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

IRC log for #mojo, 2016-06-16

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

All times shown according to UTC.

Time Nick Message
00:06 sri hey there mr fancy pants
00:30 tempire \o
00:58 zivester joined #mojo
01:07 Kripton joined #mojo
01:14 good_news_everyon joined #mojo
01:14 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vo0UN
01:14 good_news_everyon mojo/master 3be34dd Sebastian Riedel: fix support for CONNECT requests without keep alive connections in Mojo::UserAgent (closes #970)
01:14 good_news_everyon left #mojo
01:15 sri "6.66  2016-06-16" is funny
01:22 genio https://channel9.msdn.com/Blogs/Seth-Juarez/Windows-Subsystem-for-Linux-File-System?WT.mc_id=dlvr_twitter_ch9  Some good information in there
01:35 Grinnz but it's only the 15th! :o
01:38 sri here it's already the 16th
01:41 Grinnz do they hve flying cars yet??
01:42 good_news_everyon joined #mojo
01:46 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vo0kD
01:46 good_news_everyon mojo/master ddae356 Sebastian Riedel: check the method a little earlier
01:46 good_news_everyon left #mojo
01:52 sri nope, just inflated bitcoin prices so far :o
01:53 sri $731 now
01:53 sri that's +$40 since i last looked
02:28 noganex_ joined #mojo
02:29 marty joined #mojo
03:11 mcsnolte joined #mojo
03:13 tempire man
03:13 tempire I wish I would have kept my bitcoins from when they were 300
03:13 tempire I just bought 10 shares of apple stock for that reason
03:27 tempire Oh hey
03:27 tempire I have 1.25 bitcoins
03:27 tempire 💰
04:42 jberger genio: in the beginning they mention a previous video about processes
04:42 jberger Here it is
04:42 jberger https://channel9.msdn.com/Blogs/Seth-Juarez/Windows-Subsystem-for-Linux-Process-Architecture
04:42 jberger And if is really interesting too
04:42 jberger It
04:43 jberger For example NT always had fork but it wasn't exposed to the win32 subsystem
04:43 jberger Because NT originally had a posix subsystem which never really came to being
05:04 inokenty-w joined #mojo
05:08 preaction i thought they got that thing POSIX-certified even. they just stopped caring when nobody else cared
05:52 che-quest joined #mojo
06:02 dod joined #mojo
06:06 elrey joined #mojo
06:07 dod joined #mojo
06:11 elrey Hey. I'm trying to get the "over" routing working, to block users that are not authorized.
06:14 elrey im using: $r->get('/route')->over(user=>'authorized')->to('Status#status'); but even if i pass wrong "user", its going through anyway
06:49 Grinnz joined #mojo
06:53 anparker elrey: I'm using "under" for that. http://mojolicious.org/perldoc/Mojolicious/Guides/Routing#Under
06:55 McA joined #mojo
07:01 cpan_mojo Mojolicious-Plugin-Google-Cloud-UserAgent-0.02 by SCOTTW https://metacpan.org/release/SCOTTW/Mojolicious-Plugin-Google-Cloud-UserAgent-0.02
07:10 trone joined #mojo
07:14 che-quest elrey: show your user sub for routes->add_condition
07:16 trone joined #mojo
07:17 che-quest elrey: or use the Mojolicious::Plugin::Authentication
07:42 Vandal joined #mojo
07:43 elrey solved. the problem was in my add_condition, didnt realized that i need to return nothing to block
07:43 elrey thanks
08:23 icjs joined #mojo
08:25 Adura_ joined #mojo
08:30 Adura_ joined #mojo
08:54 che-quest hi everyperl, i have action without explicite $c->render and in template there isn't my controller, see https://github.com/kraih/mojo/commit/794bdf2206b278f12ab8ded3722859a1024f5b9f#commitcomment-17856576
08:54 trone joined #mojo
09:03 stephan joined #mojo
09:20 mpa joined #mojo
09:21 mpa left #mojo
09:21 mpa joined #mojo
09:59 punter joined #mojo
10:09 meshl joined #mojo
10:45 sri anyone disagree? https://github.com/kraih/mojo/issues/972
10:46 kaare joined #mojo
10:48 pink_mist no, I agree completely
10:49 Kripton joined #mojo
10:54 mpa left #mojo
10:55 dvinciguerra_ joined #mojo
10:59 tchaves joined #mojo
11:04 tchaves joined #mojo
11:10 lluad joined #mojo
11:13 sri oh, another interesting book bundle https://www.humblebundle.com/books/make-electronics-book-bundle
11:13 osfabibisi joined #mojo
11:21 sri che-quest: that's documented behavior
11:26 che-quest sri: please explain the logic of diffrent controllers switch render
11:26 sri no time i'm afraid
11:27 che-quest ok
12:05 nic ooh, I fancy the Synthesizers book
12:05 dod joined #mojo
12:06 good_news_everyon joined #mojo
12:06 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vo0AI
12:06 good_news_everyon mojo/master 6d5cf25 Sebastian Riedel: fix a few bugs in proxy tests
12:06 good_news_everyon left #mojo
12:09 sri this change might be arguable https://github.com/kraih/mojo/commit/6d5cf259b42ce17981e1b213656deb316e14e736#diff-955b07afef3c0b8632a50118b370262cR129
12:09 sri it puts the CONNECT request into the ->previous attribute of the real transaction
12:10 sri CONNECT requests do get a start event emitted for them too though, so this seems kinda consistent
12:10 sri and it's good for testing
12:12 nic In Minion, when I call $minion->add_task(...), at what point can I be confident it won't be lost (ie it has been committed to backend)?  Is it when a worker claims it as a job?
12:13 nic oh, I bet I'm confusing two different things
12:13 good_news_everyon joined #mojo
12:13 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vo0AF
12:13 good_news_everyon mojo/master d5a5b63 Sebastian Riedel: change order a little bit
12:13 good_news_everyon left #mojo
12:28 jberger nic: you really can only call it during startup
12:28 good_news_everyon joined #mojo
12:28 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vo0pz
12:28 good_news_everyon mojo/master aba6466 Sebastian Riedel: a few more CONNECT tests
12:28 good_news_everyon left #mojo
12:29 sri hmm, hope the name of this method is not misleading now https://github.com/kraih/mojo/blob/master/lib/Mojo/Transaction/HTTP.pm#L44-L49
12:29 nic jberger: cheers
12:30 jberger sri: i wish i could contribute on this but i know almost nothing on the topic
12:30 sri no need to know internals, it's api design
12:31 sri right now we use $tx->previous to link redirects
12:31 jberger Really i mean i know very little about CONNECT
12:31 jberger Sorry, go on
12:32 sri not much to know really
12:32 sri this is a CONNECT request https://gist.github.com/anonymous/7dd97b16d2d666d7e17bebac068f0403
12:32 jabberwok saw the bit about json booleans.  the real question is why anyone would create a protocol that uses four characters "t", "r", "u", and "e" or five characters to write out the English word for false to mean 1 and 0.  grrrr
12:33 sri our user agent connects to the proxy server, performs a CONNECT request to make the proxy connect to the real target and responds with a 200 if it was successful
12:33 sri and then the user agent sends the real request over the same connection
12:33 sri the proxy just forwards raw bytes from then on
12:34 jberger Ok
12:34 sri CONNECT is an http method that creates a tunnel basically
12:35 jberger So now what is the change at issue
12:43 sri CONNECT requests are now included in $tx->previous
12:43 sri before it was only followed redirects
12:43 sri and we have this method using ->previous https://github.com/kraih/mojo/blob/master/lib/Mojo/Transaction/HTTP.pm#L44-L49
12:49 CW joined #mojo
12:49 anparker what can you do with a connection without handle?
12:51 nic ok, it's a day for me to level-up my perl
12:51 nic my $worker = $guard->_workers->{$id} if $id;
12:51 nic That does _not_ do what I expected
12:52 pink_mist you're using an internal method there ... you might not want to do that
12:52 Adura joined #mojo
12:52 nic my $worker = $id ? $guard->_workers->{$id} : undef;
12:52 nic That's the correct version
12:52 nic I know sri rewrote some code to fix that problem, but I don't understand the problem
12:53 zivester joined #mojo
12:53 nic In the first example, $worker can become defined even if $id is undef
12:53 jabberwok hmm, writing tests.  since Command::test effectively uses File::Find::find() which is "depth first", can I assume anything about the default order of *.t versus random? or what is recommended way when i have create_db.t, create_users.t, test_users.t and similar dependent cascading tests
12:53 pink_mist nic: the problem you just had? it's because if $id isn't true, not even the *my* gets run, so the variable can give you nasal demons
12:54 pink_mist nic: in this case, it retained its previous value
12:54 pink_mist instead of the nasal demons
12:54 pink_mist but only because you were lucky
12:54 nic I'm not sure that's right
12:54 pink_mist nic: it's an old trick to create "state" variables, which could lead to perl segfaulting
12:55 nic there was no $worker in scope other than that declaration
12:55 pink_mist nic: yes, the value from the previous time *that* my was run
12:55 nic oh, something internal?
12:55 pink_mist no
12:56 pink_mist nic: https://metacpan.org/pod/Perl::Critic::Policy::Freenode::ConditionalDeclarations
12:56 * nic reads
12:58 nic heh, their use of "is unclear" is unclear
12:58 pink_mist and where that document mentions "undefined", it's "undefined" in the C sense of the word
12:58 nic They make it sound like it's a problem for people reading, but it sounds from what you're saying that the 'my' can be short-circuited
12:58 pink_mist ergo, this sense of the word: http://catb.org/jargon/html/N/nasal-demons.html
12:59 pink_mist nic: it can potentially lead to the internal state of the perl interpreter to be fucked up royally and there's no telling what it may cause.
12:59 nic any links to back that up?
13:00 pink_mist no, I don't have time to look for links right now
13:10 ramortegui joined #mojo
13:14 nic np, thanks for the Critic link
13:14 sri looks like nobody really cares about including CONNECT requests in follow-up requests
13:14 sri http://mojolicious.org/perldoc/Mojo/Transaction/HTTP#redirects
13:15 sri so i guess it's fine, the ->redirects description is vague enough to include CONNECT too
13:16 nic I can't see a problem with doing that
13:17 nic Nor can I think of a concise way of mentioning it :)
13:25 marty joined #mojo
13:25 anparker > with all previous transactions - it's alreday documented behaviour)
13:28 jberger sri: sorry i had to step away, i think I'm OK with that resolution if it fixes the other problem
13:28 gizmomathboy joined #mojo
13:28 sri in this case it made connection leaks easy to test
13:33 dod joined #mojo
13:33 bwf joined #mojo
13:34 marty_ joined #mojo
13:38 Vandal joined #mojo
13:43 jacoby joined #mojo
13:50 ptolemarch joined #mojo
14:08 good_news_everyon joined #mojo
14:08 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/voEOu
14:08 good_news_everyon mojo/master 1098b87 Sebastian Riedel: another leak test
14:08 good_news_everyon left #mojo
14:14 Grinnz jabberwok, true/false in JSON is directly from true/false in javascript
14:14 Grinnz many languages have direct boolean types like that
14:17 Bean joined #mojo
14:24 Atog so we got any evil feature for mojo 6.66? heh
14:25 jnbek yea, like, when a Mojo::UA call is made, it plays randomly selected songs from premier norwegian black metal bands?
14:25 jnbek that'd be, so cool!
14:25 * jnbek runs away
14:26 * coolo just had to find out the hard way that perl 5.24 now calls AUTOLOAD for DESTROY - that gave an interesting result on shutdown ;(
14:26 Atog some random redirects to those cheerful Immortal videos
14:26 jnbek yes!
14:26 jnbek damn it, now I need to hear some...
14:30 jberger coolo: that isn't new
14:31 jberger and it is why if you have a base class that implements AUTOLOAD you should probably implement a bare DESTROY too
14:31 coolo jberger: not to you, but it's new for someone who updates from 5.22 to 5.24 :)
14:31 coolo (documented in perldelta)
14:31 jberger https://github.com/kraih/mojo/blob/master/lib/Mojo/Base.pm#L19
14:32 jberger I'll have to read that then, I was pretty sure that that behavior had existed for ages
14:32 coolo the exact wording is "The AUTOLOAD method is now called when searching for a DESTROY method, and correctly sets $AUTOLOAD too.  [perl #124387]  [perl #127494]"
14:32 coolo but I bet it was best practise to have a DESTROY {} even before
14:34 jberger asking in #p5p because I'm curious
14:37 jberger yeah, https://rt.perl.org/Public/Bug/Display.html?id=124387
14:38 jberger an unintended change removed the behavior, it was reintroduced in 5.24
14:38 jberger I'm kinda amazed it went that long
14:41 * coolo wonders why I follow Gabor Szabo on G+ - I don't remember the story :)
15:05 zivester joined #mojo
15:13 sri hmm, all the recent activity on the mailing-list cost us 5 subscribers
15:14 jberger really?
15:15 * jberger finally signed up for daily digest to keep a closer eye on the mailing list without all the actual traffic
15:17 sri most people just have email notifications disabled
15:18 jberger until about a week ago I just went to the web view when something came up :P
15:18 sri hehe, i still have email enabled but use the web view for posting ;p
15:20 Grinnz_ i just have a folder in gmail, like anything else :P
15:20 Grinnz_ sorry, "label"
15:21 genio email lists get annoying to me pretty quickly.  I end up shunting them into a label and pretending like they don't exist.
15:23 Grinnz_ shunting to a label works pretty well, so i can immerse myself in fedora drama when I choose ;)
15:25 disputin joined #mojo
16:23 sri ouch, so dell keeps the silly windows key on their developer edition xps, but adds a silly ubuntu sticker http://cdn.arstechnica.net/wp-content/uploads/2016/06/dell-xps-13_2016-05-28_174308.jpg
16:24 Grinnz_ lol
16:25 * sri just stumbled over this review http://arstechnica.com/gadgets/2016/06/the-xps-13-de-dell-continues-to-build-a-reliable-linux-lineage/
16:25 Grinnz_ I'm not sure how else you'd label a super key tbh
16:25 dod joined #mojo
16:25 Grinnz_ it's alwyas been the windows key on keyboards
16:25 sri penguin key?
16:25 Grinnz_ haha
16:25 sri can't be that expensive to put different key caps into the box
16:26 sri they already make key caps for like a million languages
16:26 Grinnz_ you can swap out chiclet keys?
16:27 Grinnz_ http://shop.daskeyboard.com/products/linux-key-cap-bundle-1 hehe
16:28 Grinnz_ those cost almost as much as a basic keyboard does :P
16:30 sri :)
16:35 marty joined #mojo
16:37 marty_ joined #mojo
16:43 ptolemarch joined #mojo
16:44 disputin joined #mojo
16:47 disputin joined #mojo
16:48 odc joined #mojo
16:53 sri guess i'll have to release 6.66 today, the date is just too good ;p
17:11 PryMar56 joined #mojo
17:26 disputin joined #mojo
18:04 * jberger expects an easter egg
18:05 sri ideas?
18:06 lluad Dancing satan.
18:07 preaction Robot Devil
18:07 sri https://gist.github.com/anonymous/eb4bdf9268e21f9cd1994e6f8d3363bb
18:08 sri ?
18:08 Bean it's beautiful
18:14 odc +1000
18:16 cpan_mojo Mojolicious-Plugin-Google-Cloud-UserAgent-0.03 by SCOTTW https://metacpan.org/release/SCOTTW/Mojolicious-Plugin-Google-Cloud-UserAgent-0.03
18:19 tempire joined #mojo
18:28 good_news_everyon joined #mojo
18:28 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/voued
18:28 good_news_everyon mojo/master c93afc2 Sebastian Riedel: :octopus:
18:28 good_news_everyon left #mojo
18:37 mcsnolte joined #mojo
18:42 Kupank joined #mojo
18:45 jberger I love it
18:58 dmanto joined #mojo
19:03 dmanto anyone using plugin AssetPack for jquery-mobile library?
19:05 tempire_ joined #mojo
19:05 dmanto I'm having problems with the "loader" image
19:05 tempire joined #mojo
19:05 dmanto it should be in images/ajax-loader.gif
19:21 dmanto I mean looking in dev tools, it should be at /asset/xxxxxxxxxxx/images/ajax-loader.gif, (xxxxxxxxxxxx beeing jquery-mobile assets checksum), but it is not there so I get a 404 Not Found when the library tries to get it
19:48 cfedde are there any classes or methods available for parsing the Link: header?
19:49 dave https://cthulhuforamerica.com/
19:52 martin joined #mojo
19:53 cfedde I was with you till you got to "non-Euclidian infrastructure"
19:56 mpapec pronunciation, https://www.youtube.com/watch?v=sOA5Chz3iU4
19:57 disputin joined #mojo
20:03 cfedde Still I'm not quite clear on how to apply mechanistic materialism in this context.
20:08 sri cfedde: for a short time we had a Link parser, but it got removed again due to lack of interest
20:09 cfedde kk.  I;ll just use a few splits and some s///r
20:10 Nei joined #mojo
20:10 itaipu joined #mojo
20:10 cfedde github stuffs multipage result set hints into a link header.
20:11 sri this was one attempt https://github.com/kraih/mojo/commit/b5c4de20d3f75b4e17fb583af105963d4fcce184
20:11 Nei hi, has anyone made a Mojo::UserAgent with transparent Cache of ETag?
20:11 * cfedde goes to read
20:12 sri split doesn't work, because the link can contain semicolons
20:12 sri parsing it being tricky was what made me consider a special method
20:14 sri there's more missing from the patch
20:14 * cfedde goes back to the RFC.
20:14 sri https://github.com/kraih/mojo/commit/fdd6c579b31145bed0f762544c94013f57a3851d
20:14 cfedde A bnf...
20:15 sri those patches go together
20:15 sri yea, it's really not a simple format
20:16 sri annoyed the hell out of me back then, because it looks so simple at first :)
20:16 cfedde heh.
20:16 cfedde yeah.
20:17 * cfedde finds himself in a maze of twisty production grammars.  All different.
20:17 sri have Link headers become more common since then?
20:18 sri think at the time we could pretty much only find two uses in the wild, github api and riak
20:18 cfedde I have no idea.  I have a silly little itch to scratch. Making a csv out of github issues for DaBoss.
20:18 sri ah, github again :)
20:19 Nei cant you use github api for that instead
20:19 cfedde Nei: that's what I'm doing.
20:20 Grinnz_ https://cdn.meme.am/instances/60705911.jpg
20:20 cfedde rather than giving you back all 10000 issues in one big json it breaks it up into "pages" using rel=next Link headers.
20:21 Nei ah right, it uses http headers for that
20:21 Nei <....>;rel=..., and so on. meh
20:21 cfedde my case is trivial.  But why invent a wheel if one already exists.
20:22 Nei well what about simply using a github api module on cpan surely that has solbed this problem?
20:22 cfedde maybe.  but .. Mojo is soo cool.
20:22 cfedde the bevel on the grill of the other model is not near as shiny.
20:22 Grinnz_ make a mojo one to go with my twitter one ;)
20:23 sri now i'm really curious if there are more Link header users out there
20:24 cfedde Hire Gallup to do a survey for you.
20:25 cfedde otoh rfc5988 does not seem that bad.  Either a custom recursive parser or use some other parser module.
20:25 Nei Link header is equivalent to <link> in html
20:26 Nei there are (probably) more html link users than http header link
20:26 preaction a relic from the bygone days where XML+XSL was going to take over the world
20:28 cfedde API::Github uses mojolicious...
20:28 sri google doesn't find much, just a few references to browsers now allowing Link headers to be used for prefetching in addition to the <link> tag
20:30 Grinnz_ yeah that guy released a whole bunch of modules in the API:: namespace, havent looked into them
20:31 sri hehe, i find a lot more parsers than actual users of Link headers ;p
20:31 sri http://docs.python-requests.org/en/master/user/advanced/#link-headers
20:36 cfedde it'd be great if those rocket scientists over at github just put the clues in to the payload. but no.  they put it in a header.
20:40 disputin joined #mojo
20:59 sri bitcoin price getting close to $800
21:17 genio one of those things that in hindsight I wish I'd mined for bitcoins when it first started so that I could sell them all now
21:18 genio or, apple stock, or google stock, etc.
21:33 jberger genio: you and everyone else
22:09 cfedde that's the funny thing about my epedemic stock.  and my 24/7 stock and my mercury mail stock and the rest of the dotcom companies I had stock in.
22:21 Phil21 yeah well... I had a lot of 4 letter domain names back when they were free I let expire :)
22:29 punter joined #mojo
22:30 Phil21 oh, that an IP space
22:31 Phil21 wish I would have just been evil and soaked up a /16 when it was pretty trivial to convince someone to just kinda do it for you :)
22:33 good_news_everyon joined #mojo
22:33 good_news_everyon [mojo] kraih tagged v6.66 at ec285cb: https://git.io/vouXA
22:33 good_news_everyon left #mojo
22:35 good_news_everyon joined #mojo
22:35 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vou1T
22:35 good_news_everyon mojo/master b0f8f97 Sebastian Riedel: bump version
22:35 good_news_everyon left #mojo
22:56 dmanto left #mojo
23:33 jnbek hehe, I got creative with the AUR Pkg :D
23:54 dave won't quantum computing render bitcoins fairly worthless?
23:58 sri yea, better send me all your bitcoins :)
23:59 dave I don't have any ><

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