Cleanup the test and fix mocks. Workaround pprint flakiness.

Cleanup the test and fix mocks. Workaround pprint flakiness.

TEST=unit tests
BUG=none

[M-A Ruel, 2009-11-10 18:29:34]

Reviewers: dpranke,

Description:
Cleanup the test and fix mocks. Workaround pprint flakiness.

TEST=unit tests
BUG=none

Please review this at http://codereview.chromium.org/387007

Affected files:
   M tests/gclient_test.py

[Dirk Pranke, 2009-11-10 19:53:38]

lgtm, although I don't know pymox and so I can really only spot syntactic stuff
...

http://codereview.chromium.org/387007/diff/1/2
File tests/gclient_test.py (right):

http://codereview.chromium.org/387007/diff/1/2#newcode351
Line 351: gclient.os.path.join(self.root_dir, options.config_filename),
at some point I need to learn how pxmox works ... is this really gclient.os ?

http://codereview.chromium.org/387007/diff/1/2#newcode695
Line 695: # pprint.pformat() is non-deterministic in this case!!
odd. is that a python bug?

[Dirk Pranke, 2009-11-10 19:53:56]

lgtm, although I don't know pymox and so I can really only spot syntactic  
stuff
...


http://codereview.chromium.org/387007/diff/1/2
File tests/gclient_test.py (right):

http://codereview.chromium.org/387007/diff/1/2#newcode351
Line 351: gclient.os.path.join(self.root_dir, options.config_filename),
at some point I need to learn how pxmox works ... is this really
gclient.os ?

http://codereview.chromium.org/387007/diff/1/2#newcode695
Line 695: # pprint.pformat() is non-deterministic in this case!!
odd. is that a python bug?

http://codereview.chromium.org/387007

[M-A Ruel, 2009-11-10 20:07:12]

http://code.google.com/p/pymox/

It simplifies "calls expectation" type of unit tests. It has integrated
mocking/unmocking support which helps a lot.

http://codereview.chromium.org/387007/diff/1/2
File tests/gclient_test.py (right):

http://codereview.chromium.org/387007/diff/1/2#newcode351
Line 351: gclient.os.path.join(self.root_dir, options.config_filename),
On 2009/11/10 19:53:38, dpranke wrote:
> at some point I need to learn how pxmox works ... is this really gclient.os ?

Actually, this is so we don't have multiple entry point for os. So the unit test
doesn't import os but use the import from gclient.py instead. Hence
gclient.os.path...

This is unrelated to pymox. :)

http://codereview.chromium.org/387007/diff/1/2#newcode695
Line 695: # pprint.pformat() is non-deterministic in this case!!
On 2009/11/10 19:53:38, dpranke wrote:
> odd. is that a python bug?

Yes.

[M-A Ruel, 2009-11-10 20:07:30]

http://code.google.com/p/pymox/

It simplifies "calls expectation" type of unit tests. It has integrated
mocking/unmocking support which helps a lot.


http://codereview.chromium.org/387007/diff/1/2
File tests/gclient_test.py (right):

http://codereview.chromium.org/387007/diff/1/2#newcode351
Line 351: gclient.os.path.join(self.root_dir, options.config_filename),
On 2009/11/10 19:53:38, dpranke wrote:
> at some point I need to learn how pxmox works ... is this really
gclient.os ?

Actually, this is so we don't have multiple entry point for os. So the
unit test doesn't import os but use the import from gclient.py instead.
Hence gclient.os.path...

This is unrelated to pymox. :)

http://codereview.chromium.org/387007/diff/1/2#newcode695
Line 695: # pprint.pformat() is non-deterministic in this case!!
On 2009/11/10 19:53:38, dpranke wrote:
> odd. is that a python bug?

Yes.

http://codereview.chromium.org/387007