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

IRC log for #mojo, 2017-11-12

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

All times shown according to UTC.

Time Nick Message
00:00 sri btw. another little preview https://i.imgur.com/2l6pVFP.png
00:00 sri the design is far from final
00:01 sri but that's what i have in mind for browsing jobs
00:02 sri there will be no individual pages for jobs, just a table with paginated search results and you tap for details
00:02 sri then select the ones you want and there are buttons for retry and stuff at the bottom (maybe top too)
00:03 sri stats at the top are still realtime, on every page
00:03 purl Since Fri Nov 10 21:27:45 2017, there have been 11 modifications and 10 questions.  I have been awake for 1 day, 2 hours, 35 minutes, 24 seconds this session, and currently reference 990260 factoids. Addressing is in optional mode.
00:03 sri purl: shoo
00:03 purl sri: sorry...
00:03 nicomen I would love a cure CRUD UI for anything, including jobs… not sure creating/updating makes sense, but perhaps
00:03 nicomen *cute
00:04 sri creating maybe
00:04 sri updating is out of scope for now
00:05 sri i want only plain retry for now, changes to parameters will require the command line
00:07 nicomen nice, I liked your screenshot, and I would love a similar thing I could hook up to a DB, Redis etc
00:07 sri sorry, not what i'm building at all
00:07 sri this is minion specific
00:08 sri i know what you want, a django style admin app
00:09 sri and i agree, that would be useful
00:09 nicomen yes, I was hoping jobs somewhat mapped to a similar model ;)
00:09 nicomen but a minion ui is nevertheless a huge win
00:09 sri no, this is not basic crud i'm afraid
00:10 nicomen has your hackathon started?
00:10 sri friday was the first day
00:10 sri it will go all week
00:10 nicomen ah clever
00:10 Grinnz neat stuff
00:11 sri apparently somewhere there is a public holiday (us?), so it started on a friday instead of monday :)
00:11 Grinnz i'll be sure to bring the SQLite backend up to snuff once it's solidified
00:29 okcomputer joined #mojo
00:35 sri great
00:36 sri feels good to remove something from the backend api for once
00:36 sri don't really like that it has so many methods now
00:37 sri hard to avoid with so many features though
01:16 aborazmeh joined #mojo
02:12 kaare_ joined #mojo
02:20 mtj joined #mojo
02:55 ilbot2 joined #mojo
02:55 Topic for #mojo is now 🍩 nom nom | http://mojolicious.org | http://irclog.mojolicious.org | http://code-of-conduct.mojolicious.org
03:16 noganex_ joined #mojo
03:21 Kundun joined #mojo
05:04 dboehmer joined #mojo
05:35 dugword joined #mojo
07:03 geospeck joined #mojo
07:04 geospeck joined #mojo
07:05 geospeck joined #mojo
07:06 geospeck joined #mojo
07:07 geospeck joined #mojo
07:33 Vandal joined #mojo
07:52 geospeck joined #mojo
08:01 geospeck joined #mojo
08:28 mib_kzbmzc joined #mojo
08:45 Vandal joined #mojo
09:49 mib_4hlhbq joined #mojo
09:49 mib_4hlhbq left #mojo
09:52 mr_trousers joined #mojo
10:48 dod joined #mojo
10:54 dod joined #mojo
10:58 dod joined #mojo
12:08 Vandal joined #mojo
12:44 dotan_convos joined #mojo
13:18 ghenry joined #mojo
14:08 ghenry joined #mojo
14:55 dugword joined #mojo
15:18 mib_d9hkp7 joined #mojo
15:58 dod joined #mojo
16:09 mib_d9hkp7 left #mojo
16:10 Curly joined #mojo
16:13 Sebbe joined #mojo
16:15 anon joined #mojo
16:15 anon Hey! I need a help: how i can run mojo by fastcgi in openbsd httpd?
16:16 sri http://mojolicious.org/perldoc/Mojolicious/Guides/Cookbook#PSGI-Plack
16:19 anon thx, and but i need a websocket support?
16:20 sri not possible
16:20 purl not possible is it?
16:20 pink_mist then you can't use fastcgi
16:20 pink_mist use reverse proxy
16:23 anon hm... okay, thx
17:14 anon plackup has not work, http://dpaste.com/11ADWXB
17:22 jberger plackup is not a mojolicious server, if you need support you should ask them
17:22 jberger however, I'm assuming you need to install the module it recommends
17:23 sri there is a #plack channel
17:23 ChmEarl joined #mojo
17:23 sri very few mojolicious users actually deploy with plack, so you'll have more luck finding help there
17:36 anon ouch, FCGI is a different module, did not see
17:36 pink_mist does plack support websockets?
17:37 pink_mist anon: if you wanted websocket, use a reverse proxy to hypnotoad
17:38 anon but openbsd has not support reverse proxy
17:38 pink_mist really?
17:38 pink_mist they don't have mod_proxy?
17:39 pink_mist (they also need mod_ws_tunnel for websockets though)
17:39 pink_mist err
17:39 pink_mist mod_proxy_wstunnel
17:39 pink_mist I meant
17:40 anon http://man.openbsd.org/cgi-bin/man.cgi/OpenBSD-current/man8/httpd.8?query=httpd%26sec=8
17:40 anon i run plackup, yeah)
17:41 pink_mist oh, they don't use apache's httpd? well never mind then
17:43 sri pink_mist: no, plack does not support websockets
17:44 sri some psgi servers have a way to break out of the psgi sandbox to do websockets with the raw socket, but it's really ugly and doesn't work so well
17:44 mohawk is there anything fundamental about plack that means it couldn't support ws?
17:44 sri and we do not support that
17:44 mohawk (about psgi, i think i mean)
17:44 sri psgi/plack have no native support for websockets
17:45 sri just some web servers provide access to the raw socket
17:45 sri on which you have to build your own implementation
17:46 mohawk i wonder whether psgi/plack could be persuaded to abstract the raw socket bit out, and support WS
17:46 sri we'll prolly never support it
17:46 Grinnz mohawk: the psgi protocol is based around returning a subroutine reference that is run in a blocking manner, not conducive to websockets or other async
17:46 sri the psgi/plack folks appear to have mostly stopped developing the spec
17:46 mohawk Grinnz, right
17:47 sri there are open bugs even for non-blocking http support in psgi
17:47 mohawk yes, ws without async would be a bit challenging
17:47 * mohawk dreams of FSGI, with the F standing for future
17:48 geospeck joined #mojo
17:48 sri honestly, i think plack-ish abstraction layers have failed
17:48 sri all the big players in async just use custom servers
17:49 mohawk that's a shame
17:54 mohawk sri, sorry for the dumb questions (i'm still new to modern perl web serving) - what are the big players in async now?
17:58 pink_mist for perl? just Mojolicious really, at least as far as I'm aware
17:59 pink_mist I mean, dancer is trying I guess? but from what I've read about it, their attempt is pretty bad
17:59 pink_mist not that I've read much
18:00 mohawk dancer is trying async?
18:01 pink_mist unless it was Catalyst
18:03 sri mohawk: i mean other languages, for perl i honestly don't think we have real competition
18:03 sri i'm watching projects like node.js and phoenix
18:04 sri go can also be fairly interesting
18:05 pink_mist mohawk: https://metacpan.org/pod/distribution/Dancer2/lib/Dancer2/Manual.pod#Delayed-responses-(Async/Streaming) <-- I was right the first time, it was dancer
18:05 anon joined #mojo
18:05 sri hahaha, they still have a flush
18:06 mohawk sri, gotcha
18:06 mohawk pink_mist, thanks
18:07 sri that flush is total bullshit, and where you see whoever made this doesn't really grok the topic
18:07 mohawk quite
18:07 sri of course the abstraction for async in psgi is not great and you could get the idea something like flush actually might work
18:07 mohawk i think someone mentioned to me that D2 (and catalyst) are, because of psgi, fundamentally tied to being synchronous
18:07 mohawk possibly i got that wrong
18:08 * mohawk looks at http://p3rl.org/Catalyst::PSGI to try to wrap his head round it
18:09 sri there have been attempts to fix psgi at least for async http, and like 3 years i joined some discussions about it, but nothing happened since
18:10 sri think there were two big showstoppers, no standard way for the web server to signal client connection close, and handling of backpressure
18:11 mohawk backpressure?
18:11 sri what we handle with finish and drain events
18:11 sri application code generates more content than the server can write, filling up the write buffer
18:11 sri server running out of memory and so on
18:12 mohawk gotcha
18:12 sri we just say ->write('some chunk' => sub {... do more stuff once the chunk is written... })
18:13 sri since mojo is non-blocking to the core we have observables that stuff is written around
18:14 mohawk and that means it can be controlled by the server, so no risk of drowning it
18:14 sri yea
18:14 sri you can actually know when the server write buffer is empty
18:14 mohawk and since psgi doesn't have a concept of events, it can't really do that
18:15 sri and that's what made the dancer flush so silly, like a blocking call could wait for an event loop to finish emptying its buffer asynchronously
18:16 mohawk reminds me of what i read about throw away vs rewrite: firefox has taken 5 years (and counting) to get towards multi process, while chrome did that from the start
18:17 sri doesn't throw away and rewrite go together?
18:18 mohawk in that sentence, rewrite = incremental
18:18 geospeck joined #mojo
18:18 mohawk sorry to be unclear
18:18 sri i'm a strong believer in incremental
18:18 mohawk me too
18:20 mohawk insert joke here about incremental vs excremental
18:30 jberger merged my url branch into Mojo::Chrome, getting close to a release
18:30 jberger error handling is still the wild west
18:31 jberger but I don't think that's going to change much for a while
18:57 brunoramos joined #mojo
18:59 good_news_everyon joined #mojo
18:59 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vFrmV
18:59 good_news_everyon mojo/master 61dae0f Sebastian Riedel: improve built-in templates with high resolution logos
18:59 good_news_everyon left #mojo
19:01 sri yea, that looks nicer :)
19:01 sri i lost the original image and had to redraw the logos though
19:02 sri so there are small differences
19:02 sri doubt many people will notice though
19:02 mohawk what did you draw with?
19:03 trone joined #mojo
19:03 sri affinity designer
19:03 mohawk cool
19:04 sri drawing is one of the few things i still do with commercial apps
19:05 mohawk i have actually had good experiences using GIMP
19:05 mohawk and i found it more scriptable (using Scheme) than photoshop (using god knows what, never really tried hard)
19:06 mohawk ah, supports JS these days
19:06 mohawk not a huge surprise
19:06 mohawk GIMP doesn't
19:06 mohawk "someone should" make that possible
19:07 mohawk (no factoid for purl with someone should -> well volunteered?)
19:08 trone joined #mojo
19:08 sri hmm, looks like the picture tag doesn't work correctly in firefox :(
19:16 mohawk sri, your usage of it doesn't match this? https://www.w3schools.com/tags/tag_picture.asp
19:17 mohawk you have <img src=... srcset=.../>
19:17 mohawk whereas the above seems to say you'd need <img src=.../><source srcset=.../> ?
19:18 mohawk (i say, having just learned what a <picture> element is, and is for)
19:19 sri that is the opposite of useful
19:20 mohawk um?
19:20 sri totally irrelevant
19:20 mohawk ok then
19:20 sri and a shitty site to learn from actually
19:20 sri https://developer.mozilla.org/en-US/docs/Learn/HTML/Multimedia_and_embedding/Responsive_images
19:22 sri it works perfectly fine in chrome and safari, while firefox shows the 2x image, but makes it take up extra space
19:26 mohawk https://html.spec.whatwg.org/multipage/embedded-content.html#the-img-element says "If the srcset attribute is present and has any image candidate strings using a width descriptor, the sizes attribute must also be present, and is a sizes attribute."
19:35 karjala_ joined #mojo
19:58 [0xAF] joined #mojo
20:01 jabberwok joined #mojo
20:12 jamesaxl joined #mojo
20:22 jamesaxl joined #mojo
20:24 good_news_everyon joined #mojo
20:24 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vFrG0
20:24 good_news_everyon mojo/master e5cee74 Sebastian Riedel: use the full length of the menu bar
20:24 good_news_everyon left #mojo
20:51 sri not sure i like the menu bar like this
20:52 sri i guess it would make sense to redo the whole menubar css
20:53 sri if anyone feels like doing it, all the code is here https://github.com/kraih/mojo/blob/master/lib/Mojolicious/resources/templates/mojo/menubar.html.ep
20:53 sri you can test it with a mojo generate app and browsing the docs at /perldoc
20:54 sri especially the overflow handling needs to be better
21:29 good_news_everyon joined #mojo
21:29 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vFrcH
21:29 good_news_everyon mojo/master db02801 Sebastian Riedel: more consistent formatting
21:29 good_news_everyon left #mojo
22:11 [1]mohawk joined #mojo
22:29 brunoramos joined #mojo
23:07 good_news_everyon joined #mojo
23:07 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vFrB4
23:07 good_news_everyon mojo/master 4eeb062 Sebastian Riedel: stringify default paths
23:07 good_news_everyon left #mojo
23:08 sri new menu bar is live on http://mojolicious.org btw.
23:26 jberger Seconded: w3 schools is terribad

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