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

IRC log for #mojo, 2016-08-29

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

All times shown according to UTC.

Time Nick Message
00:16 mtj joined #mojo
00:19 meshl joined #mojo
00:23 mtj joined #mojo
00:34 * sri adds Grinnz on battlenet and expects to be carried to higher ratings
00:53 Pierre joined #mojo
02:42 noganex joined #mojo
04:26 rnkn joined #mojo
04:47 Pierre joined #mojo
05:05 Pierre joined #mojo
05:13 inokenty-w joined #mojo
05:22 romel joined #mojo
05:24 Pierre joined #mojo
05:38 che-quest joined #mojo
05:42 Pierre joined #mojo
06:00 Pierre joined #mojo
06:06 Pierre joined #mojo
06:21 Lee joined #mojo
06:40 dod joined #mojo
06:46 dod joined #mojo
06:55 irctc582 joined #mojo
06:59 mbudde joined #mojo
07:03 salva joined #mojo
07:17 cpan_mojo Mojo-IRC-0.35 by JHTHORSEN https://metacpan.org/release/JHTHORSEN/Mojo-IRC-0.35
07:27 Jonis oo
07:27 Jonis 0.35
07:44 salva joined #mojo
07:47 dux joined #mojo
07:48 Vandal joined #mojo
08:45 mvorg joined #mojo
09:03 cstamas joined #mojo
09:05 pink_mist https://twitter.com/bbotezatu/status/768756952528723969 lol
09:12 [0xAF] joined #mojo
09:14 [0xAF] jberger: hi, i've just read the irclog and way your suggestion on '&verify=0x00', thanks, i will try it later, to confirm if it works. meanwhile i've made a setup with nginx as you suggested and found the way to pass the certificate information to mojo with this setting 'proxy_set_header X-SSL-User $ssl_client_s_dn;' in nginx.conf. thanks again for your help.
09:14 [0xAF] s/and way your/and saw your/
09:17 cuechan joined #mojo
10:27 kaare joined #mojo
10:42 salva joined #mojo
11:06 tchaves joined #mojo
11:13 tchaves joined #mojo
11:34 cafe joined #mojo
11:37 Pierre joined #mojo
12:28 jberger Nice
12:28 jberger I almost like the nginx way better ;p
12:49 gizmomathboy joined #mojo
12:55 sri guess i'll release 7.04 later today
13:03 Averna joined #mojo
13:03 lluad joined #mojo
13:04 jacoby joined #mojo
13:08 cuechan_ joined #mojo
13:17 marty joined #mojo
13:20 dantti_laptop joined #mojo
13:37 dod joined #mojo
13:39 sri some interesting talks to watch in the videos from yapc::eu
13:49 cpan_mojo Mojolicious-Plugin-AssetPack-1.21 by JHTHORSEN https://metacpan.org/release/JHTHORSEN/Mojolicious-Plugin-AssetPack-1.21
13:59 good_news_everyon joined #mojo
13:59 good_news_everyon [mojo] kraih tagged v7.04 at 3ae7e70: https://git.io/vieU3
13:59 good_news_everyon left #mojo
14:00 good_news_everyon joined #mojo
14:00 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vieU4
14:00 good_news_everyon mojo/master 153a606 Sebastian Riedel: bump version
14:00 good_news_everyon left #mojo
14:12 marty joined #mojo
14:16 CW joined #mojo
14:17 jberger thanks for the release!
14:17 jberger and yeah, I intend to watch some videos too, I just hope they split them out
14:18 jberger the continuous stream thing is tough
14:18 jberger mst's lightning talk was funny because he seriously misjudged his amount of time :-P
14:30 dod joined #mojo
14:46 mcsnolte joined #mojo
14:49 anton joined #mojo
14:52 jacoby joined #mojo
14:55 jberger srI: did 7.04 index?
14:55 kes joined #mojo
14:55 sri yes
14:55 jberger oh, nm
14:55 jberger the "fast sync" mirror is behind :s
14:55 jberger preaction: poke
14:55 sri heh
14:56 tchaves1 joined #mojo
14:59 sri subprocess performance on os x seems pretty decent actually
14:59 cuechan_ joined #mojo
14:59 sri perl -Mojo -E 'a(sub { my $c = shift; Mojo::IOLoop->subprocess(sub { "Hello World!" }, sub { $c->render(data => pop) }) })->start' daemon -m production -l http://*:8080
14:59 sri at a concurrency of 100 i can serve about 150 rps
15:01 jberger i've been thinking about the initial next_tick and I think it is best to leave it as you have it without that
15:01 jberger since you can always wrap it in a next_tick if that's what you want
15:01 jberger but you can't do the opposite
15:01 sri ohoh, i did just find a problem though
15:02 sri looks like EV doesn't like fork and the subprocess randomly crashes every now and then
15:02 jberger I've never seen that before
15:02 jberger remember I yanked the original code from AnyEvent::Util
15:02 sri i'm seeing "Non-zero exit status (72057594037927935)"
15:02 jberger Oo
15:06 sri or wait a minute
15:06 tchaves joined #mojo
15:06 sri i think ev might be stealing our exit status
15:09 sri yea, that's it
15:10 zivester joined #mojo
15:11 good_news_everyon joined #mojo
15:11 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vieGd
15:11 good_news_everyon mojo/master c7a7b66 Sebastian Riedel: EV might steal our exit status
15:11 good_news_everyon left #mojo
15:12 good_news_everyon joined #mojo
15:12 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vieZJ
15:12 good_news_everyon mojo/master 255b108 Sebastian Riedel: just check the process id
15:12 good_news_everyon left #mojo
15:13 sri jberger: imagine you were not having the problem since you were not checking the exit status
15:14 kaare joined #mojo
15:15 sri "More precisely: when the process receives a SIGCHLD, EV will fetch the outstanding exit/wait status for all changed/zombie children and call the callback."
15:16 sri yea, EV just goes around stealing exit statuses
15:22 good_news_everyon joined #mojo
15:22 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vienK
15:22 good_news_everyon mojo/master c9816ac Sebastian Riedel: just use a variable
15:22 good_news_everyon left #mojo
15:24 sri question is we actually want to check the exit status
15:25 sri considering it's unreliable
15:25 sri it will only work half of the time
15:25 jberger oh, I do remember that
15:25 jberger :s
15:26 jberger IIRC it was a bigger deal for ReadWriteFork
15:27 jberger isn't there some way to disable the EV child watcher
15:27 * jberger digs
15:27 sri don't think so
15:29 sri this is basically the error you would get instead if i remove the exit status check https://gist.github.com/anonymous/c578334558487b541da52276222e8a67
15:29 sri and you would get 50% of the time if you have EV loaded
15:30 jberger many I really dislike how schmorp messes with other people's code and yet throws a fit when other people mess with his
15:30 jberger s/many/man,/
15:30 sri +1
15:30 purl 1
15:31 sri guess consistency might win here :S
15:32 jberger http://lists.schmorp.de/pipermail/libev/2008q1/000098.html
15:35 jberger digging through the EV.xs source there are build flags to disable it
15:35 jberger but that's not something we can rely on
15:38 sri yea, removing it
15:38 good_news_everyon joined #mojo
15:38 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vieWR
15:38 good_news_everyon mojo/master 2055ae7 Sebastian Riedel: with modules like EV the exit status is just not reliable enough
15:38 good_news_everyon left #mojo
15:39 Grinnz sri, if you're not catching exit status, you should probably trap serialize exceptions
15:41 sri what do you mean?
15:41 Grinnz if you return a code ref in the subprocess, the parent won't get an error
15:42 sri yes it will
15:42 sri not a good one, but an error
15:43 sri "Magic number checking on storable string failed..."
15:44 sri and you should see a message on STDERR from the forked process actually
15:47 jberger not serialize but deserialize
15:47 Grinnz no serialize is what i meant
15:47 jberger why?
15:47 Grinnz sri, you mean because storable will try to deserialize nothing?
15:47 sri yes
15:48 Grinnz you might not have the same error with other deserializers
15:48 jberger an empty document, no matter what type of serializer is always invalid
15:48 sri not my problem ;p
15:48 jberger because [] is the smallest message it can try to send (no matter what serializer)
15:49 jberger if there were some serializer who's default serialization of nothing and an emtpy array were the same, then it would fall through the cracks
15:49 jberger I don't see that happening
15:51 kaare joined #mojo
16:14 Grinnz_ jberger: or if it's undef
16:15 Grinnz_ but sure, probably not a problem other than confusing error messages
16:19 jberger if what is undef?
16:23 Grinnz_ the default serialization
16:24 jberger well, that's kinda what I meant
16:25 disputin joined #mojo
16:42 good_news_everyon joined #mojo
16:42 good_news_everyon [mojo] kraih tagged v7.05 at 0515528: https://git.io/vie2X
16:42 good_news_everyon left #mojo
16:43 pink_mist wow, that was a quick version bump :P
16:44 sri like in old times :)
16:44 good_news_everyon joined #mojo
16:44 good_news_everyon [mojo] kraih pushed 1 new commit to master: https://git.io/vie2x
16:44 good_news_everyon mojo/master c3694c7 Sebastian Riedel: bump version
16:44 good_news_everyon left #mojo
16:48 mpapec joined #mojo
16:52 mpapec joined #mojo
16:59 mpapec joined #mojo
17:11 mpapec joined #mojo
17:15 meshl joined #mojo
17:22 mpapec joined #mojo
17:26 tianon joined #mojo
17:31 CW joined #mojo
17:37 dod joined #mojo
17:49 mpapec joined #mojo
17:59 meshl joined #mojo
18:03 PryMar56 joined #mojo
18:13 orev joined #mojo
18:15 gizmomathboy joined #mojo
18:38 gizmomathboy joined #mojo
19:18 sri somethign came up on the mailing-list
19:18 sri someone was asking for a subprocess limit
19:19 sri i think forkcall in AnyEvent actually has that, once you reach the limit it puts future forkcalls into a fifo queue to be processed later
19:19 jberger yes, it has that
19:19 jberger I never did that
19:19 jberger since that really has to be a global queue to make valuable
19:20 jberger suppose you could store that in the singleton if you wanted
19:20 sri no, in the Mojo::IOLoop instance
19:20 sri a per ioloop limit
19:21 jberger oh, sure, sorry
19:21 jberger a level higher than the subprocess though
19:21 jberger and forkcall didn't implement any higher level than that that it owned
19:22 sri mlehmann likes globals
19:22 jberger I could have put it in a package global like AE ... and you just said that
19:22 * sri doesn't
19:23 jberger in this case (as with most singletons) it almost makes sense since the thing you are doing it limiting the global resource of memory
19:42 mishanti1 Anyone else here using Mojolicious::Plugin::AccessLog and want to log outermost client ip from the XFF-header? If so I created a (somewhat crude) pull-request for that.
19:43 mishanti1 https://github.com/augensalat/mojolicious-plugin-accesslog/pull/11 for anyone that's interested.
19:51 Grinnz i think the biggest challenge for subprocess limit for this feature is that unlike other applications, requests can't wait for a connection to free up
19:51 Grinnz er, s/connection/process/
19:51 Grinnz so you'd have to decide, would new requests that need a subprocess just error? timeout?
20:18 batman joined #mojo
20:29 tchaves joined #mojo
20:47 Janos joined #mojo
21:08 romel joined #mojo
21:38 mvorg joined #mojo
21:42 margeas joined #mojo
21:46 cpan_mojo Mojolicious-Plugin-ClosedRedirect-0.09 by AKRON https://metacpan.org/release/AKRON/Mojolicious-Plugin-ClosedRedirect-0.09
21:58 doc joined #mojo
22:06 doc joined #mojo
22:06 doc hello everyone, I posted a simple question to perlmonks: http://www.perlmonks.org/?node_id=1170750
22:10 Grinnz_ ew, perlmonks
22:15 jberger doc dropped already?
22:15 jberger grrr
22:55 sri yay, subroutine signatures are now fast http://perl5.git.perl.org/perl.git/blob/HEAD:/pod/perl5254delta.pod
22:58 Grinnz_ \o/
23:01 pink_mist awesome =)
23:08 jberger Sweet
23:09 jberger I would still love a trailing coderef option
23:09 jberger But i don't that would fly with the powers that be
23:09 jberger Doubt even
23:10 pink_mist unsure what you mean by that?
23:11 Grinnz_ some way to parse variable args with the last one being a coderef, like a lot of async functions do
23:12 pink_mist ah
23:13 jberger Grinnz_++
23:13 Grinnz_ like ($self, @args, $cb)
23:13 jberger http://irclog.perlgeek.de/mojo/2016-06-06#i_12614011
23:13 Grinnz_ but it kinda doesnt work with how signatures are supposed to directly map to perl's list assignment
23:14 sri i remember an old proposal for ruby-ish blocks in perl
23:15 sri $ua->get('mojolicious.org') { my ($ua, $tx) = @_; ... }
23:15 Grinnz_ heh
23:15 sri wonder what that would have looked like with signatures
23:16 sri $ua->get('mojolicious.org') ($ua, $tx) { ... }
23:16 sri that would have been kinda cute
23:16 * genio is not a fan
23:16 Grinnz_ poor perl parser :P
23:16 haarg https://metacpan.org/pod/PerlX::MethodCallWithBlock
23:16 Grinnz_ neat
23:17 Grinnz_ except for its depencencies, but the idea :P
23:17 Grinnz_ probably would be tricky to do in a better way, since its not really a keyword of any sort
23:18 jberger purl parser?
23:18 purl rumour has it parser is somehow telling us "need more" or "got enough"
23:18 sri need more beer
23:19 jberger sri: I'm pretty sure Germany has more ;p
23:27 fitnerd joined #mojo

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