grunt will fail using the latest commit of polymer-all

[dinhvh]

I used:

$ git submodule update --recursive

Then, ran grunt.

$ grunt
Running "uglify:Polymer" (uglify) task
Source Map "polymer.min.js.map" created.
File "polymer.min.js" created.

Running "uglify:PolymerNative" (uglify) task
Source Map "polymer.native.min.js.map" created.
File "polymer.native.min.js" created.

Running "uglify:PolymerSandbox" (uglify) task
Source Map "polymer.sandbox.min.js.map" created.
File "polymer.sandbox.min.js" created.

Running "audit:polymer" (audit) task
Warning: fatal: Not a git repository: '/Users/dvh/Polymer/polymer-all/TemplateInstances/.git' Use --force to continue.

Aborted due to warnings.

There was an issue with "TemplateInstances".
I couldn't find that folder or repository.

Before I updated, using grunt used to work.

[sjmiles]

polymer-all is now deprecated, primarily because it was too difficult to keep all the various submodules up to date.

Instead, we recommend you use the pull-all.sh script to access the live git repositories (or use one of the other methods of distribution, i.e. zip file or bower).

The pull-all.sh is pretty nice, you should be able to use it in your existing polymer-all checkout to update everything, although at this point there is really no need for polymer-all.

[dinhvh]

You can update the instructions on http://www.polymer-project.org/getting-the-code.html then.

[dinhvh]

I can't see a "TemplateInstances" repository or is there something I missed?

[dfreedm]

TemplateInstances was renamed to TemplateBinding. The "getting the code" steps say to use git submodule update --init --recursive in polymer-all because the submodule changed folder names and must be reinitialized, which is why the --init flag is important.

[dinhvh]

ok.
I fixed it using the following diff:

diff --git a/gruntfile.js b/gruntfile.js
index ec1a24e..0c25bb3 100644
--- a/gruntfile.js
+++ b/gruntfile.js
@@ -144,7 +144,7 @@ module.exports = function(grunt) {
             '../polymer-expressions',
             '../observe-js',
             '../NodeBind',
-            '../TemplateInstances'
+            '../TemplateBinding'
           ]
         },
         dest: 'build.log',

Now grunt works.

[dfreedm]

Hmm, I wonder how those got so out of sync.

This is a good data point for why polymer-all needs to go away.

[dfreedm]

BTW, the build files are still being generated. This error happens after they are written. Only the build.log file will be missing.

[dinhvh]

gruntfile.js is in polymer repository.
I don't think pull-all.sh scripts would have helped here.

[sjmiles]

Your problem gruntfile.js is simply old. If you use pull-all.sh you will not have old stuff.

Don't use pull-all.sh in polymer-all as I originally suggested, use it in a fresh folder.

[dinhvh]

Here's the output I got when using pull-all.sh:
https://gist.github.com/dinhviethoa/6360311
Is there some arguments I should use when running pull-all.sh?

I got lots of: fatal: Remote branch not found in upstream origin

[sjmiles]

Did you do it from a fresh folder, as per my last instruction?

[dinhvh]

I did the following:

• I created a new folder
• cd into that folder.

$ curl -O https://raw.github.com/Polymer/tools/master/bin/pull-all.sh
$ chmod 755 pull-all.sh
$ ./pull-all.sh

[sjmiles]

That's supposed to work. azakus is looking into it.

[dinhvh]

Thanks!

[sjmiles]

Sorry for all of the trouble today.

[dfreedm]

@dinhviethoa that's some crazy errors. What version of bash do you have?

[dinhvh]

I'm running it on a Mac:

$ bash --version
GNU bash, version 3.2.48(1)-release (x86_64-apple-darwin12)
Copyright (C) 2007 Free Software Foundation, Inc.

[dfreedm]

Ah, found the issue: Polymer/tools@f8fda2e.

A clean clone should work now.

[dinhvh]

It worked!