Update gulp-sass version.

[Garbee]

Upgrading sass builder to try and fix some reports of builds breaking. Issue #1118 for instance.

[addyosmani]

@zeakd could you try this out?

[Garbee]

@addyosmani From their report:

I changed to same gulp-sass version of your project ( ^2.0.4 -> ^1.3.3 ) and it works.

I just changed from ^ on the start to an * on the patch to make sure we only update within patch releases. Minor updates we should re-assess in case of accidental breakage.

[addyosmani]

@Garbee Thanks. I missed that in the original report. Manually verifying locally.

[Garbee]

OH wait, I misread the arrow direction.

They are using 2.0.4 in their project and it was breaking on them. #1075 may have fixed the problems with 2.0.4, so the master will work but 1.0.0 tag still won't.

Either way though, if 2.x works now we should go with it since it does lead to a faster build (plus new goodies.)

[addyosmani]

Works for me locally. Regarding the tag: we're hopefully going to tag and publish 1.0.1 this week. @surma can we get a second verification from you that this doesn't cause any breakage?

[surma]

On it

[googlebot]

We found a Contributor License Agreement for you (the sender of this pull request) and all commit authors, but as best as we can tell these commits were authored by someone else. If that's the case, please add them to this pull request and have them confirm that they're okay with these commits being contributed to Google. If we're mistaken and you did author these commits, just reply here to confirm.

[surma]

npm install on node 0.12 only worked for me after I added phantomjs explicitly to package.json. Added a commit for that.

Microsite, components and templates look fine.

We are breaking node 0.10 btw, which I think we used to platform on.

$ gulp serve
...
Error: `libsass` bindings not found in /Users/surma/src/github.com/google/Material-Design-Lite/node_modules/gulp-sass/node_modules/node-sass/vendor/darwin-x64-14/binding.node. Try reinstalling `node-sass`?

Is node 0.10 still relevant? If not, LGTM.

[Garbee]

Ah, the phantom thing may have been what others were hitting as well. The install/updates worked for me (aside from once), then again I could have had a cache somewhere.

Node 0.10 is years old. Good to not support imo. Whether it is in LTS repos or not, NVM should be used to manage and keep node updated.

[Garbee]

P.S. Googlebot is drunk again.

[surma]

Node 0.10 is years old. Good to not support imo.

Agreed. Good to merge, imo.

[Garbee]

Assigning to Addy for a quick review and merge of he agrees with us dropping 0.10 support.

[addyosmani]

Current Node is 0.12.7 and 0.10.x is pretty old at this point. I did have a question around us opting for 2.0.* vs ^2.0.0, but @Garbee already included a comment about that. Before we merge this, two things:

• Can we verify via VM that this will build on Windows 8.1/10 (per mocha-phantomjs error when building under Windows 8.1 #1089)
• Please include a note in the README calling out that Node 0.12.x will be required to build the source

@Garbee or @surma can either of you look at these items?

[Garbee]

I'll handle it. I can test windows 10 on actual hardware, 8.1 in a vm. A quick note on this should be simple as well.

[Garbee]

Just added a quick blockquote, "MDL requires NodeJS 0.12." underneath the "npm install" step of the instructions. Direct in point and focused where developers building will be looking.

[surma]

👍

[Garbee]

FYI, found this issue on mocha-phantomjs's repo as to why the peerdep change occured. Maybe someone who knows that aspect better can look through real quick and see if there is anything we could improve with our package requirements around it.

[Garbee]

I'll dig into that issue actually. The phantomjs version specified is not satisfying the dependency.

And here we go down the rabbit hole!

[Garbee]

So, seems to work fully on Mac and Linux. It is Windows where we are having problems. But, those don't seem to be related to this update, but instead some phantomjs stuff that happened last week.

This patch functions and is good to merge. We should assess windows breakage on its own since that seems unrelated.

[surma]

Agreed.

[addyosmani]

I'm starting to get more direct reports of the PhantomJS issue arising. Looking at https://github.com/google/material-design-lite/commits/master/package.json, I can't see what could have caused this other than an upstream dep issue @Garbee mentioned.

Merging this and continuing to investigate issues with Phantom.