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

IRC log for #mojo, 2017-05-26

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

All times shown according to UTC.

Time Nick Message
01:08 disputin joined #mojo
01:18 disputin joined #mojo
01:20 disputin joined #mojo
01:21 disputin joined #mojo
01:22 disputin joined #mojo
01:23 disputin joined #mojo
01:25 disputin joined #mojo
01:26 disputin joined #mojo
01:28 disputin joined #mojo
01:30 disputin joined #mojo
01:31 disputin joined #mojo
01:32 disputin joined #mojo
01:33 disputin joined #mojo
01:35 disputin joined #mojo
01:36 disputin joined #mojo
01:37 disputin joined #mojo
01:38 disputin joined #mojo
01:40 disputin joined #mojo
01:41 disputin joined #mojo
01:42 disputin joined #mojo
01:46 disputin joined #mojo
01:46 disputin joined #mojo
01:47 disputin joined #mojo
01:50 disputin joined #mojo
01:51 disputin joined #mojo
01:52 disputin joined #mojo
01:53 disputin joined #mojo
01:55 disputin joined #mojo
01:56 disputin joined #mojo
01:57 disputin joined #mojo
01:58 disputin joined #mojo
02:00 disputin joined #mojo
02:01 disputin joined #mojo
02:03 disputin joined #mojo
02:06 disputin joined #mojo
02:08 disputin joined #mojo
02:10 disputin joined #mojo
02:11 disputin joined #mojo
02:13 disputin joined #mojo
02:15 disputin joined #mojo
02:16 disputin joined #mojo
02:17 disputin joined #mojo
02:18 disputin joined #mojo
02:21 disputin joined #mojo
02:23 disputin joined #mojo
02:26 disputin joined #mojo
02:28 disputin joined #mojo
02:30 disputin joined #mojo
02:32 disputin joined #mojo
02:57 noganex_ joined #mojo
03:03 tchaves joined #mojo
03:42 tchaves joined #mojo
03:51 aborazmeh joined #mojo
03:55 punter joined #mojo
04:04 dboehmer_ joined #mojo
05:50 inokenty-w joined #mojo
06:50 AndrewIsh joined #mojo
06:52 Vandal joined #mojo
06:57 dod joined #mojo
07:02 dod joined #mojo
07:09 ladnaV joined #mojo
07:27 trone joined #mojo
07:41 karjala_ joined #mojo
07:49 Vandal joined #mojo
07:55 karjala_ joined #mojo
07:57 karjala_ There's no way that a Subprocess (with Mojo::IOLoop::Subporcess) can itself launch a subprocess, right? Because the subroutine that the subprocess executes (and immediately exits afterwards) is syncronous, am I correct here?
07:58 karjala_ or maybe it can#
07:58 karjala_ but it won't capture the return value, will it?
09:08 [0xAF] jberger: hi, I'm trying Mojo::ACME, I've setup everything and I'm on my way to register account in letsencrypt with the command "./mojo-acme.pl acme account register", but I get "Could not get nonce", any ideas?
09:09 [0xAF] same error w/ and w/o the -t option
09:13 PopeF joined #mojo
09:21 karjala_ subprocess inside other subprocess was not a good idea, because once parent subprocess finished, the child subprocess died with it in the middle of its function
09:22 rshadow joined #mojo
09:24 stryx` joined #mojo
09:48 kes joined #mojo
09:49 kes Hi, is this OK to use validators from plugin?
10:00 gregf_ joined #mojo
10:33 [0xAF] jberger:
10:33 purl jberger: are the tests failing into infinite loop ever acceptable?
10:34 [0xAF] jberger: I've changed the line here https://github.com/jberger/Mojo-ACME/blob/master/lib/Mojo/ACME.pm#L109 with : "die "Could not get nonce\n".$tx->res->error->{message} unless @$nonces;"   to get more info on the error
10:35 [0xAF] jberger: in my case it was "IO::Socket::SSL 1.94+" is required for TLS support
10:38 punter joined #mojo
10:40 rshadow joined #mojo
10:48 punter joined #mojo
11:00 tchaves joined #mojo
11:29 jberger [0xAF] yeah, I've been a little fast and loose with error checking in that code
11:30 jberger The new result method would help though too
11:31 jberger Can you open a bug or a PR for me so that I don't forget?
11:34 rshadow joined #mojo
11:50 [0xAF] jberger: what's the new result method?
11:50 [0xAF] jberger: you want the PR for this line or for something else?
11:53 gizmomathboy joined #mojo
11:54 stryx` joined #mojo
11:54 [0xAF] jberger: btw, i'm still fighting something. The test server generates the cert file, but the real server gives me errors and creates no cert. I can see the real server, which makes 3 GET request but the result from the plugin command is error report. see here: http://pb.0xaf.org/?066fa2f841b0c59b#BjM24ueaJPMJUh7t+UshOfcAkeeH5u7cEYNjmQdWVSA=
12:10 Petru joined #mojo
12:19 dod joined #mojo
12:25 jberger Did you get an account key for the production server?
12:25 [0xAF] yes
12:25 jberger Oh it makes the requests .... hmmmm
12:25 [0xAF] the strange thing is that the staging (test) server is working
12:26 [0xAF] though the test server does not make requests to mojo app
12:27 [0xAF] i cant remember for sure, but i believe it was making requests, then it stopped (perhaps it cached something?!?)
12:27 jberger Yeah it doesn't always re validate
12:28 jberger Which was annoying during testing
12:28 [0xAF] i guess so
12:28 jberger Perhaps test against the test server and use a new subdomain
12:29 jberger That should force a real request
12:29 [0xAF] but the test server is working, i do not need any more certs from the test server...
12:29 [0xAF] the real one is not working and i cannot see the reason :(
12:30 jberger I'm not sure what to tell you :(
12:30 jberger It looks like the plugin is responding correctly
12:31 jberger Maybe LE is having some troubles today?
12:31 [0xAF] btw another thing is that the real server makes only 3 request from (~10 sudomains)... i guess it gives up because it sees errors and does not go for the rest subdomains
12:31 [0xAF] i shall try it later then
12:32 [0xAF] jberger: thanks for the help, i will try it later and will report to you of the results.
12:34 jberger Good luck!
12:34 purl You'll need it.
12:36 [0xAF] thanks ;)
12:50 vicash left #mojo
12:52 [0xAF] jberger: for some reason when i changed the order of the subdomains it worked... who knows... or LE fixed their problem if they had one...
12:59 jberger Hey, good deal!
13:00 [0xAF] jberger: thank you very much for the help and the great plugin, helps a lot. thanks man.
13:11 irqq joined #mojo
13:12 webart left #mojo
13:18 Petru joined #mojo
13:27 Armen good <time of day>!
13:30 punter joined #mojo
13:30 ashimema joined #mojo
13:59 jberger and good $timezone_corrected_day_phase to you too!
13:59 zen :D
14:00 nic If anyone can tell us how Saturday is shaping up, that would be useful.  I've heard it's going to be crazy
14:01 zen life is what you make it
14:01 pink_mist how very zen of you :P
14:03 jberger nic: saturday in the American midwest looks to be ok (other than, you know, the constant dumpster fire that is the US)
14:04 jberger but graded on a curve, looks nice
14:05 punter joined #mojo
14:20 Pyritic joined #mojo
14:39 punter joined #mojo
15:14 Janos joined #mojo
15:15 stryx` joined #mojo
15:30 michael hi, is the plain content (not form/json) in POST transactions (e.g. Mojo::UserAgent) expecting Perl character strings or bytes? Can't see anything in t/ or POD to give an example
15:34 jberger looking in the code, it should be bytes
15:35 jberger I'm still looking for if it is properly documented, not seeing it yet (as you say)
15:35 Grinnz yeah, i don't think that parameter is documented well in general, only a couple examples for it
15:43 michael thanks, I thought that too from code, but wasn't 100% sure
15:52 maschine joined #mojo
16:06 Petru joined #mojo
16:16 bwf joined #mojo
16:19 disputin joined #mojo
16:24 bwf joined #mojo
16:31 Janos joined #mojo
16:39 bwf joined #mojo
16:39 rshadow joined #mojo
16:54 stryx` joined #mojo
16:57 Armen I'm still looking at better ways to handle my models in mojolicious. I've moved my project to a public repo on github, so I can more easily share what I'm working on. I'm trying to define database relationships in my model by doing something like belongs_to(Other::Model) and has_one(Other::Model), but the way I got it to work looks rather hacky: __PACKAGE__->SUPER->belongs_to(Other::Model). I wonder if
16:57 Armen anyone has an idea how to improve this
16:57 Armen relevant file: https://github.com/Armen138/pastebeest.pm/blob/master/app/Pastebeest/Model.pm and https://github.com/Armen138/pastebeest.pm/blob/master/app/Pastebeest/Model/Paste.pm
16:59 dod joined #mojo
17:00 Armen apologies if this is more of a perl-help kind of question
17:17 sh14 joined #mojo
17:35 dod joined #mojo
17:35 bwf joined #mojo
17:42 jnap not 100% sure this is correct place to ask, but the Mojar dist pointed to #mojo.  Was looking at CPAN modules related to google Analytics and trying o figure out Mojar::Google::Analytics versus Net::Google::Analytics
17:43 jnap and trying to figure out what the Mojar version is doing better... Its slightly newer but seems to cover most of the same ground.  any takers?
17:48 jnap ok sorry its there, 'Net::Google::Analytics is similar, main differences being dependencies and means of getting tokens.'  So its really just about using the Mojo approved networking stuff.
17:49 Grinnz and other stuff, yeah, probbly
17:50 Grinnz looks like it uses RSA rather than standard oauth2
17:55 jnap well, the Net::Google::Analytics author probably would love a patch to support RSA.
17:56 Grinnz no idea on that, but not a fan of the JSON.pm and LWP dependencies
17:58 jnap LWP was good enough to have its own book though :)  not important I guess
17:58 Armen anyone can have a book! :P
17:58 Grinnz I don't really put much worth into how historically significant a module is
17:59 Grinnz just like I think led zeppelin is overrated :P
17:59 Armen now hold on just a second there
18:00 jnap they have one or two songs that are ok if you are stoned enough
18:00 Grinnz heh
18:00 Armen you guys are barbarians
18:02 Grinnz for more concrete reasons, I prefer HTTP::Tiny over LWP in 95% of cases because it's cleaner, has better defaults w.r.t. timeout and proxies, is core in 5.14+, and doesn't have really weird and ugly backcompat code involving Net::SSLeay
18:03 Grinnz the other 5% are things that LWP does that HTTP::Tiny doesn't, but I usually end up using Mojo::UA for that weird stuff anyway
18:03 jnap HTTP::Tiny++
18:04 Grinnz oh also, the widespread use of LWP::Simple where people should really be checking for HTTP errors
18:04 Grinnz that grinds my gears a little
18:05 jnap too bad we can't have a core module HTTP::Total that covers the rest, taking the best ideas from Mojo, LWP etc.
18:05 Grinnz well there are modules building on HTTP::Tiny already, that could probably be continued
18:05 Grinnz like HTTP::Tiny::Multipart
18:06 jnap I don't think thats i core though... I've used HTTP::Thin as well
18:06 Grinnz right, they arent core
18:08 jnap the real problem I had was when doing https://github.com/jjn1056/WebService-Auth0 wanting to be able to support non blocking if thats' there.  That was supposed to be my work related excuse to hack on Mojolicious
18:08 Grinnz yeah, that's when i jump to Mojo::UA
18:16 Grinnz er, when i said Net::SSLeay earlier I meant Crypt::SSLeay, which is a broken OpenSSL binding only used by LWP when all of your modules are out of date
18:17 Grinnz or if you 'use Net::SSL;' which is a terrible idea
18:43 jnap yeah no doubt its confusing to use SSL if you are not aware of what is going on :(
18:50 PryMar56 joined #mojo
18:58 pink_mist jnap: that's not what he's talking about - using the module named specifically "Net::SSL" is a terrible idea.
19:00 rshadow joined #mojo
19:20 rshadow joined #mojo
19:21 rshadow joined #mojo
19:30 rshadow joined #mojo
19:41 jnap glad AFAIK I don't then :)
19:42 rshadow joined #mojo
19:43 rshadow joined #mojo
19:51 trone joined #mojo
20:04 nic jnap:   o/
20:04 nic I can add some detail on the GA question
20:05 jnap it only came to mind to me since CPAN tends to have 10 of everything and it would be great to have more help to figure out which one, other than 'the one that seems maintained' :)
20:07 nic Yeah, agreed.  The original motivators for me were (a) be leaner on dependencies, (b) have more rigorous tests that it does what I needed for reporting
20:08 nic The Mojar v happens to be more efficient for inserting into a MySQL db, which is a major use-case for me
20:08 nic but...
20:09 nic In July there will be a major version bump.  There'll be more thought into making even more convenient to use and it'll be focused on GA v4 (rather than v3 currently)
20:10 nic (I haven't looked at the Net version in years, so not sure what version it targets)
20:27 BigDoubleZ joined #mojo
20:29 kaare joined #mojo
20:33 Trumpet joined #mojo
20:35 new_student joined #mojo
20:42 sri coolo: i don't get it, he's right
20:42 sri ;p
20:51 genio heh. we're working on fixing up LWP.
20:51 genio Lots of room for improvement, to be sure
20:53 Grinnz genio++ # and its a good effort so far
21:16 sri jnap: CGI.pm has a book too
21:18 * sri has spent quite some time reading the LWP internals, and things are bad, really bad in some places
21:24 jnap sri: CGI, back when I could install Perl and have everything I needed to work...
21:24 * sri would have fixed more stuff in LWP had he not had such a bad experience contributing to the project
21:27 sri finding and fixing a bug in HTTP::Headers that could be used to kill any catalyst app out there and not getting credit did sting a little
21:29 jnap sri: sorry :(
21:31 jnap its just saddens me that we all can't seem to get it together, considering the state of Perl in the wider world.  wish there was more I could usefully do
21:32 sri i mean, it's not much different in other programming communities
21:32 preaction it's a lot of effort, and social work is not prioritized (derisively called "soft skills")
21:32 sri btw. this was my bug https://github.com/libwww-perl/HTTP-Message/commit/5f70571af4f1d69aa91d0baa5478bca8e67ab80c
21:33 sri until 2012 you could make any catalyst app go out of memory with a carefully crafted requests
21:33 sri *+few
21:34 jnap yeah I can see that, in a persistent app
21:34 jnap tons of core Perl bits were written when persistent apps were not the norm
22:02 sri i do wish we had more well-run (big) perl projects
22:03 sri like, it is sad to see DBIx::Class and the surrounding projects stagnate
22:12 sri anyway... i don't care very much for melancholy at the moment
22:13 sri so, i kinda want to make the server (daemon/prefork) available to the mojo app
22:13 sri i know, i said we don't want to do that
22:13 sri but i changed my mind
22:14 sri there are reasonable use cases, like using hooks to extend the server
22:15 sri similar to how you use hooks to extend minion workers
22:17 sri how to do it... i'm not so sure
22:17 sri there's special cases, like hypnotoad, psgi and cgi
22:19 jnap not sure your thinking on it, but after playing with Go and having access to a low level http server in my framework has been fun.  I always shied from that in Catalyst because PSGI
22:20 sri i've been telling you it's great fir a few years now i think
22:20 genio sri: Contributing should be MUCH better now.
22:20 sri s/fir/for/
22:20 genio we'd be happy to have your help cleaning things up
22:21 genio The only problem is with us being scared to change some things for fear of breaking the many things that depend on LWP
22:34 jnap genio: thank part I get, which is why it might not be terrible to have something like HTTP::LessTiny in core.  just my opinion.  People seem to dig Mojo:UA
22:36 jnap sri: the difference is that Go comes with a production grad, event aware server in core.
22:36 pink_mist jnap: nah, HTTP::Tiny is all you need in order to download something more comprehensive :P
22:36 Grinnz pink_mist: not if you want to do it with https
22:37 pink_mist that's a fair point
22:37 sri genio: hope it works out
22:37 jnap I'd like a more complete core, but that's just my pref.
22:38 sri i'd prefer cleaned up and consistent versions of all modules perl currently ships
22:38 pink_mist jnap: most of the people I've talked to seem to want the opposite - only the very minimal set of things you need in order to get more stuff from CPAN ... I've not got many opinions on that myself though
22:39 sri pink_mist: i think so too
22:39 Grinnz the more-minimal-core vision fits better with distros trying to slim down to the smallest core perl they can package for their cloud images
22:39 jnap I used to feel like that, because the idea was you could show more agility and not be tied to a releas
22:39 Grinnz docker docker webscale etc
22:40 jnap well that's a packaging thing, I build minimal Perl + only my application dependencies as part of deployment.
22:40 sri the perl core should come with stuff like FFI instead of web servers
22:40 jnap reasonable point
22:40 Grinnz mmm
22:42 sri i'm still sad that there's no good perl bindings for tensorflow and nghttp2 yet
22:43 sri getting good bindings is such a slow process with the focus being on XS
22:53 genio XS is difficult and providing well done access to external C libraries is hard.
22:54 genio but, there are several of us now working on the LWP side of things, just trying to clean up as much as possible without breaking too much. There's lots of really, really old stuff in there
22:55 genio Hell, just cleaning up HTTP::Cookies is a massive feat and that's a relatively small dist
22:55 Grinnz genio: which is why making FFI easier to access would be nice, as it's supposed to be much simpler for binding external libs
22:57 genio right, only having a handfull of devs who are good enough with Perl, C, and XS to do those is not a great state to be in
23:26 jnap genio: in theory there's funding from TPF

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