rebaseline_server: clean up thread locks

rebaseline_server: clean up thread locks
followup to https://codereview.chromium.org/66803004/ ('rebaseline_server: improve thread locks to allow read access during updates')
(SkipBuildbotRuns)

R=jcgregorio@google.com

Committed: https://code.google.com/p/skia/source/detail?r=12323

[epoger, 2013-11-14 20:21:55]

Ready for review at patchset 1.

https://codereview.chromium.org/64273011/diff/1/gm/rebaseline_server/server.py
File gm/rebaseline_server/server.py (right):

https://codereview.chromium.org/64273011/diff/1/gm/rebaseline_server/server.p...
gm/rebaseline_server/server.py:109: self._actuals_repo = svn.Svn(actuals_dir)
Now, the self._actuals_repo object is shared between threads, and is in charge
of synchronizing access to itself.  (Same for self._expectations_repo)

https://codereview.chromium.org/64273011/diff/1/gm/rebaseline_server/server.p...
gm/rebaseline_server/server.py:159: self.results = results.Results(
I *think* this is fine, because I assume self.results is not pointed away from
the old Results object and at the new Results object until the new one is
completely formed.

Any idea where I should look in Python's language specification for
documentation of that behavior?

[jcgregorio, 2013-11-18 16:50:35]

https://codereview.chromium.org/64273011/diff/1/gm/rebaseline_server/server.py
File gm/rebaseline_server/server.py (right):

https://codereview.chromium.org/64273011/diff/1/gm/rebaseline_server/server.p...
gm/rebaseline_server/server.py:123: if not os.path.isdir(expectations_dir):
Duplicated code from above, break out into a function.

https://codereview.chromium.org/64273011/diff/1/gm/rebaseline_server/server.p...
gm/rebaseline_server/server.py:159: self.results = results.Results(
Yes, assignment won't happen until the RHS is evaluated completely.

On 2013/11/14 20:21:56, epoger wrote:
> I *think* this is fine, because I assume self.results is not pointed away from
> the old Results object and at the new Results object until the new one is
> completely formed.
> 
> Any idea where I should look in Python's language specification for
> documentation of that behavior?

[epoger, 2013-11-20 19:51:34]

Thanks, PTAL.

https://codereview.chromium.org/64273011/diff/1/gm/rebaseline_server/server.py
File gm/rebaseline_server/server.py (right):

https://codereview.chromium.org/64273011/diff/1/gm/rebaseline_server/server.p...
gm/rebaseline_server/server.py:123: if not os.path.isdir(expectations_dir):
On 2013/11/18 16:50:35, jcgregorio wrote:
> Duplicated code from above, break out into a function.

Fixed in patchset 2.

https://codereview.chromium.org/64273011/diff/60001/gm/rebaseline_server/serv...
File gm/rebaseline_server/server.py (right):

https://codereview.chromium.org/64273011/diff/60001/gm/rebaseline_server/serv...
gm/rebaseline_server/server.py:71: def _get_routable_ip_address():
made this function private (it's only used internally; let's not encourage
others to use it and thus be on the hook for maintaining it)

[jcgregorio, 2013-11-20 19:53:50]

On 2013/11/20 19:51:34, epoger wrote:
> Thanks, PTAL.
> 
> https://codereview.chromium.org/64273011/diff/1/gm/rebaseline_server/server.py
> File gm/rebaseline_server/server.py (right):
> 
>
https://codereview.chromium.org/64273011/diff/1/gm/rebaseline_server/server.p...
> gm/rebaseline_server/server.py:123: if not os.path.isdir(expectations_dir):
> On 2013/11/18 16:50:35, jcgregorio wrote:
> > Duplicated code from above, break out into a function.
> 
> Fixed in patchset 2.
> 
>
https://codereview.chromium.org/64273011/diff/60001/gm/rebaseline_server/serv...
> File gm/rebaseline_server/server.py (right):
> 
>
https://codereview.chromium.org/64273011/diff/60001/gm/rebaseline_server/serv...
> gm/rebaseline_server/server.py:71: def _get_routable_ip_address():
> made this function private (it's only used internally; let's not encourage
> others to use it and thus be on the hook for maintaining it)

LGTM

[epoger, 2013-11-20 19:58:10]

Committed patchset #2 manually as r12323 (presubmit successful).