23:27 jberger now I think there is more impetus to keep up with what things are or are not experimental
23:27 jberger why were lvalue subs "experimental" for 15 years, because no one ever went back and un-marked them
23:27 jberger because it didn't matter
23:28 sri if nobody uses signatures now, it won't matter either
23:28 jberger since the experimental pragma and warning etc things are different
23:28 jberger I think people will now argue for them
23:28 sri i see your point, and i strongly disagree
23:29 jberger but if too many people (and major projects) start enabling them by default, p5p will get gunshy about changing them in order to get them into a good state
23:29 jberger so, yes, use them in your non-cpan work, but I don't think we should see any cpan modules using it
23:29 jberger then again, that is my personal belief
23:30 sri honestly, if that's truly how perl evolves, i would be done with the community and leave
23:30 jberger the other side of the argument is that mojo is a good candidate for using it on cpan
23:30 sri there is just no way i can wait 5 years to use a feature like signatures
23:31 jberger since the dev team is responsive and if p5p did change it we would react very quickly
23:31 sri maybe i should fork mojolicious and create a more bleeding edge framework
23:31 jberger and I think they know that that is true
23:32 jberger sri: I'm all for being on the bleeding edge, especially now that there is carton
23:32 sri tempire, batman, marcus, crab: is that the general consensus? wait until signatures become stable?
23:32 jberger has there been any chatter about going stable?
23:32 sri nothing at all
23:32 purl nothing at all is probably in core. or what stupid sexy flanders says
23:32 sri not a single thread on p5p
23:32 jberger start one?
23:33 jberger sawyer had his rather silly "where are we going" thread, adding absolutely nothing
23:33 jberger that one made me really sad
23:33 sri that would be counter productive, some people want to disagree with me just for the sake of it
23:34 jberger would it be better if I did it?
23:34 sri ideally someone like peter martini did
23:34 jberger or else, does anyone know where peter martini hangs out? I would happily chat with him
23:34 sri no clue
23:36 sri but seriously, maybe it is time to out mojolicious into maint mode and do development in a much more aggressive fork with a new name
23:36 sri s/out/put/
23:38 jberger what would be the idea?
23:38 jberger futures?
23:38 purl futures is a zero-sum game.
23:38 sri first of all signatures everywhere
23:38 * jberger farts in purl's general direction
23:39 sri there is so much value in just having the documentation use signatures everywhere
23:39 saki stevan little's work on p5-mop(-redux) seems to have stalled, but it was looking like the signatures in that would quickly make non-experimental core. haven't heard anything in months on it, so that seems to have stalled :(
23:39 sri saki: stevan is moving to amsterdam
23:39 jberger sri: the day signatures go core, I say we bump to 6.0, require 5.22, and use signatures everywhere sri
23:39 sri umm, they are in core
23:40 jberger non-experimental core
23:40 sri that sounds bad actually, it's so easy to support 5.20 with a small workaround too
23:40 jberger I have a deeper dream of making a fully libuv subsystem for perl/mojo (ala node)
23:42 jberger I would say that signatures everywhere (especially in docs) would make mojo much more approchable for non-perl users
23:43 woz joined #mojo
23:43 jberger sri: how many places would the mojo codebase benefit from s///r?
23:43 sri no idea
23:43 jberger I'm game for bumping to 5.16 or so for 6.0 anyway
23:44 sri i don't care enough about those small features
23:44 jberger __sub__
23:44 jberger but no, signatures is obviously the prize
23:44 purl okay, jberger.
23:44 jberger purl, signatures
23:44 purl signatures is obviously the prize
23:44 jberger nice
23:45 sri there is going to be a huge fight
23:45 sri and signatures is the feature i'm personally willing to fight and sacrifice users for
23:45 sri nothing else comes close
23:45 jberger as you say, its basically feature complete at this point
23:46 jberger so if users dont want to upgrade then fine
23:47 jberger sri: ONLY as a conversation point, what else would do/add/change in a fork/major break?
23:47 preaction that reminds me that i should add 5.20 to $work as a techdebt ticket
23:48 jberger preaction: you think $work cares about techdebt?
23:48 preaction keep mojolicious, let people fork 5.x if they want backcompat
23:48 preaction jberger: $work doesn't care about my team, so i can do basically whatever i want ;) and what i want is to get rid of techdebt
23:48 * jberger wishes he was on preaction's team
23:49 preaction they don't care about us, so they're not going to give us money to hire people
23:49 jberger my job seems to be about creating techdebt
23:49 jberger :(
23:49 * jberger should probably shut up
23:49 jberger everything is awsome :D
23:50 sri jberger: dunno, look through perldelta for more stuff we might benefit from i guess
23:50 preaction that's all programming though. it's all techdebt, or techdebt-in-waiting
23:50 sri i would prolly use postderef too
23:50 jberger oh postderef
23:50 preaction i'd say let the users create a Mojolessious if they really want to stay in the past
23:50 sri get rid of the @{} everybody hates
23:50 jberger yummy
23:51 sri seriously, it's mostly about making mojo apps look better in the documentation
23:51 sri i would base the docs on the latest perl release i guess
23:51 preaction and that to the non-perl programmer, which is who all perl projects should be going after
23:52 jberger sri: I guess I'm starting to soften at least for the docs
23:52 sri "This documentation will always use the latest Perl features, you may have to adapt examples for versions before 5.20.0."
23:52 jberger remember, my realization at mojoconf was the "dark user" experience
23:53 jberger it could go right in the "Learning Perl" section
23:53 * jberger reminds himself to make his "dark user" post
23:53 preaction alternatively, we've got a project that has tabs to switch between various language's examples. you could do similar for perl versions
23:53 * jberger worries its just going to spark another perl 7 firestorm
23:54 preaction more work, more annoying, but helpful for those stuck in the dark ages
23:54 sri preaction: how do you do that in POD?
23:54 preaction probably =for sections and some processing/templating
23:55 preaction yeah. i like that idea less now
23:55 * jberger wishes he liked podweaver
23:56 sri =for Legacy Perl (you should be ashamed of yourself!)
23:56 preaction podweaver is awesome! :p
23:57 sri __SUB__ is a good example for something we should be teaching in the docs btw.
23:58 sri YEA LETS TEACH PEOPLE HOW TO LEAK BECAUSE WE CAN'T USE MODERN FEATURES YET!
23:58 jberger I still love Mojo::FriendFeed, might be my proudest cpan module
23:58 jberger that or Tie::Array::CSV
23:58 jberger which is nuts

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