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

IRC log for #mojo, 2014-09-01

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

All times shown according to UTC.

Time Nick Message
00:07 hesperaux joined #mojo
00:22 sri i guess this was the fix https://github.com/Perl/perl5/commit/ffdb8b167
00:40 sri haha, lets see if we have smart people in the list ;p https://groups.google.com/forum/#!topic/mojolicious/4lqjcY4vdD4
00:43 good_news_everyon joined #mojo
00:43 good_news_everyon [mojo] kraih created faster_templates (+1 new commit): http://git.io/kLMpWQ
00:43 good_news_everyon mojo/faster_templates 3b0a7ff Sebastian Riedel: optimize successive text lines in Mojo::Template
00:43 good_news_everyon left #mojo
00:55 laouji joined #mojo
01:00 ivan joined #mojo
01:12 ivan joined #mojo
01:26 klapperl joined #mojo
01:26 ivan joined #mojo
01:32 disputin joined #mojo
01:42 jb360 joined #mojo
01:58 jberger oh wow, that was only fixed in 5.20
02:04 jberger looking at the (perl) patch, I don't think there's much that can be done
02:05 sri i cant replicate the problem outside of Mojo::Template, so it might be possible to work around it
02:07 jberger is the problem the #line directive?
02:07 jberger if so, do we need it?
02:07 sri i don't know :)
02:07 sri yes, we certainly do need it
02:08 jberger hmmmm, yeah, I suppose
02:09 jb360 joined #mojo
02:11 jberger what do the line numbers end up being without the line directive?
02:11 jberger #line
02:11 sri one off
02:11 sri oh, you mean completelt
02:12 sri some garbage
02:12 jberger ok, that's no good
02:12 sri eval yada yada or so
02:12 sri with wrong line numbers
02:13 sri the last line thing is completely impossible without it i believe
02:13 sri you may not ever notice it, but that hack is a real miracle
02:18 sri MOJO_TEMPLATE_DEBUG=1 perl -Ilib -Mojo -E 'say app->build_controller->render_to_string(inline => qq{% for (1.. 3) {\ntest\n% } # loop})'
02:18 sri that comment at the end of the last line
02:18 sri to make it work i had to split up line 3 into two lines
02:19 sri and use a line directive for line 4 to keep errors correct and hide it
02:29 jberger ahhhhh
02:29 * jberger has dawning realization
02:31 noganex_ joined #mojo
02:34 Eke- joined #mojo
02:48 sri wow, roy fielding is writing some great criticism on http/2... but i have a feeling even he will be ignored
02:48 jberger link?
02:49 sri he's starting multiple threads on the http wg list
02:49 sri http://lists.w3.org/Archives/Public/ietf-http-wg/2014JulSep/1907.html
02:49 sri like that one
02:52 sri or this one http://lists.w3.org/Archives/Public/ietf-http-wg/2014JulSep/1927.html
02:55 sri funny thing is... the spec is in last call... so the super accelerated timeline the internet giants have been pushing for would be disrupted if any of the points he brought up don't get completely ignored by the editors
02:55 preaction add those to the "reasons I don't understand the web standards bodies anymore" :p
02:55 preaction no "CONNECT" would completely destroy $work's web proxy
02:56 jberger hehe, and that's a "problem" ;-P
02:57 preaction it's a pain in the ass, but it fufills a completely reasonable regulatory requirement
02:57 preaction we're not HIPPA, by any means, but we're close
03:00 sri the server side needs more power in the working group :S
03:01 crab internet giants!
03:02 sri it's only browser vendors rushing stuff, varnish, jetty... authors are the ones speaking sense
03:02 sri yea... and certain internet giants
03:04 jberger "do no evil" ... right
03:08 * jberger just noticed that o/ make ???? in gmail chat
03:08 preaction the girl is raising the wrong hand!
03:09 * jberger fires preaction out of a cannon
03:09 preaction https://dl.dropboxusercontent.com/u/226725/Screen%20Shot%202014-08-31%20at%2010.08.33%20PM.png
03:09 preaction wrong. hand.
03:09 * jberger slaps purl with a fish
03:09 * purl slaps jberger with comically large fish in return
03:10 jberger that's not at all the font I have, but yes I also see the wrong hand, then again, perspective
03:10 preaction ... clever... if i hold up o/, then you should see \o, but then why do i look like i'm holding up \o if in fact i am holding up o/?
03:18 * jberger spins in circles
03:19 * jberger flails wildly
03:19 * preaction dances!
03:50 jberger sri: I'm trying to post a gist and I keep only getting to '-- Connect (http:api.github.com:80)' and then nothing more happens
03:50 jberger any ideas?
04:32 jberger egad! it was https!
04:33 jberger one stupid *^%)*Q% s!
04:43 lsm joined #mojo
05:07 irq joined #mojo
05:31 zivester joined #mojo
05:43 bobkare joined #mojo
05:50 Eke- joined #mojo
06:00 laouji joined #mojo
06:07 davido___ joined #mojo
06:16 denis_boyun joined #mojo
06:20 laouji joined #mojo
06:28 laouji joined #mojo
06:30 laouji joined #mojo
06:32 laouji_ joined #mojo
06:37 Vandal joined #mojo
06:37 cpan_mojo Convos 0.83 by Jan Henning Thorsen - http://metacpan.org/release/JHTHORSEN/Convos-0.83 (depends on Mojolicious)
06:41 KCL joined #mojo
06:41 marcus batman++
06:41 laouji joined #mojo
06:44 batman \o/
06:44 batman marcus: i was about to tell you
06:44 batman https://twitter.com/convosby/status/506331703707983872
06:44 batman jberger: ^ enjoy :)
06:45 batman marcus: hope there's not too many spelling mistakes...
06:48 mr-foobar joined #mojo
06:49 marcus batman: looks good. Today I *will* blog about convos 0.8 series
06:49 batman that's just crazy talk!!
06:49 purl http://www.sinfest.net/archive_page.php?comicID=63
06:49 batman :)
06:50 batman marcus++
06:51 batman ARG! the perltidy git hook mess with git merge :(
06:52 rawler joined #mojo
06:53 laouji joined #mojo
06:57 laouji joined #mojo
06:58 laouji_ joined #mojo
07:07 edestler joined #mojo
07:09 pete joined #mojo
07:14 batman Do you have that problem marcus?
07:17 bobkare joined #mojo
07:20 dod joined #mojo
07:23 pete Has anyone ever done a speculative audit to see how many "in the wild" mojolicious apps don't change the session secrets?
07:25 Dandre joined #mojo
07:25 fhelmber_ joined #mojo
07:27 forke joined #mojo
07:32 odc joined #mojo
07:37 trone joined #mojo
07:56 dexteruk joined #mojo
08:04 woz joined #mojo
08:09 marcus batman: yes, I've had that, but it was pretty easy to fix.
08:09 batman marcus: how?
08:09 marcus by resolving the merge?
08:10 marcus I didn't mean it was easy to fix permanently.
08:13 batman i think you need to show me some day...
08:22 bobkare joined #mojo
08:28 dexteruk_ joined #mojo
08:30 marcus Maybe.
08:32 batman :)
08:32 edestler joined #mojo
08:35 bobkare joined #mojo
08:40 laouji joined #mojo
08:49 dod joined #mojo
08:53 laouji joined #mojo
08:54 cpan_mojo Mojolicious-Plugin-BootstrapHelpers 0.008 by CSSON - http://metacpan.org/release/CSSON/Mojolicious-Plugin-BootstrapHelpers-0.008
08:57 laouji joined #mojo
09:03 sri pete: how?
09:10 shadowpaste joined #mojo
09:10 Jonis joined #mojo
09:10 bayashi joined #mojo
09:10 stokachu joined #mojo
09:10 al joined #mojo
09:10 Foxcool joined #mojo
09:10 b^_^d joined #mojo
09:10 moritz joined #mojo
09:10 tianon joined #mojo
09:10 HtbaaPi joined #mojo
09:10 sivoais joined #mojo
09:10 diegok joined #mojo
09:10 mtj| joined #mojo
09:10 tardisx joined #mojo
09:10 jnthn joined #mojo
09:10 mishantil joined #mojo
09:10 jwang joined #mojo
09:10 rwp joined #mojo
09:10 nicomen joined #mojo
09:10 stephan48 joined #mojo
09:10 michael joined #mojo
09:10 KindOne joined #mojo
09:10 nic joined #mojo
09:10 lb joined #mojo
09:10 jberger joined #mojo
09:10 D4RK-PH0ENiX joined #mojo
09:10 basiliscos joined #mojo
09:10 preaction joined #mojo
09:10 lsm joined #mojo
09:10 denis_boyun joined #mojo
09:10 forke joined #mojo
09:10 bobkare joined #mojo
09:10 dod joined #mojo
09:10 edestler joined #mojo
09:10 woz joined #mojo
09:10 trone joined #mojo
09:10 odc joined #mojo
09:10 fhelmber_ joined #mojo
09:10 pete joined #mojo
09:10 KCL joined #mojo
09:10 Eke- joined #mojo
09:10 noganex_ joined #mojo
09:10 ivan joined #mojo
09:10 klapperl joined #mojo
09:10 hesperaux joined #mojo
09:10 meshl joined #mojo
09:10 dotan joined #mojo
09:10 marty joined #mojo
09:10 Claudio_ joined #mojo
09:10 og01 joined #mojo
09:10 Kundun joined #mojo
09:10 Zx3 joined #mojo
09:10 priodev joined #mojo
09:10 hahainternet joined #mojo
09:10 mikegrb_ joined #mojo
09:10 ralesk_ joined #mojo
09:10 davido__ joined #mojo
09:10 rwf joined #mojo
09:10 abhishekisnot joined #mojo
09:10 kedare joined #mojo
09:10 saki joined #mojo
09:10 jmmills joined #mojo
09:10 btyler joined #mojo
09:10 howitdo joined #mojo
09:10 Lee joined #mojo
09:10 marcus joined #mojo
09:10 Adura joined #mojo
09:10 sri joined #mojo
09:10 jberger_ joined #mojo
09:10 batman joined #mojo
09:10 crab joined #mojo
09:10 genio joined #mojo
09:10 jojo joined #mojo
09:10 Gasseus joined #mojo
09:10 Guest joined #mojo
09:10 elik_ joined #mojo
09:10 kablamo_ joined #mojo
09:10 avenj joined #mojo
09:10 kthakore_ joined #mojo
09:10 romel_ joined #mojo
09:10 phillipadsmith joined #mojo
09:10 ua1 joined #mojo
09:10 ceej joined #mojo
09:10 bc547_ joined #mojo
09:10 Jonneh_ joined #mojo
09:10 ssm joined #mojo
09:10 kmx joined #mojo
09:10 Wim joined #mojo
09:10 Gedge joined #mojo
09:10 doby joined #mojo
09:10 go|dfish joined #mojo
09:10 sawtooth joined #mojo
09:10 cpan_mojo joined #mojo
09:10 onur joined #mojo
09:10 kbenson1 joined #mojo
09:10 mtths joined #mojo
09:10 mattp joined #mojo
09:10 cosimo joined #mojo
09:10 adhoc joined #mojo
09:10 purl joined #mojo
09:10 omega joined #mojo
09:10 rj11 joined #mojo
09:10 dahl joined #mojo
09:10 jalie joined #mojo
09:10 alnewkirk joined #mojo
09:10 agorman joined #mojo
09:10 lestrrat joined #mojo
09:10 charsbar joined #mojo
09:10 mudler joined #mojo
09:10 xtetsuji joined #mojo
09:10 DaTa joined #mojo
09:10 yusukebe joined #mojo
09:10 mfontani_ joined #mojo
09:10 caeles_ joined #mojo
09:10 cfedde joined #mojo
09:10 allison_ joined #mojo
09:10 suy joined #mojo
09:10 inokenty joined #mojo
09:10 bjoernfan joined #mojo
09:10 xevz joined #mojo
09:10 mgrimes joined #mojo
09:10 mst joined #mojo
09:10 DaniBunny joined #mojo
09:10 jzawodn joined #mojo
09:10 rblackwe joined #mojo
09:10 yourname joined #mojo
09:10 Mikey joined #mojo
09:10 ashleydev joined #mojo
09:10 throughnothing joined #mojo
09:10 BinGOs joined #mojo
09:10 bd joined #mojo
09:10 jeff joined #mojo
09:10 memowe joined #mojo
09:10 rofl_ joined #mojo
09:10 cotto joined #mojo
09:10 tempire joined #mojo
09:10 Bender joined #mojo
09:10 callumgare joined #mojo
09:10 robinsmidsrod joined #mojo
09:10 gabiruh joined #mojo
09:10 bzero joined #mojo
09:10 Phil21 joined #mojo
09:10 garu joined #mojo
09:14 sri that escalated quickly http://lists.w3.org/Archives/Public/ietf-http-wg/2014JulSep/1935.html
09:19 denny joined #mojo
09:34 pete sri: My understanding reading the docs was that the secret defaults to a known key. So visit sites suspected to be running mojolicious, and see if the digest in the session was created using that default
09:34 dp_ joined #mojo
09:47 dod1 joined #mojo
09:50 michael appears to be a semi-known key (->moniker), so depends how easy is to guess that (might be the domain name etc). Might make sense to be fatal when running under production?
09:57 laouji joined #mojo
09:57 dod joined #mojo
10:00 laouji joined #mojo
10:02 laouji joined #mojo
10:05 dotan1 joined #mojo
10:16 sri pete: that's not the case
10:16 laouji joined #mojo
10:16 pete sri: What's not the case?
10:17 sri everything you said
10:17 purl i heard everything you said was based on wrong assumptions
10:17 sri botsnack!
10:17 purl thanks sri :)
10:27 good_news_everyon joined #mojo
10:27 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/MFqcEQ
10:27 good_news_everyon mojo/master 37fd93c Sebastian Riedel: mention shared options in command recipe
10:27 good_news_everyon left #mojo
10:29 good_news_everyon joined #mojo
10:29 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/90M4ag
10:29 good_news_everyon mojo/master 44e331e Sebastian Riedel: fixed typos in command recipe
10:29 good_news_everyon left #mojo
10:30 dexteruk joined #mojo
10:31 d4rkie_ joined #mojo
10:31 sri pete: see also http://mojolicio.us/perldoc/Mojolicious/Guides/FAQ#What-does-Your-secret-passphrase-needs-to-be-changed-mean
10:31 irq joined #mojo
10:33 cosimo joined #mojo
10:41 dotan1 joined #mojo
10:44 d4rkie joined #mojo
10:47 bobkare joined #mojo
10:54 bobkare joined #mojo
11:18 meshl joined #mojo
11:35 cpan_mojo Mango 1.09 by Sebastian Riedel - http://metacpan.org/release/SRI/Mango-1.09 (depends on Mojolicious)
11:41 good_news_everyon joined #mojo
11:41 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/JECb_w
11:41 good_news_everyon mojo/master 016c62a Sebastian Riedel: fixed a few typos
11:41 good_news_everyon left #mojo
12:10 good_news_everyon joined #mojo
12:10 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/PFn5JQ
12:10 good_news_everyon mojo/master e639d64 Sebastian Riedel: mention that additional tests can run on the transaction
12:10 good_news_everyon left #mojo
12:20 good_news_everyon joined #mojo
12:20 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/L_fzgA
12:20 good_news_everyon mojo/master a9b9f77 Sebastian Riedel: describe Mojolicious as a web toolkit
12:20 good_news_everyon left #mojo
12:23 good_news_everyon joined #mojo
12:23 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/NFZ1fA
12:23 good_news_everyon mojo/master 1c23846 Sebastian Riedel: use more words
12:23 good_news_everyon left #mojo
12:39 marcus https://translate.google.com/translate?sl=auto&tl=en&js=y&prev=_t&hl=en&ie=UTF-8&u=http%3A%2F%2Fhabrahabr.ru%2Fpost%2F233991%2F&edit-text=&act=url
12:41 bobkare joined #mojo
12:42 sri someone dissected the mojo coding style :)
12:43 marcus I wonder if he knows we're using perltidy for the formatting.
13:07 Kripton joined #mojo
13:22 cpan_mojo Mojolicious-Plugin-BootstrapHelpers 0.009 by CSSON - http://metacpan.org/release/CSSON/Mojolicious-Plugin-BootstrapHelpers-0.009
13:23 neilhwatson joined #mojo
13:30 punter joined #mojo
13:37 sri he noticed a lot of things
13:38 sri even newer stuff i only picked up recently, like using slices everywhere
13:40 sri i don't think perltidy matters much in that context
13:42 sri hehe, aside from #17, which is a direct result of perltidy not being able to make that code pretty otherwise :)
13:43 __eugen joined #mojo
13:58 Kripton joined #mojo
14:07 sh4 joined #mojo
14:09 sri too bad nobody has a workaround for the template problem
14:09 sri https://groups.google.com/forum/#!topic/mojolicious/4lqjcY4vdD4
14:16 basiliscos joined #mojo
14:25 sri got a minimal test case now too https://gist.github.com/anonymous/e0282b998cb996d54e50
14:32 augensalat joined #mojo
14:38 lipizzan joined #mojo
14:49 mr-foobar joined #mojo
14:50 zivester joined #mojo
14:53 punter joined #mojo
14:55 marty I have set the env variables MOJO_MODE=production; MOJO_LOG_LEVEL=debug;   but calls to $self->app->log->debug("ooops"); are not logging.  Any ideas what I am doing wrong?
14:56 marty My logging statements work fine when in development mode
15:01 jberger joined #mojo
15:02 jberger o/ from Convos 0.83!
15:02 jberger batman++
15:02 jberger marcus++
15:07 sri hahaha
15:07 sri i have a solution
15:07 batman jberger: cool :)
15:11 sri this patch actually works across all versions of perl ;p https://gist.github.com/anonymous/fb6b70d78e954238f527
15:12 sri the escaped newlines inside the multiline quoted string were the problem
15:15 cpan_mojo Mojolicious-Command-nopaste 0.04 by Joel Berger - http://metacpan.org/release/JBERGER/Mojolicious-Command-nopaste-0.04
15:15 jberger batman: in turn I give you this ...
15:16 jberger argh! it just beat me to it
15:16 batman jberger++ # gist :)
15:16 GabrielVieira joined #mojo
15:16 batman i want to use that code in https://github.com/Nordaaker/convos/issues/90
15:16 jberger sri++ very cool
15:17 jberger PLUS PLUS ALL THE THINGS!!!!
15:17 GabrielVieira sri: have you tested the  optimization in Perl 5.14?
15:17 jberger batman: I hadn't though about reuse
15:17 jberger hmmmmm
15:17 batman jberger: you don't need to. i'm just going to use it as inspiration
15:19 sri GabrielVieira: why do you ask?
15:20 jberger <3 the red (read?) line!
15:20 GabrielVieira looks like the test is passing here, or I'm doing it wrong.. or the problem started before 5.14, I saw in your gist you tested in 5.16.
15:20 GabrielVieira sri: ^
15:20 GabrielVieira in 5.14*
15:20 sri GabrielVieira: then you're doing it wrong
15:21 sri anyway, i know what caused the problem, and have a fix
15:21 GabrielVieira good
15:24 jberger batman: do you think that what I did is sufficient? requiring you to get an access token and not supporting the full OAuth flow?
15:25 batman jberger: yes, i think so... anyway, i'm not going to do that until we have https://github.com/Nordaaker/convos/issues/89
15:26 batman right now, i'm focusing on http://home.thorsen.pm/private/raw/notverysecret/convos/channel-list-2014-08-31-15-11-41.ogv and https://cloud.githubusercontent.com/assets/45729/4106241/7baf77b6-31b7-11e4-9dda-f5b3eb00b7ea.png
15:26 batman oh. that looked a bit funny... :)
15:29 jberger purl: convos?
15:29 purl i guess convos is really great on Android.
15:31 jberger purl: convos is also teh awesomez
15:31 purl okay, jberger.
15:32 jberger purl: convos is also at http://convos.by/
15:32 purl okay, jberger.
15:32 jberger convos?
15:32 purl well, convos is really great on Android. or teh awesomez or at http://convos.by/
15:32 jberger purl: botsnack
15:32 purl :)
15:35 Akron joined #mojo
15:36 jberger hmmm a couple tweaks are needed to gist support I think
15:42 jberger there, that's better
15:44 cpan_mojo Mojolicious-Command-nopaste 0.05 by Joel Berger - http://metacpan.org/release/JBERGER/Mojolicious-Command-nopaste-0.05
15:47 noganex joined #mojo
15:47 Dandre joined #mojo
15:51 odc joined #mojo
16:00 sri hehe, discovered another special case
16:02 jberger Galileo is one star from 100 stars on github
16:02 * sri unstars ;p
16:03 * btyler stars
16:03 neilhwatson ++
16:03 sri oh noes
16:06 jberger \o/
16:06 jberger made it
16:06 jberger neilhwatson: did 0.036 release fix your issue?
16:07 neilhwatson Yes, I see the config page now. Thanks.
16:07 jberger cool
16:08 jberger kinda gratifying to get three bug reports when things break, means at least three people tried installing Galileo when it was broken (which wasn't all that long)
16:08 neilhwatson I know the feeling.
16:15 disputin joined #mojo
16:17 neilhwatson That's odd, setup worked, then it didn't when I reran it.
16:18 jberger why did you rerun it?
16:18 jberger it 404ed on the rerun or it bailed out?
16:19 neilhwatson 404
16:19 jberger hmmmm
16:19 jberger I suspect browser caching
16:19 jberger refresh
16:20 neilhwatson that was it.
16:20 * jberger wipes brow
16:21 * jberger considers proposing a grant for Mojolicious book: http://blogs.perl.org/users/makoto_nozaki/2014/08/september-grant-round-is-just-around-the-corner.html
16:23 Akron jberger++
16:23 Akron Great idea!
16:23 neilhwatson ++
16:23 sri do it!
16:24 marcus +1
16:24 purl 1
16:24 batman jberger++
16:44 * jberger brushes up his .latexmkrc file
16:49 neilhwatson jberger: Galileo built in editor does not play well with vimperator, a mozilla plugin. ctrl-i usually opens external editor in text fields. Galileo editor captures ctrl-i for its own uses.
16:51 good_news_everyon joined #mojo
16:51 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/Y1tN6w
16:51 good_news_everyon mojo/master ce1a62e Sebastian Riedel: optimize successive text lines in Mojo::Template
16:51 good_news_everyon left #mojo
16:51 sri phew, getting this right was rather tough
16:54 jberger neilhwatson: the editor is the same as the one from stack overflow. Do you have the same trouble there?
16:54 jberger sri++
16:54 marcus I use pentadactyl - it's far superior to vimperator :)
16:55 chankey joined #mojo
16:55 neilhwatson Yes, have seen that editor elsewhere, do not like it. Can't do any serious editing in it.
16:55 sri after hacking on Mojo::Template so much, i get the feeling it should be a little smarter about handling newlines
16:55 neilhwatson I can add some feature requests if you like.
16:56 jberger neilhwatson: I don't know if there is much I can do about it
16:58 neilhwatson FYI, The default Wordpress editor does not behave this way, but I'm not trying to be adversarial.
16:58 * neilhwatson does not like Wordpress
17:00 jberger the place to discuss would be here: http://code.google.com/p/pagedown/issues/list
17:05 jberger neilhwatson: if you could make that configurable I would be all ears, but since the target demographic is the less technical end of the spectrum, I don't see myself changing the default, even if it were possible
17:13 Akron sri: Do you mean "syntax wise" smarter, e.g. with code areas?
17:28 sri nope
17:28 sri just internals
17:40 tempire grant proposal for a book is a good idea.
17:40 tempire Wish I would have thought of it first.
17:41 meshl joined #mojo
17:42 jberger tempire: as you are sans employer, if you want it, I defer to you
17:42 jberger my fear would be finding enough time
17:43 tempire hmm
17:44 tempire Basics. Non-blocking basics.
17:44 tempire Maybe a touch of delay
17:44 jberger I even have some latex tooling worked out if you want: https://github.com/jberger/MojoBook
17:45 jberger I didn't really work on the book, I was trying to sort out things like highlighting etc
17:45 tempire I would probably use a mac tool that did most of that for me
17:45 tempire Or indesign, actuall
17:45 tempire y
17:45 jberger you mac users and your mac things
17:50 jberger this is the current state: https://drive.google.com/file/d/0B-gBN2ue1FmVaDFjbWhvbWxONWs/edit?usp=sharing
17:54 sri tempire has some experience too, from the course material
17:56 sri (re tooling)
17:56 sri although i guess turning it all into ebook form is harder than it might look
17:57 tempire /me wonders how much grant $ is reasonable
17:57 sri this one comes to mind http://news.perlfoundation.org/2014/05/grant-proposal-modern-object-o.html
17:59 jberger I was looking into the LaTeX to ebook problem too
17:59 jberger :(
18:00 jberger the tools I find don't seem to like the subimport package that I like for large projects
18:01 jberger LaTeX likes to be monolithic and if not then at least flat folder structure, subimport helps that but I suspect it is rather evil
18:02 Eke- joined #mojo
18:02 Akron jberger: Use org-mode! ;)
18:03 Akron However, I don't think the tools will be the problem.
18:27 good_news_everyon joined #mojo
18:27 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/iO5alA
18:27 good_news_everyon mojo/master 794de43 Sebastian Riedel: better template tests
18:27 good_news_everyon left #mojo
18:40 berov joined #mojo
18:42 KCL_ joined #mojo
19:03 Eke- joined #mojo
19:21 MojoNoPaste joined #mojo
19:28 denis_boyun joined #mojo
19:30 denis_boyun_ joined #mojo
19:31 MojoNoPaste joined #mojo
19:31 MojoNoPaste jberger pasted https://gist.github.com/e6f58d032862a2b72b66
19:31 jberger \o/
19:32 marcus wowpaste
19:38 batman jberger: cool :)
19:38 denis_boyun joined #mojo
19:41 Dandre left #mojo
19:43 jberger https://github.com/jberger/Mojolicious-Command-nopaste/commit/1efbdaf109466b84d559fdfa62814e3fcae41da9
19:44 denis_boyun_ joined #mojo
19:52 denis_boyun joined #mojo
20:00 jberger batman: can I get a quick glance at this: https://github.com/jberger/Mojolicious-Command-nopaste/compare/15597595c14b8c12760959848d7548563e2ed6d9...master
20:00 jberger does it seem sane?
20:01 jamesaxl joined #mojo
20:12 irq joined #mojo
20:31 bc547 joined #mojo
20:33 woz joined #mojo
20:36 Kripton_ joined #mojo
20:41 meshl joined #mojo
20:44 jberger I'm feeling lucky, going for it
20:46 sri that's... new
20:50 good_news_everyon joined #mojo
20:50 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/FSgdqw
20:50 good_news_everyon mojo/master 638a3a6 Sebastian Riedel: use a better AST in Mojo::Template
20:50 good_news_everyon left #mojo
20:51 cpan_mojo Mojolicious-Command-nopaste 0.06 by Joel Berger - http://metacpan.org/release/JBERGER/Mojolicious-Command-nopaste-0.06
20:52 jberger ^^ will use Mojo::IRC to paste to a channel if you specify one and if the service doesn't already to it (i.e. shadowcat)
21:05 woz joined #mojo
21:08 sri oh funny, md5_hex is actually a bottleneck when rendering inline templates
21:09 sri this line specifically https://github.com/kraih/mojo/blob/master/lib/Mojolicious/Plugin/EPRenderer.pm#L25
21:15 sri wow, just not generating a checksum increases performance quite a bit
21:16 sri from 1.37s down to 1.11s
21:16 sri MOJO_TEMPLATE_DEBUG=1 perl -Ilib -Mojo -E 'my $c = app->build_controller; my $inline = "<p>test</p>\n" x 1000; n { $c->render_to_string(inline => $inline) } 10000'
21:17 elik joined #mojo
21:17 sri only problem is we still need a name for caching
21:25 marcus 200ms checksum :o
21:37 preaction joined #mojo
21:41 sri haha, it actually gets much faster
21:42 sri 0.94s with an old hack i found
21:42 sri the shittiest hash algorithm i could find https://gist.github.com/anonymous/79df365c1a7c9a7a2ff8
21:43 sri from 1.37s down to 0.94s
21:43 woz joined #mojo
21:46 sri is perl's core hash algorithm exposed somehow?
21:48 trone joined #mojo
21:51 sri with Digest::OAT it goes down to 1.13s
21:52 sri oooh, Digest::SipHash is great
21:53 sri 0.96s
21:58 mishantil SipHash seems like a nice fit.
22:04 jberger Digest::xxHash?
22:06 punter joined #mojo
22:07 sri 0.90s :)
22:08 jberger Mojo::Util::hash which delegates to available hashing algos?
22:10 sri according to my tests it is sadly not really worth it
22:10 sri outside of micro benchmarks there is no real gain
22:11 jberger probably true, these are being used to render and cache once, yeah?
22:11 jberger to so unless you have gazillion templates, no one cares
22:11 jberger still, interesting to think about
22:11 sri they are used for every render call
22:12 sri to get the key for looking up the template in the cache
22:12 jberger hmmmmmmm
22:12 jberger you could just make the "key" the template string itself
22:12 sri i actually did add a Mojo::Util::fast_sum and replaced all uses of md5_sum with it :)
22:13 jberger that relies on the core hash algo
22:14 sri that's what i started with
22:14 sri the 1.11s
22:14 jberger oh ok
22:15 sri but now that you mention it
22:15 purl hmmm... now that you mention it is a todo the point is moot
22:15 Akron joined #mojo
22:16 sri we can totally not use a checksum there
22:16 sri i shuffled around stuff a few weeks ago... and now it's clean enough... the problem i was thinking of earlier just doesn't exist anymore :D
22:22 good_news_everyon joined #mojo
22:22 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/3fBIRQ
22:22 good_news_everyon mojo/master ff47d9c Sebastian Riedel: no need to use a checksum for caching
22:22 good_news_everyon left #mojo
22:23 sri the benchmark went from 1.9s earlier today to 1.1s now :)
22:26 sri jberger++
22:26 Akron So this is just a problem with lots of long inline templates, right?
22:26 sri i don't think it is actually
22:27 sri like jberger said, we just let perl create the checksum now
22:27 sri it's better than now in every way
22:28 sri oh wait
22:28 Akron But before the string content was removed, now it will be stored for inline templates.
22:28 sri where do you mean?
22:28 Akron I mean - it was removed after compilation.
22:29 sri it actually gets stored twice
22:30 sri once in the cache hash, and once in the stack of current cache keys ;p
22:30 sri well... shit
22:32 Akron Oh, true, inline stays intact.
22:33 sri jberger: or am i overlooking something?
22:36 good_news_everyon joined #mojo
22:36 good_news_everyon [mojo] kraih pushed 1 new commit to master: http://git.io/XTebfw
22:36 good_news_everyon mojo/master aec221f Sebastian Riedel: caching inline templates requires too much memory
22:36 good_news_everyon left #mojo
22:39 Akron How does the Hash::Util::hash_value() perform instead?
22:40 Akron (Sorry - can't test it now ...)
22:43 sri 1.12s
22:44 sri ah, it's a very new function
22:47 trone joined #mojo
22:48 sri new in 5.18 i believe
22:49 Akron Nice little function. I can see use for it. But it's a pitty it's so new ... and making this a preference if perl >= 5.18 instead of md5_sum?
22:50 sri not sure it's worth extra trouble
22:51 Akron Oh - but you could just md5_sum inline templates - so there's at least a speedup for file templates.
22:52 sri again not sure it's worth the extra trouble
22:52 Akron What's the extra trouble?
22:52 sri extra code
22:55 Akron It's not an extra loc, as far as I can see. ;)
22:56 sri it isn't?
22:57 Akron Well - I would do "my $name = defined $options->{inline} ? md5_sum($options->{inline}) : $renderer->template_name($options);" and then remove the md5_sum later on. However, I don't know why you need the encoding. But maybe I'm completely wrong.
22:57 sri yes, you need the encoding
22:57 sri md5 likes to die with wide characters
22:58 Akron Oh.
22:58 Akron That's good to know.
22:59 sri not encoding would be a nice performance boost as well
22:59 sri anyway, if it can be done with the same amount of code, please send a patch! otherwise i think it needs a measurable performance gain
23:01 Akron Okay.
23:12 preaction joined #mojo
23:34 sri btw. getting rid of the encode is a bigger gain than getting rid of the md5_sum

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