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

IRC log for #mojo, 2017-02-23

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

All times shown according to UTC.

Time Nick Message
00:18 PopeFelix Does the proxy in t/mojo/websocket_proxy.t *only* work for websocket requests?
00:18 PopeFelix Or rather does it only proxy those requests?
00:34 PopeFelix no, the proxy is only being *called* for those requests. If you look at https://metacpan.org/source/SRI/Mojolicious-7.26/t/mojo/websocket_proxy.t#L142, I don't think the proxy is being called for that request
00:34 purl okay, PopeFelix.
00:35 PopeFelix Sorry to spam the channel like this; I'm just trying to understand what's going on there.
00:38 jberger this is your periodic reminder that you should have git push.default set to simple in your .gitconfig
00:38 jberger https://github.com/jberger/dotfiles/blob/master/gitconfig#L5
00:39 PopeFelix Why is the CONNECT only being triggered in the web socket query?
00:42 jberger PopeFelix: I can't look at the moment, but a websocket handshake **is** a GET request
00:42 PopeFelix jberger, OK, that makes sense, but why aren't the regular GET requests triggering the proxy?
00:43 jberger as I said, I'm actually heading out just exactly now
00:43 jberger sorry
00:43 PopeFelix That's OK.
00:43 PopeFelix Thanks all the same
00:43 * PopeFelix is just confused and it makes him grumpy
00:47 lluad joined #mojo
02:11 aborazmeh joined #mojo
02:19 Crisewng joined #mojo
03:20 Crisewng left #mojo
03:26 noganex_ joined #mojo
03:39 ashimema joined #mojo
03:59 marty joined #mojo
05:02 mattp_ joined #mojo
05:04 dboehmer_ joined #mojo
06:14 hartenfels joined #mojo
06:36 hartenfels1 joined #mojo
07:01 polettix joined #mojo
07:32 Vandal joined #mojo
07:53 kaare joined #mojo
08:39 trone joined #mojo
09:09 crab joined #mojo
09:19 child035 joined #mojo
09:21 sugar joined #mojo
09:25 sri PopeFelix: there are different kinds of proxy servers
09:25 sri CONNECT proxy, HTTP proxy and SOCKS5 proxy are the ones we support
09:36 sh14 joined #mojo
09:47 sri hmm, unix domain socket support still has many many bugs :(
10:02 irqq joined #mojo
10:26 good_news_everyon joined #mojo
10:26 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vyeYb
10:26 good_news_everyon mojo/master 00625bc Sebastian Riedel: add support for WebSockets over UNIX domain sockets
10:26 good_news_everyon left #mojo
10:26 sri i started out looking into proxy support
10:27 sri while it is true that http+unix proxies kinda work, that's not the case for CONNECT and SOCKS
10:28 sri so yea
10:28 sri $ua->websocket('ws+unix://%2Ftmp%2Fmyapp.sock/echo.json' => sub {....});
10:28 sri that's a thing now
10:28 nic wow
10:30 nic I'd love to see the impact for cooperating real-time apps (grouped micro services)
10:32 nic securing partner apps from other ips should become trivial
10:55 sugar joined #mojo
11:00 sugar joined #mojo
11:08 good_news_everyon joined #mojo
11:08 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vyeZZ
11:08 good_news_everyon mojo/master bd8685e Sebastian Riedel: improve Mojo::UserAgent to complain about unsupported protocols
11:08 good_news_everyon left #mojo
11:08 sri since we support quite a few more protocols now... i suppose it's time to do that
11:09 sri so much code added this week :S
11:09 tchaves joined #mojo
11:16 sugar joined #mojo
11:51 good_news_everyon joined #mojo
11:51 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vyeWg
11:51 good_news_everyon mojo/master 55ab414 Sebastian Riedel: place protocol check a little more at the center of connection management
11:51 good_news_everyon left #mojo
12:05 good_news_everyon joined #mojo
12:05 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vyely
12:05 good_news_everyon mojo/master a48d0e6 Sebastian Riedel: expect failure
12:05 good_news_everyon left #mojo
12:14 stryx` joined #mojo
12:50 aborazmeh joined #mojo
12:53 sugar joined #mojo
12:54 good_news_everyon joined #mojo
12:54 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vyeEG
12:54 good_news_everyon mojo/master db337ec Sebastian Riedel: test the host_port method too
12:54 good_news_everyon left #mojo
13:19 aborazmeh joined #mojo
13:35 sugar joined #mojo
13:57 hartenfels joined #mojo
14:07 gryphon joined #mojo
14:11 itaipu joined #mojo
14:15 genio joined #mojo
14:44 ashimema joined #mojo
14:45 itaipu joined #mojo
14:50 khfeng joined #mojo
14:55 sugar joined #mojo
15:16 gregf_ joined #mojo
15:41 marty_ joined #mojo
15:51 sri i guess unix domain sockets have turned into my hackweek project after all :S
15:51 sri now i'm thinking about how to do proxies correctly
15:55 asarch joined #mojo
16:07 gizmomathboy joined #mojo
16:08 coolo sri: I have a text[] column and I do insert { keywords => [ sort @words ] }
16:09 coolo but I get Cannot bind unknown placeholder 4 (4) at /usr/lib/perl5/vendor_perl/5.24.0/Mojo/Pg/Database.pm line 97.
16:09 coolo what is wrong?
16:09 purl but it feels so right
16:10 spicyjack purl: botsnack
16:10 purl :)
16:11 * coolo assumes sql::abstract is not supporting it
16:13 coolo that's it
16:16 sri coolo: yea, would have been my first guess too
16:17 sri more and more feature requests for SQL::Abstract :S
16:19 jabberwok coolo: what does the expanded sql statement from SQL::Abstract look like?
16:24 jberger it'd be pretty hard to put array functionality into sqlabstract wouldn't it
16:24 jberger do any other rdbmses have native arrays?
16:27 jabberwok the other gotcha is: do you want that as native array or as a json-encoded thing?  in many cases the right sql depends on the table definition
16:29 sri jabberwok: DBD::Pg handles arrayrefs passed to it transparently
16:29 sri so, just pass the arrayref along verbatim
16:29 sri there is nothing to do
16:30 * coolo switched to crafted sql and it works - it just feels so January :)
16:31 sri the more i look into DBIx::Class the more grim things look... that project has been so mismanaged it's ridiculous
16:31 spicyjack the code looks grim, or the people situation is grim?
16:32 sri both
16:32 sri riba apparently managed to become the only person who could still navigate most of the code
16:33 sugar joined #mojo
16:34 sri so now there's a bunch of not very motivated folks trying to make sense of it, and everything is basically frozen in time
16:40 sri there was a dev! release planned for SQL::Abstract last friday... it still hasn't happened
16:41 sri a dev release
16:41 purl hmmm... a dev release is for pussies!
16:41 sri !
16:41 lluad joined #mojo
16:41 ribasushi sidestepping the DBIC part as offtopic, I can respond to where my SQLA fork is headed: while update returning will be included for historical reasons, it will be discouraged as a RBDMS-specific piece. SQLA::<undecided> will basically focus on the least common denominator, and on interface stability, with only efficiency/disambiguity gains in the future
16:42 ribasushi on where SQLA itself is headed - that'd be a question for mst
16:43 ribasushi to clarify "least common denominator": rdbms-specific pieces should be implemented as rdbms-specific subclasses ( like DBIC itself does for the most part )
16:44 ribasushi the update/delete returning syntax as done by sqla master is only supported this way only by Pg: both MSSQL and Oracle have a different incompatible way
16:44 ribasushi multi-group inserts are not widely supported either
16:45 sri how did SQL::Abstract end up not getting a release since 2014?
16:45 ribasushi for limits: there is this: https://metacpan.org/pod/distribution/DBIx-Class/lib/DBIx/Class/SQLMaker/LimitDialects.pm#SQL-LIMIT-DIALECTS
16:45 ribasushi sri: as Lee said elsewhere: there was no compelling reason
16:46 ribasushi sri: we generally have a *radically* different view on how to "do CPAN", so I am really not in a position to answer your question in a manner you'll find satisfactory/agreeable
16:50 jabberwok what about SQL::Abstract::More
16:53 jberger the idea of "least common denominator" is fine generally, but when you have the power of postgres I find it distinctly less than compelling
16:53 jberger even most of my CRUD work is more complex than could be generally abstractable
16:54 ribasushi sure thing, this is why having something like SQL::Abstract::Pg is likely the way to go for this audience
16:54 * ribasushi dislikes postgress for the record ;)
16:54 * jberger could have guessed :-P
16:55 jabberwok Postgres is MySQL on "acid"  pththtbbbt
16:57 mullagain joined #mojo
16:58 stryx` joined #mojo
17:10 rshadow joined #mojo
17:19 rshadow joined #mojo
17:28 sri never trust a man who doesn't like postgres
17:34 rshadow joined #mojo
17:34 stryx` joined #mojo
17:37 ribasushi sri: here, have me on record: https://youtu.be/5d0umrIFPR0?t=2724 ;)
17:38 genio get the pitchforks :)
17:45 dustinm joined #mojo
17:51 sri well, i suppose that's the problem with ORMs, they're all made by people who don't like databases
17:55 lluad like or understand, mostly.
17:56 lluad Though there are a few exceptions.
17:56 coolo sri: damn, I did it again - I called a function import and debugged for an hour why the class wouldn't load
17:56 sri coolo: i warned you ;p
17:56 stephan48 sounds way too familar
17:56 stephan48 i once did that back when i used catalyst
17:57 coolo sri: since today I'm stephan41 and I guess it shows :)
17:57 sri haha
17:57 stephan48 i think i managed do override some internal method leaving the whole thing in a broken state :D
17:57 stephan48 haha
17:57 sri coolo: happy birthday!
17:57 purl rumour has it happy birthday is for (('to you', 'dear '.shift)[0,0,1,0]) { print "Happy birthday $_" }
17:58 stephan48 coolo: happy birthday!
17:58 purl happy birthday is, like, for (('to you', 'dear '.shift)[0,0,1,0]) { print "Happy birthday $_" }
17:58 coolo sri: did I tell? SUSE gives a party! :)
17:59 coolo and thanks
17:59 stephan48 and btw. the number behind my nick is not my age (yet, by far) and totally meaningless :D
17:59 sri hope there's better beer this time
18:00 coolo sri: better beer? they put alcohol into it so it stays fresh - so much about Lebensmittelqualitaet
18:02 coolo sri: and read your email before monday - the mothership is calling
18:17 stryx` joined #mojo
18:18 sri coolo: oh, interesting
18:19 kes joined #mojo
18:22 Peppard joined #mojo
18:23 lluad joined #mojo
18:27 lluad joined #mojo
18:35 stryx` joined #mojo
18:44 lluad joined #mojo
18:58 good_news_everyon joined #mojo
18:58 good_news_everyon [mojo] kraih deleted ioloop_tls at b863ad4: https://git.io/vyvzO
18:58 good_news_everyon left #mojo
19:00 sri jberger, batman, marcus: btw. would be great to get your opinion on what we have so far https://github.com/kraih/mojo/compare/v7.26...master
19:00 jberger coolo: happy birthday!
19:00 purl happy birthday is, like, for (('to you', 'dear '.shift)[0,0,1,0]) { print "Happy birthday $_" }
19:00 jberger purl: you've said that already!
19:00 purl jberger: what?
19:00 mcsnolte joined #mojo
19:00 rshadow joined #mojo
19:00 lluad joined #mojo
19:00 sri specifically the use of the custom URL schemes, http+unix and ws+unix
19:01 sri if we want full support i might look into socks+unix tomorrow
19:05 jberger looks fine to me
19:05 jberger this is a little over my head since I don't use unix sockets for anything
19:05 jberger but from what I understand it might be useful to have the application running on a unix socket and reverse proxy to that from nginx
19:06 sri people who containerize a lot will like it
19:07 sri securing ports on shared servers can be a bit painful, with unix sockets you just got file permissions, and it's trivial
19:07 sri also a bit faster, since there's no protocol overhead
19:08 sri server side support is already feature complete btw.
19:09 sri client side is interesting for testing and ipc uses
19:10 sri at work we have a project (openQA) that uses dbus for ipc, that could just as well be websockets over unix sockets
19:13 irqq joined #mojo
19:18 Grinnz also nice to not have to deal with random port numbers, i think
19:19 sri wonder if IO::Socket::UNIX supports this linux thing where domain sockets don't require an actual file
19:20 sri where you start the name with \0
20:06 Dandre joined #mojo
20:09 khfeng joined #mojo
20:09 sugar joined #mojo
20:18 PopeFelix sri, thanks for explaining that. Now I know better where to look.
20:18 * PopeFelix shakes his head.
20:18 PopeFelix All this trouble for a bloody test case.
20:24 noganex joined #mojo
20:34 sri the expanse is soooo good
20:35 sri this episode... just wow
20:37 genio haven't watched in a while
20:37 genio I don't recall if I finished season one
20:38 PryMar56 joined #mojo
20:38 sri apparently this episode was the end of the first book
20:38 itaipu joined #mojo
20:39 sri season 2 has been phenomenal so far
20:40 sri i've not enjoyed scifi this much since firefly and bsg
20:43 jberger I've watched s01e01, am I living the high life?
20:43 sri it starts a bit slow
20:44 jberger I could use a good space opera
20:44 sri but a few episodes in it should get you hooked
20:44 jberger but I've been so consumed by podcasts etc lately
20:44 jberger and actually I just started an audiobook too
20:45 jberger with the news of the new Philip Pullman trilogy I decided it was finally time to read His Dark Materials
20:45 sri it's *the* space opera of the last 10 years imo
20:45 jberger that's high praise
20:46 sri i hope cyberpunk will make a comeback this year
20:47 jberger of course I reread ASOIAF last year on the assumption that Winds of Winter was coming out soon, so whenever that comes out I'll have to reread that again
20:47 jberger cause now its been too long again
20:47 jberger :-P
20:50 marty joined #mojo
20:50 Grinnz i only saw whatever syfy let me watch on their website free when i last checked
20:50 Grinnz i dont think it works anymore
20:51 orev joined #mojo
20:52 marty_ joined #mojo
21:00 stryx` joined #mojo
21:05 castaway joined #mojo
21:13 jberger Grinnz: amazon prime has it
21:14 jberger though that did annoy me because amazon stubbornly doesn't support chromecast
21:14 Grinnz nice. added
21:14 Grinnz that's part of the reason i got a fire TV stick :P
21:14 jberger you bowed to their pressure?!
21:14 Grinnz i also find the fire TV remote a little less annoying than using my phone
21:14 Grinnz well, it was that or keep using my xbox
21:14 jberger I like the idea of chromecast
21:15 PopeFelix I want one thing that plugs into an HDMI port and does amazon, chromecast, whatever.
21:15 jberger a little button that apps can add
21:15 PopeFelix Instead of all these different thingies.
21:15 Grinnz i use fire TV stick for most things, and i still have the chromecast for casting arbitrary things
21:15 jberger though I will say, it seems like chromecast app integrations have been flaky lately, like they seem to lose connection to the chromecast during playback
21:15 PopeFelix jberger, That happens to me too.
21:16 jberger so hitting pause sometimes can take a while while you fumble around to get it to reconnect
21:16 Grinnz the netflix app has gotten a lot better, but it still works better overall on the fire tv
21:16 * PopeFelix has a lot of stuff in Plex.
21:16 PopeFelix Can you do Plex over a Fire stick?
21:18 Grinnz it does have a plex app, but ive never used plex
21:18 PopeFelix I've got all of my DVD's encoded and stored on a little FreeNAS box.
22:07 PopeFelix Could I get some eyes on a pull request for Mojo::UserAgent::Mockable? https://github.com/PopeFelix/Mojo-UserAgent-Mockable/pull/10
22:27 Grinnz ooh, Iron Fist comes to netflix march 17
22:27 Grinnz just before Mass Effect Andromeda
22:31 itaipu joined #mojo
22:47 Dandre joined #mojo
22:54 sri luke cage kinda ruined netflix marvel shows for me
22:58 preaction why?
23:01 sri because it was sooo bad
23:02 sri and the problem with the netflix marvel shows is that they are all related and leading up to a defenders series
23:03 sri i really liked jessica jones and daredevil
23:05 preaction i've heard it said that Luke Cage, somewhat like Jessica Jones, has a more-specific target audience, and i can believe it. i liked it, but i could see where certain parts might not resonate deeply like they could
23:06 sri there was not a single thing i liked about luke cage
23:06 preaction Luke Cage as a character has always kind of been about the black experience in the US
23:07 preaction i liked the music, the first villain was played quite well (really made me hate him). i liked the barber guy (he was in The Wire, and was excellent there)
23:08 preaction Luke's journey was appealing, and the story about the community were endearing
23:08 pink_mist bubbles
23:09 preaction Cottonmouth was the villain i was thinking of, and "Pop" was the barber (who played Commissioner Burrell in The Wire)
23:10 Grinnz I'm still kidna miffed about what they did with Cottonmouth, not sure if spoilers are still in effect :P
23:10 elik joined #mojo
23:11 Grinnz I'm a sucker for good villains
23:11 Grinnz it's why i like batman so much
23:11 preaction yeah, the other villain got a bit too comic-book for my tastes
23:12 preaction Cottonmouth had a reality to him, and a grounded realistic motivation
23:12 preaction Diamondback kind of got silly by the end...
23:13 PopeFelix I've been watching DC's Legends of Tomorrow on Netflix. It's fun.
23:14 Grinnz me too. in step with the seasons of Flash and Arrow, though they aren't related that much after the first episode
23:17 Grinnz it's... a little awkward. requires a bit more suspension of disbelief than i'd like. but overall story is cool
23:18 PopeFelix Yes, I'll agree with you there. But it's a comic book.
23:30 pink_mist preaction: oh yeah, my bad ... bubbles did a role in 'agent carter'
23:30 preaction yeah, andre ... something... he's been in a couple good things since The Wire
23:31 Grinnz havent seen agent carter, but if it pops up on netflix i'd like to
23:35 sri preaction: purple man and kingpin were great villains, that washed up power ranger from luke cage was a total joke
23:35 preaction yeah. i wish they'd kept going with Cottonmouth...
23:36 sri guess that i don't enjoy that whole black culture thing doesn't help either
23:37 sri then again i didn't really like the hand either
23:39 sri PopeFelix: yea, legends is just cheesy fun
23:40 itaipu joined #mojo
23:44 sri not sure why exactly, but i've been really enjoying that lethal weapon show, also silly fun
23:49 dave Mass Effect Andromeda?
23:50 Grinnz yes, March 21
23:50 dave Really? EA?
23:50 Grinnz huh?
23:50 dave it's an EA game?
23:50 dave still?
23:51 Grinnz yes, EA has always published mass effect and dragon age games
23:51 dave after that last ending...I'm scared
23:51 dave er I thought the first dragon age was just bioware
23:51 Grinnz i still have liked all of them
23:51 * sri never got around to playing mass effect
23:51 dave I personally think bioware died when EA bought them
23:51 Grinnz Bioware is the developer, not like EA has much say in things
23:52 Grinnz eh. they make consistently interesting stories and fun combat systems, IMO
23:52 dave hm, I've known EA as a company since the 80s ... they always get their hands involved and screw the company up
23:52 dave or developer
23:52 purl i heard developer was Developer only developes under vi vim or pico
23:52 dave purl that made no sense
23:52 purl dave: what?
23:53 dave anyway, that's my opinion on EA FWIW
23:53 dave no offense intended
23:53 Grinnz I have no love for EA either, but it doesn't stop me from playing bioware games or the sims
23:53 Grinnz and yeah, the first dragon age was EA published as well
23:54 Grinnz they made a misstep with the lack of ending on ME3, but I feel they rectified it with the free ending update
23:55 dave lol .. I was so disgusted with the ending I stopped paying attention
23:57 dave well if I can ever stop playing the new master of orion or developing in javascript, I'll have to check out the new mass effect
23:57 Grinnz the teasers look neat so far. based on bioware's track record I'm looking forward to it, but we'll see
23:59 sri i wouldn't mind another fallout game
23:59 dave oh yeah, thanks for reminding me I still have fallout 4 to finish

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