Use SassC to compile Sass #2176
Conversation
boffbowsh
force-pushed
the
use-sassc-for-assets
branch
from
d63564a to
df8747e
Compare
Relies on a fork at the moment, hoping to have the change merged upstream: sass/sassc-rails#17
boffbowsh
force-pushed
the
use-sassc-for-assets
branch
from
df8747e to
0999648
Compare
boffbowsh
force-pushed
the
use-sassc-for-assets
branch
from
5a243df to
8f6b454
Compare
| end | ||
| end | ||
|
|
||
| Sprockets::Base.send(:prepend, Extensions::Sprockets::Engines) |
There was a problem hiding this comment.
This is a bit gross, but I assume there's no better way?
There was a problem hiding this comment.
Should have mentioned, this is the current recommended workaround from sass/sassc-rails#6 (comment)
There was a problem hiding this comment.
Worth a comment, or at least adding to the commit message?
|
👍 |
| @@ -67,7 +67,8 @@ else | |||
| end | |||
|
|
|||
| gem 'sass', '3.4.9' | |||
| gem 'sass-rails' | |||
| gem 'sass-rails' # required by bootstrap-sass, but not listed as a dependency of it | |||
| gem 'sassc-rails', github: 'bolandrm/sassc-rails', branch: 'master' | |||
There was a problem hiding this comment.
Forgive my possible misunderstanding on how Gemfiles and Gemfile.locks work but if the master branch changes on this repo is there a chance that we will get that when we bundle? Would it be safer to point this at a sha rather than a branch?
There was a problem hiding this comment.
Good spot @edds.
We'd have to bundle update to change the version, and the lock file would change and need to be checked in to take effect. In theory that means both the committer and the reviewer would be encouraged to ask "why is this?", but it would definitely be better to specify a tag or a sha here.
There was a problem hiding this comment.
Yeah, hopefully this is a short-term thing, the gem should be released at some point soon at which point we can point it at a release.
Relies on a fork at the moment, hoping to have the change merged upstream:sassc-railsGem not been released yet, relies on a bugfix in master only: sass/sassc-rails#17I've compared the output of both engines in
:expandedmode. It's not the easiest diff in the world to review, but it does seem to be mostly whitespace, wrapping of long selectors and leading zeroes, e.g..5emversus0.5em: https://gist.github.com/boffbowsh/28da0cdd02022b26eada.<is ruby Sass,>is SassC.Speedup on my laptop is 47s versus 3m11s.