What is the plan for updating libsass? #233

Closed
domenic opened this Issue Jan 22, 2014 · 38 comments

Comments

Projects
None yet

domenic commented Jan 22, 2014

Apologies for opening an issue on this, but I couldn't quite find it documented anywhere...

We're trying to move our toolchain over to libsass, but are blocked on hcatlin/libsass#246. That was fixed five days ago, but the changes haven't propagated to node-sass yet: it appears you guys are using a submodule revision that's a month old.

I guess you guys are pegged to the 1.0.1 version from a month ago? Is there any chance of getting a newer bleeding-edge version, or instructions on how to compile such a version?

Contributor

nschonni commented Jan 22, 2014

Is there any chance of getting a newer bleeding-edge version, or instructions on how to compile such a version?

https://github.com/andrew/node-sass#rebuilding-binaries

Contributor

nschonni commented Jan 22, 2014

There has been alot of awesome work in libsass lately, but a chunk of the @extend tests are now failing. I haven't had a chance to dig through whether the are just new tests or that latest fixes will break currently deployed SCSS. It's up to @andrew though.

PS: updating libsass means we need to co-ordinate cutting new binaries for Windows, Linux, and OSX.

domenic commented Jan 22, 2014

Yeah, I can certainly appreciate wanting to wait for a new release from upstream. We'll try rebuilding the binaries ourselves and using that. It'll get a bit tricky doing it for Linux on our build server and Windows for our dev servers, but hey, somebody's got to do it---and we're the ones in such a rush, not you guys :)

Member

andrew commented Jan 22, 2014

@domenic which platforms are you using it on? Would you be ok for running on master for a little while? The problem with trying to push new releases when ever libsass updates is that we get their bug reports and lots of unhappy users.

Contributor

nschonni commented Jan 22, 2014

PS: The binaries live in https://github.com/andrew/node-sass-binaries if you find that it's stable and want to share

domenic commented Jan 22, 2014

@andrew We're running via a gulp-sass -> node-sass -> libsass on both Windows and Linux. We'd definitely be OK running on master.

Member

andrew commented Jan 22, 2014

I just tried pulling master from libsass and the tests are all kinds of broken right now 😦

glebm referenced this issue in twbs/bootstrap-sass Jan 28, 2014

Closed

libsass support #494

3 of 3 tasks complete

When it will update?
I have error with official bootstrap-sass
…/bower_components/bootstrap-sass-official/vendor/assets/stylesheets/bootstrap/mixins:365: error: non-terminal statement or declaration must end with ';'
but twbs/bootstrap-sass#494 say it fixed in libsass.

Member

andrew commented Jan 30, 2014

@theaqua we've got some failing tests with master version of libsass, I'm not going to have any time to update it for at least a week right now.

Contributor

mgol commented Jan 30, 2014

I may want to spend some time on improving the test but I was trying to run tests on current node-sass master first and I got the error:

$ npm run test  

> node-sass@0.8.1 pretest /Users/mgol/Documents/projects/public/node-sass
> jshint .


> node-sass@0.8.1 test /Users/mgol/Documents/projects/public/node-sass
> mocha test


  ․․․․․․․․node(67436,0x7fff77ac0310) malloc: *** error for object 0x10080f8b0: pointer being freed was not allocated
*** set a breakpoint in malloc_error_break to debug
[1]    67433 abort      npm run test

What could be it?

Contributor

mgol commented Jan 30, 2014

Also, I've tried to run test with libsass from master with the following result:

$ cd libsass
$ git pull origin master
$ cd ..
$ npm install
...
$ node-gyp rebuild
...
$ npm run test

> node-sass@0.8.1 pretest /Users/mgol/Documents/projects/public/node-sass
> jshint .


> node-sass@0.8.1 test /Users/mgol/Documents/projects/public/node-sass
> mocha test


  ․․․․․․․․․․․․․․․․․․․․․․․․․․․․․․

  29 passing (277ms)
  1 failing

  1) cli should compile with the --output-style:

      actual expected

      #navbar {width:80%;height:23px;}#navbar ul {list-style-type:none;}#navbar li {float:left;}#navbar li a {font-weight:bold;}

      at EventEmitter.<anonymous> (/Users/mgol/Documents/projects/public/node-sass/test/cli.js:82:14)
      at EventEmitter.emit (events.js:106:17)
      at /Users/mgol/Documents/projects/public/node-sass/lib/render.js:28:17
      at /Users/mgol/Documents/projects/public/node-sass/node_modules/mocha/node_modules/glob/node_modules/graceful-fs/graceful-fs.js:103:5
      at Object.oncomplete (fs.js:107:15)




npm ERR! node-sass@0.8.1 test: `mocha test`
npm ERR! Exit status 1
npm ERR! 
npm ERR! Failed at the node-sass@0.8.1 test script.
npm ERR! This is most likely a problem with the node-sass package,
npm ERR! not with npm itself.
npm ERR! Tell the author that this fails on your system:
npm ERR!     mocha test
npm ERR! You can get their info via:
npm ERR!     npm owner ls node-sass
npm ERR! There is likely additional logging output above.
npm ERR! System Darwin 13.0.0
npm ERR! command "node" "/usr/local/bin/npm" "run" "test"
npm ERR! cwd /Users/mgol/Documents/projects/public/node-sass
npm ERR! node -v v0.10.25
npm ERR! npm -v 1.3.25
npm ERR! code ELIFECYCLE
npm ERR! 
npm ERR! Additional logging details can be found in:
npm ERR!     /Users/mgol/Documents/projects/public/node-sass/npm-debug.log
npm ERR! not ok code 0

Doesn't sound like everything broke and the error is trivial to fix. Am I missing something?

Member

andrew commented Jan 30, 2014

@mzgol I'm getting a lot more failures than that locally, I've just pushed the latest libsass to master, are you getting the same results with that?

Contributor

mgol commented Jan 30, 2014

@andrew The same.

EDIT: FWIW, my setup is: OS X 10.9.1, node v0.10.25, npm 1.3.25.

Member

andrew commented Jan 30, 2014

@mzgol yeah getting the same on travis, so it must be my local setup now.

Contributor

mgol commented Jan 30, 2014

@andrew \o/ Does that mean it can be updated after fixing this one test? :) I can submit a PR though it's trivial enough you might want to do it by yourself.

Contributor

mgol commented Jan 30, 2014

@andrew You've changed the wrong file. ;)

Member

andrew commented Jan 30, 2014

Just pushed a fix for that test, we'll need to get binaries updated for mac, linux and windows then we can push a new release.

Contributor

mgol commented Jan 30, 2014

@andrew I fixed the errors correctly in #235.

Contributor

nschonni commented Jan 30, 2014

👍 Awesome @mzgol, but there are still failing @extend tests in libsass https://travis-ci.org/hcatlin/libsass/jobs/17783813#L252-L260 which is what I was referring to before. These are new tests from @akhleung but I'm not sure if the fixes are going to break people's currently compiling Sass.

Contributor

mgol commented Jan 30, 2014

@nschonni Ah, OK, I get it. Sure, no rush with the release if it might break someone's code. I'll live on master for now.

Contributor

nschonni commented Jan 30, 2014

It may be fine, I'm just not sure the best way to test the improvements.

PS: I do want this for Bootstrap-sass too 😉

nschonni referenced this issue in wet-boew/wet-boew Jan 30, 2014

Closed

Evaluate Bootstrap 3.1.0 release #4448

Contributor

mgol commented Feb 3, 2014

What steps are needed to re-build the binaries? I assume just putting andrew/node-sass#master won't fly because of the build step needed?

I've read https://github.com/andrew/node-sass#rebuilding-binaries but there is no precompiled folder; also, it's not clear to me what string exactly does *your-platform* represent. Should it be named like in the https://github.com/andrew/node-sass-binaries repository, e.g. darwin-x64-vVERSION? Are there any constraints on what VERSION can be?

I feel README could be a little improved in this area. ;)

EDIT: It seems binaries are generated into the bin directory so it's enough to do git add -f bin/*-v8-* and commit on a branch.

Anyway, @nschonni, you're right about these extend tests. I compared output from original Sass & libsass in our large project and extends are completely messed up. I guess we'll have to wait longer until libsass fixes those errors.

Member

andrew commented Feb 4, 2014

@mzgol yeah it definitely needs updating, I'll see if I can make some time for it later in the week.

Member

LaurentGoderre commented Feb 4, 2014

Let me know when you need the Windows one.

ocombe commented Feb 23, 2014

Any news on this ?
When I build the master it seems to be working fine for me on windows, but an official release would be nice so that I can include it in my package.json

Contributor

nschonni commented Feb 23, 2014

libsass still has failures in master around the @extend feature

ocombe commented Feb 23, 2014

oh well, I'll stick with jlong's bootstrap port a little bit longer then since I can't use the official bootstrap port :(

Member

andrew commented Feb 25, 2014

I'm in the middle of moving house at the moment and don't have any free time to work on a new release but if anyone else is interested in pushing it forward I'm happy to give them commit access.

Pradeek referenced this issue in stefanpenner/ember-app-kit Feb 25, 2014

Closed

Proposal: Replace ruby based sass/compass for node-sass #514

Contributor

nschonni commented Feb 25, 2014

libsass is down to 2 failing @extend tests, but one is a rather important. It still fails for the nested extend.

%x {
  width: 100px;

  %y {
    height: 100px;
  }
}

.foo {
  @extend %x;

  .bar { @extend %y }
}

nschonni referenced this issue in sass/libsass Feb 25, 2014

Closed

SCSS compiler bug #297

Hmm, I don't think that case ever worked in LibSass -- I'm not sure how that test got moved out of the to-do folder. Given that @extend works better now than before (not to mention all of the other bug fixes), I would encourage everyone to use the latest master.

Contributor

nschonni commented Feb 25, 2014

It was added 4 months ago when libsass was still passing the suite
sass/sass-spec@a5b0ff9

genexp commented Mar 3, 2014

Thanks folks for the hard work. I'm keeping an eye on this thread too, as I'm blocked by several issues fixed in the latest libsass, namely: hcatlin/libsass#305 fixed broken scss spritesheets generated by the node spritesmith module.

👍 Thanks for your hard work. Excited to see this working again...having the same issue as @theaqua mentioned above with bootstrap-official-sass, but also documented (strangely) in the grunt-sass repo:

https://github.com/sindresorhus/grunt-sass/issues/33#issuecomment-37377580

Member

andrew commented Mar 12, 2014

I've just published v0.8.3 of node-sass which should solve some of these issues, there is still one failing test with @extend but it includes fixes for lots of other things: https://github.com/andrew/node-sass/releases/tag/v0.8.3

andrew closed this Mar 12, 2014

genexp commented Mar 12, 2014

Hey @andrew - You're awesome. Also, haven't seen you since HybridConf.
Cheers bro!

On Wed, Mar 12, 2014 at 1:56 PM, Andrew Nesbitt notifications@github.comwrote:

Closed #233 andrew#233.

Reply to this email directly or view it on GitHubhttps://github.com/andrew/node-sass/issues/233
.

jtribble commented Dec 5, 2014

Not sure if this is helpful for anyone, but figured I'd post just in case it's useful to someone!

I had a very similar error to @mzgol when attempting to run my gulp-sass task:

gulp(5396,0x7fff70715300) malloc: *** error for object 0x10489ee44: pointer being freed was not allocated
*** set a breakpoint in malloc_error_break to debug
Abort trap: 6

After some debugging, I figured out that this error was triggered because I had an empty .scss file that I was running through the gulp-sass task. Removing the empty .scss file (or adding text to it) fixed the error.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment