Fixed debug mode

Fixed debug mode

BUG=420727
R=dnj@chromium.org

Committed: https://chromium.googlesource.com/infra/testing/expect_tests/+/ced75d2e9aef3d124129442ae4d7727bb770dbec

[pgervais, 2014-12-02 19:17:55]

pgervais@chromium.org changed reviewers:
+ dnj@chromium.org, iannucci@chromium.org

[pgervais, 2014-12-02 19:17:56]

And the debug mode was broken as well. Here's the fix.

[dnj, 2014-12-02 20:47:26]

https://chromiumcodereview.appspot.com/774783002/diff/1/expect_tests/pipeline.py
File expect_tests/pipeline.py (right):

https://chromiumcodereview.appspot.com/774783002/diff/1/expect_tests/pipeline...
expect_tests/pipeline.py:333: with use_chdir(cwd):
Is the reason for this that the 'chdir' is necessary for 'run_ctx' to work? Is
it reasonable to have 'run_ctx' perform the 'chdir'? (If so, rename it to
something more obviously associated with chdir, such as 'run_ctx_in_chdir')

[pgervais, 2014-12-02 21:17:00]

https://chromiumcodereview.appspot.com/774783002/diff/1/expect_tests/pipeline.py
File expect_tests/pipeline.py (right):

https://chromiumcodereview.appspot.com/774783002/diff/1/expect_tests/pipeline...
expect_tests/pipeline.py:333: with use_chdir(cwd):
On 2014/12/02 20:47:26, dnj wrote:
> Is the reason for this that the 'chdir' is necessary for 'run_ctx' to work? Is
> it reasonable to have 'run_ctx' perform the 'chdir'? (If so, rename it to
> something more obviously associated with chdir, such as 'run_ctx_in_chdir')

TBH, I don't know. There's only an issue with the debugger, nothing else. It
does not seem to be related to the chdir but rather to the coverage, because I
remember seeing the same issue in the past. I'll try to invert chdir and run_ctx
to see if it still works.

[pgervais, 2014-12-02 21:32:06]

https://chromiumcodereview.appspot.com/774783002/diff/1/expect_tests/pipeline.py
File expect_tests/pipeline.py (right):

https://chromiumcodereview.appspot.com/774783002/diff/1/expect_tests/pipeline...
expect_tests/pipeline.py:333: with use_chdir(cwd):
On 2014/12/02 21:17:00, pgervais wrote:
> On 2014/12/02 20:47:26, dnj wrote:
> > Is the reason for this that the 'chdir' is necessary for 'run_ctx' to work?
Is
> > it reasonable to have 'run_ctx' perform the 'chdir'? (If so, rename it to
> > something more obviously associated with chdir, such as 'run_ctx_in_chdir')
> 
> TBH, I don't know. There's only an issue with the debugger, nothing else. It
> does not seem to be related to the chdir but rather to the coverage, because I
> remember seeing the same issue in the past. I'll try to invert chdir and
run_ctx
> to see if it still works.

Yep, it works with use_chdir at the lowest level. I'll send another patchset.

[dnj, 2014-12-03 00:02:32]

Thanks for looking into that. LGTM!

[pgervais, 2014-12-03 00:12:06]

Committed patchset #2 (id:20001) manually as
ced75d2e9aef3d124129442ae4d7727bb770dbec (presubmit successful).