Docserver: achieve online vs offline (cron vs instance) behaviour at the object

Docserver: achieve online vs offline (cron vs instance) behaviour at the object
store level, not at the caching file system level. This fixes the problem where
the branch data and github data wasn't updating.
Add a whole bunch of tests.

BUG=167432

[not at google - send to devlin, 2013-05-02 16:29:56]

I should really test this...

I'd like to write RenderServletTest and CronServletTest, and this would be a
good opportunity to test effectiveness. But I need to see how far I get today
with the C++ bug work. Ping me before you start reviewing this.

[not at google - send to devlin, 2013-05-02 21:57:42]

I ended up adding a whole bunch of tests, and needed to move some more code
around to make that easier - but in the end failed to actually test the bug
here. Couldn't figure out a meaningful way to do it.

[cduvall, 2013-05-03 02:54:51]

Started reviewing this, not sure what it was but this broke the branch chooser
on preview.py.

[方觉(Fang Jue), 2013-05-03 06:03:31]

https://chromiumcodereview.appspot.com/14856006/diff/8001/chrome/common/exten...
File chrome/common/extensions/docs/server2/future.py (right):

https://chromiumcodereview.appspot.com/14856006/diff/8001/chrome/common/exten...
chrome/common/extensions/docs/server2/future.py:22: if self._value is not
_no_value:
If you're going to use _no_value = (None,), be sure to compare with ==.
(None,) is not (None,)
(None,) == (None,)

[not at google - send to devlin, 2013-05-03 16:04:25]

ugh, thanks Clark. I need to slow down and sit with a pen and paper and figure
out what the dependency graph should look like here, it's getting especially
muddled as I add the svn pinning stuff in a follow-up. I'll get back to with
that. The bug here is because SubversionFileSystem's constructor thing strips
off the branch, but it needs to happen further up the stack.

https://chromiumcodereview.appspot.com/14856006/diff/8001/chrome/common/exten...
File chrome/common/extensions/docs/server2/future.py (right):

https://chromiumcodereview.appspot.com/14856006/diff/8001/chrome/common/exten...
chrome/common/extensions/docs/server2/future.py:22: if self._value is not
_no_value:
On 2013/05/03 06:03:31, 方觉(Fang Jue) wrote:
> If you're going to use _no_value = (None,), be sure to compare with ==.
> (None,) is not (None,)
> (None,) == (None,)

"is not" is correct because the refer to the same value, no? python doesn't copy
by value, it copies by reference?

[cduvall, 2013-05-07 23:47:03]

Ben: should I wait to review this one?

https://chromiumcodereview.appspot.com/14856006/diff/8001/chrome/common/exten...
File chrome/common/extensions/docs/server2/future.py (right):

https://chromiumcodereview.appspot.com/14856006/diff/8001/chrome/common/exten...
chrome/common/extensions/docs/server2/future.py:22: if self._value is not
_no_value:
On 2013/05/03 16:04:25, kalman wrote:
> On 2013/05/03 06:03:31, 方觉(Fang Jue) wrote:
> > If you're going to use _no_value = (None,), be sure to compare with ==.
> > (None,) is not (None,)
> > (None,) == (None,)
> 
> "is not" is correct because the refer to the same value, no? python doesn't
copy
> by value, it copies by reference?

'is not' is correct.