Fix file manager to better handle missing roots and scan failure

Fix file manager to better handle missing roots and scan failure

On device, FileManager should always see at least one root (Downloads), but when doing development on Linux some engineers may not have any local paths setup.  This CL improves error handling in this case.

Also I realized that our directory scan logic is setup to periodically retry (silently dropping the completion callback on the floor) rather than have any hard failure notification.  I don't know when this could happen in practice, but if it did it could break my redraw reduction optimization. Similarly, if a scan is aborted (eg. because the active root is quickly changed and a new scan is started) we may not get the callback.  I don't see any trivial low-risk solutions to this problem, so given where we are in the schedule I decided it was better to remove this optimization for now rather than risk any further regression.

BUG=chromium-os:27241
TEST=

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=126968

[rginda, 2012-03-15 17:42:09]

LGTM with a nit.

I'm still working on getting my full-commiter bit set, so this lgtm may or may
not actually allow you to commit :/

https://chromiumcodereview.appspot.com/9701063/diff/1/chrome/browser/resource...
File chrome/browser/resources/file_manager/js/directory_model.js (right):

https://chromiumcodereview.appspot.com/9701063/diff/1/chrome/browser/resource...
chrome/browser/resources/file_manager/js/directory_model.js:331: * Cancels
waiting and scheduled rescans and starts new scan.  If the scan
I had been following python documentation guidelines, where comments start with
a single sentence, and the clarifying text in a new paragraph when necessary.

[bshe, 2012-03-15 17:59:17]

On 2012/03/15 17:42:09, rginda wrote:
> LGTM with a nit.
> 
> I'm still working on getting my full-commiter bit set, so this lgtm may or may
> not actually allow you to commit :/
> 
>
https://chromiumcodereview.appspot.com/9701063/diff/1/chrome/browser/resource...
> File chrome/browser/resources/file_manager/js/directory_model.js (right):
> 
>
https://chromiumcodereview.appspot.com/9701063/diff/1/chrome/browser/resource...
> chrome/browser/resources/file_manager/js/directory_model.js:331: * Cancels
> waiting and scheduled rescans and starts new scan.  If the scan
> I had been following python documentation guidelines, where comments start
with
> a single sentence, and the clarifying text in a new paragraph when necessary.

This LGTM should work :)

[Rick Byers, 2012-03-15 19:13:31]

Thanks guys!

https://chromiumcodereview.appspot.com/9701063/diff/1/chrome/browser/resource...
File chrome/browser/resources/file_manager/js/directory_model.js (right):

https://chromiumcodereview.appspot.com/9701063/diff/1/chrome/browser/resource...
chrome/browser/resources/file_manager/js/directory_model.js:331: * Cancels
waiting and scheduled rescans and starts new scan.  If the scan
On 2012/03/15 17:42:09, rginda wrote:
> I had been following python documentation guidelines, where comments start
with
> a single sentence, and the clarifying text in a new paragraph when necessary.

Thanks, Done.  I agree that's better.

[commit-bot: I haz the power, 2012-03-15 19:26:43]

No LGTM from a valid reviewer yet. Only full committers are accepted.
Even if an LGTM may have been provided, it was from a non-committer or
a lowly provisional committer, _not_ a full super star committer.
See http://www.chromium.org/getting-involved/become-a-committer
Note that this has nothing to do with OWNERS files.

[Rick Byers, 2012-03-15 19:31:40]

On 2012/03/15 19:26:43, I haz the power (commit-bot) wrote:
> No LGTM from a valid reviewer yet. Only full committers are accepted.
> Even if an LGTM may have been provided, it was from a non-committer or
> a lowly provisional committer, _not_ a full super star committer.
> See http://www.chromium.org/getting-involved/become-a-committer
> Note that this has nothing to do with OWNERS files.

This should have been enough (bshe is a full committer as of today, and rginda
is an owner).  I'm guessing some cache of full committers is out of date. 
There's probably no point in using the CQ anyway - these are CrOS-specific files
and I've already run CrOS try jobs.  Direct submitting...