07:19:44 <KwBot> [nikola] Kwpolska assigned issue #1904 to Kwpolska: Post.title() returns chunks of JSON for some IPython NB https://github.com/getnikola/nikola/issues/1904 07:30:45 <KwBot> [nikola] Kwpolska closed issue #1904: Post.title() returns chunks of JSON for some IPython NB https://github.com/getnikola/nikola/issues/1904 07:31:00 -GitHub[nikola]:#nikola- [nikola] Kwpolska pushed 1 new commit to master: http://git.io/vYTeE 07:31:00 -GitHub[nikola]:#nikola- nikola/master 7031c54 Chris Warrick: Fix #1904 -- compiler meta as good as 2-file... 07:34:50 -travis-ci:#nikola- getnikola/nikola#5810 (master - 7031c54 : Chris Warrick): The build is still failing. 07:34:51 -travis-ci:#nikola- Change view: https://github.com/getnikola/nikola/compare/81fad1cd7567...7031c54479a9 07:34:51 -travis-ci:#nikola- Build details: https://travis-ci.org/getnikola/nikola/builds/72072905 07:41:43 -GitHub[nikola]:#nikola- [nikola] Kwpolska pushed 1 new commit to master: http://git.io/vYTJu 07:41:43 -GitHub[nikola]:#nikola- nikola/master e7b5b79 Chris Warrick: Create a compiler_meta dict if there isn’t any... 07:45:37 -travis-ci:#nikola- getnikola/nikola#5811 (master - e7b5b79 : Chris Warrick): The build is still failing. 07:45:38 -travis-ci:#nikola- Change view: https://github.com/getnikola/nikola/compare/7031c54479a9...e7b5b79207dc 07:45:38 -travis-ci:#nikola- Build details: https://travis-ci.org/getnikola/nikola/builds/72073913 07:49:00 -GitHub[nikola]:#nikola- [nikola] Kwpolska pushed 1 new commit to master: http://git.io/vYTTe 07:49:00 -GitHub[nikola]:#nikola- nikola/master ff62213 Chris Warrick: Use webiso instead of iso in mako, too (cc @Aeyoun)... 07:52:46 -travis-ci:#nikola- Invariant build by Chris Warrick: The build passed. See https://travis-ci.org/getnikola/invariant-builds/builds/72074705. 07:53:29 -travis-ci:#nikola- getnikola/nikola#5812 (master - ff62213 : Chris Warrick): The build is still failing. 07:53:30 -travis-ci:#nikola- Change view: https://github.com/getnikola/nikola/compare/e7b5b79207dc...ff6221354c1f 07:53:30 -travis-ci:#nikola- Build details: https://travis-ci.org/getnikola/nikola/builds/72074584 08:00:45 -travis-ci:#nikola- getnikola/nikola#5811 (master - e7b5b79 : Chris Warrick): The build is still failing. 08:00:45 -travis-ci:#nikola- Change view: https://github.com/getnikola/nikola/compare/7031c54479a9...e7b5b79207dc 08:00:46 -travis-ci:#nikola- Build details: https://travis-ci.org/getnikola/nikola/builds/72073913 08:09:25 -travis-ci:#nikola- getnikola/nikola#5812 (master - ff62213 : Chris Warrick): The build was fixed. 08:09:26 -travis-ci:#nikola- Change view: https://github.com/getnikola/nikola/compare/e7b5b79207dc...ff6221354c1f 08:09:26 -travis-ci:#nikola- Build details: https://travis-ci.org/getnikola/nikola/builds/72074584 08:38:11 -GitHub[site]:#nikola- [nikola-site] Kwpolska pushed 1 new commit to redesign: http://git.io/vYTni 08:38:11 -GitHub[site]:#nikola- nikola-site/redesign be6bf58 Chris Warrick: Document MacPorts and Fink... 12:52:21 <ChrisWarrick> In other news: jekyll is still a pile of crap 12:55:34 <ralsina> ChrisWarrick: what happened specifically? Or is it just a periodic reminder? ;-) 12:56:17 <ChrisWarrick> ralsina: it needs so much tinkering just to get a post out of it 12:56:22 <ChrisWarrick> ralsina: and it wants weird file names 13:03:42 <ChrisWarrick> AND it needs special post-processing, which is awesome 13:07:35 <ralsina> ChrisWarrick: yet, 20K stars in GH :-( 13:07:57 <ralsina> hey, we have 800! Nice round number ;-) 13:08:11 <ChrisWarrick> I’m actually playing with some speed comparison 13:08:39 <erdgeist> Ohh, maybe I should star nikola, too 13:09:56 <erdgeist> co-co-combo breaker 13:12:04 <ChrisWarrick> gour: hello, mister hugo fan. how the hell do you get this thing to work and produce actual websites? 13:13:32 <gour> ChrisWarrick: no idea - no more Hugo here, only Tesla ;) 13:13:40 <ChrisWarrick> gour: s/Tesla/Nikola/ 13:13:48 <gour> :-) 13:14:01 <ChrisWarrick> speking of which, tesla the car company should totally use Nikola for their website 13:14:54 <gour> lol 13:16:26 <ChrisWarrick> Okay, I can’t set hugo up, so we probably don’t have them as contenders. What other SSGs should I include in my test? 13:18:45 <ChrisWarrick> Average time, from 4 rebuilds: PELICAN 4.29; NIKOLA 3.02; JEKYLL 2.285 13:19:19 <gour> ChrisWarrick: what's wrong with Hugo? when you pull repo, you can just use e.g. its doc folder as example site...iow. i believe Hugo should be included if you measure performance 13:19:42 <ChrisWarrick> gour: please send me a working config for a site at least as functional as Nikola. 13:20:10 <gour> ChrisWarrick: ok, will try...sometime later, ok? 13:20:24 <ChrisWarrick> gour: I don’t really care. You don’t have to do it. 13:21:25 <ralsina> gour: really? Want to share the story of your recovery? ;-) 13:21:53 <ralsina> so Jekyll is fast now? 13:22:20 <ralsina> ChrisWarrick: perhaps create a site with 100 posts and change 1 to see how that woorks out 13:22:41 <gour> ralsina: sure. when i find time to properly re-vive my old blog, i'll write about it 13:23:21 <ChrisWarrick> ralsina: I have 178 (logs from this channel) already and am doing blank rebuilds right now 13:23:30 <ralsina> ah ok 13:23:40 <ChrisWarrick> ralsina: (first build: 4.42/8.47/2.60) 13:27:42 <ChrisWarrick> let’s see how bad hexo works 13:34:31 <ChrisWarrick> hexo crashed with hexo generate 83.90s user 0.89s system 97% cpu 1:26.52 total 13:36:53 <ChrisWarrick> successful build hexo generate 90.13s user 1.17s system 98% cpu 1:33.04 total 13:38:49 <ralsina> whoa 13:39:15 <ralsina> BUT HEXO IS "fast, simple & powerful" !!!oneeleven! 13:39:30 <superfly> bwahaha 13:39:48 <ralsina> "Node.js brings you incredible generating speed. Hundreds of files take only seconds to build." 13:39:58 <ralsina> just 100, 200 seconds maybe. 13:40:05 <superfly> I'm so over Node.js 13:40:08 <ChrisWarrick> Can you believe this? Hexo is actually doing INCREMENTAL REBUILDS 13:40:16 <ChrisWarrick> And the second build was ultra fast! hexo generate 89.97s user 1.31s system 97% cpu 1:33.52 total 13:40:17 <ralsina> <blink> 13:40:37 <ralsina> saved 2 whole seconds? 13:40:51 <ralsina> ah, no 13:40:55 <ralsina> 0.2 seconds? 13:40:59 <ChrisWarrick> 0.16 13:41:17 <ralsina> weeeeee 13:41:50 <ChrisWarrick> Nikola saves 64% of the original build time. Hugo saves 0.17% of the original build time. 13:51:34 <ChrisWarrick> http://t.co/Uutz9oCIkE 13:53:04 <bwinton> ChrisWarrick: Okay, but why is Nikola so much slower than Pelican and Jekyll on the first run? ;) 13:53:26 <ChrisWarrick> bwinton: we generate all the files, including a lot of crap others don’t have 13:54:09 <ChrisWarrick> bwinton: we’re still faster than Pelican in average-of-5, and I could slim it down by dropping some useless resources and source .rst files 13:56:00 <bwinton> ChrisWarrick: Yeah, I saw that. :) Also, are you changing anything in between the runs? (Do the incremental rebuilds have any work to do?) 13:56:28 <bwinton> (Not that I expect Hexo to do any better there, but it would be interesting to see how the others compare. :) 13:58:34 <ralsina> we could start making more stuff optional, but really, if the difference is < 5 seconds in a largish site I don't care 13:58:50 <ralsina> like sitemaps and whatever 13:59:08 <ChrisWarrick> we could also do jquery colorbox i18n intelligently 14:00:24 <ChrisWarrick> currently we copy a 38 files, and 35–38 are actually useless 14:00:31 <ralsina> true 14:00:50 <ralsina> but we have no filter-assets-per-locale step 14:00:58 <ralsina> and looks like cheating for a benchmark ;-) 14:01:17 <ChrisWarrick> we can build it into the core 14:05:37 <ralsina> ChrisWarrick: a large chuk is our initialization which is crazy slow. Check how long nikola help takes 14:05:46 <ralsina> which does *nothing* except load plugins 14:06:08 <ChrisWarrick> nikola help 1.02s user 0.32s system 81% cpu 1.632 total 14:06:11 <ralsina> it's 1.15s in my (rather fast, SSD) device 14:06:35 <ChrisWarrick> Jekyll does 0.49 14:06:38 <ralsina> so, we spend ~30% of a typical run doing initialization, and 60% building stuff 14:06:52 <ralsina> give or take 10% 14:07:06 <ChrisWarrick> I actually tried to do the test on my local home machine, and we were really bad compared to Pelican (24s vs 8s?) 14:07:24 <ralsina> slow disk? 14:08:08 <ralsina> I suspect the loading / parsing of 50 ini files is not fast 14:08:29 <ralsina> there is a mechanism to get rid of the .plugin files in yapsy but I never really looked into it 14:10:13 <ralsina> 57 plugins in core 14:14:46 <ralsina> removing all plugins (and changing one line so it doesn't fail because it can't load mako) it goes down to less than 0.3 seconds 14:15:20 <ralsina> so best case, if we make all features load INSTANTLY, we save .8 seconds or so 14:15:40 <ralsina> then again this is in my machine, in others it will be much slower 15:32:53 <ChrisWarrick> turns out hexo was trying to do highlighting, it is faster otherwise