stylus renders slow with nib

[acidjazz]

benchmark results for a core i5 -> all files have 100+ lines of code and 10+ includes

jade compilation: 0.08s stylus compilation: 0.2s stylus with nib compilation: 2.2s

any solutions for this? has nib been profiled?

[acidjazz]

profile to possibly help

[acidjazz]

one of the ways I got jade to run so fast was passing an option "self" that wasn't documented. are there any options in stylus not documented I could play with to possible tweak it?

[notslang]

I've never actually noticed any lag when compiling Stylus; other than huge sites like carrot.is, but that's in combination with several other compiled assets. But if you're interested in optimizing nib then I'd be happy to work with you on it. However, I don't know of any undocumented Stylus options like that.

Also, I'd love to hear more about that 'self' option in Jade.

[acidjazz]

Thanks @slang800 ,

As far as Jade goes, the trick I found is if you pass "self" as true as a render option and then refer to all variables passed to jade as self.??? then you get a 100x performance increase. This was a huge find when I stress tested my jade. In fact w/out this optimization, operations was about to pick a pretty big fight on jade being the major bottleneck, and a very measurable cost in hosting.

Here is where they found and added it

https://github.com/visionmedia/jade/pull/217

[acidjazz]

@slang800 as far as stylus goes, after doing more playing around i noticed the difference in load time is about the same even w/ no css, so an empty .styl file w/out a nib include was 0.02s load time, w/ a nib include 0.1s load time. Then as I added lines of stylus code the nib included load time increases little by little.

[amwmedia]

I was able to speed up stylus by first concatenating the styl files into 1. That way nib only has to load once.

http://blog.amwmedia.com/post/110919529872/speeding-up-stylus-with-grunt-contrib-concat