Fix GetNSExecutablePath for running tests with relative path

Fix GetNSExecutablePath for running tests with relative path

When running tests (zip_unittests.cc for instance) using
base::DIR_SOURCE_ROOT with ./<test_executable_name> they will get
incorrect directory names because ./ will be included in the return
value of GetNSExecutablePath() and it will be consider as one level
of parent directory by FilePath::DirName(). To avoid this problem we
should alway convert the value returned by GetNSExecutablePath() into
absolute path first.

BUG=258846
TEST=see bug

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=215882

[Jiang Jiang, 2013-07-10 12:16:52]

We found it while running zip_unittests while developing Opera, I think Chromium
must have the same problem so it would be nice to fix it in upstream.

[Nico, 2013-07-10 18:02:31]

lgtm, please add "TEST=see bug" to cl description

https://codereview.chromium.org/18309008/diff/1/base/base_paths_mac.mm
File base/base_paths_mac.mm (right):

https://codereview.chromium.org/18309008/diff/1/base/base_paths_mac.mm#newcode39
base/base_paths_mac.mm:39: *path =
base::MakeAbsoluteFilePath(base::FilePath(executable_path));
Hm, are you sure these paths never get persisted somewhere? Maybe callers of
this should normpath() instead?

…looks like win and linux return full paths too, so I suppose this is fine.

[commit-bot: I haz the power, 2013-07-11 08:15:06]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/jiangj@opera.com/18309008/1

[Jiang Jiang, 2013-07-12 08:37:32]

On 2013/07/11 08:15:06, I haz the power (commit-bot) wrote:
> CQ is trying da patch. Follow status at
> https://chromium-status.appspot.com/cq/jiangj%40opera.com/18309008/1

Still waiting for an OWNER to say LGTM :)

[Nico, 2013-07-15 20:54:19]

On 2013/07/12 08:37:32, Jiang wrote:
> On 2013/07/11 08:15:06, I haz the power (commit-bot) wrote:
> > CQ is trying da patch. Follow status at
> > https://chromium-status.appspot.com/cq/jiangj%2540opera.com/18309008/1
> 
> Still waiting for an OWNER to say LGTM :)

When you ping people, the best form is "someone: ping". In this case:

brettw: ping.

(…but lots of people were on vacation the last 2 weeks and now have a pile of
mail, so it's possible it'll take brettw a day or two to get to this)

[Jiang Jiang, 2013-07-30 10:53:41]

On 2013/07/15 20:54:19, Nico wrote:
> On 2013/07/12 08:37:32, Jiang wrote:
> > On 2013/07/11 08:15:06, I haz the power (commit-bot) wrote:
> > > CQ is trying da patch. Follow status at
> > > https://chromium-status.appspot.com/cq/jiangj%252540opera.com/18309008/1
> > 
> > Still waiting for an OWNER to say LGTM :)
> 
> When you ping people, the best form is "someone: ping". In this case:
> 
> brettw: ping.
> 
> (…but lots of people were on vacation the last 2 weeks and now have a pile of
> mail, so it's possible it'll take brettw a day or two to get to this)

brettw: ping

[brettw, 2013-08-06 04:52:06]

LGTM unless you can think of something better

https://codereview.chromium.org/18309008/diff/1/base/base_paths_mac.mm
File base/base_paths_mac.mm (right):

https://codereview.chromium.org/18309008/diff/1/base/base_paths_mac.mm#newcode39
base/base_paths_mac.mm:39: *path =
base::MakeAbsoluteFilePath(base::FilePath(executable_path));
We do want an absolute path here. We should be sure this is the best solution. I
think MakeAbsoluteFilePath does I/O on Posix, which we'd want to avoid if
possible (since I think this function is called on the UI thread). But I don't
have any better ideas.

I think if we're sure this is best the comment would be better worded to just
say that we expect absolute paths to be returned here, since I think DirName()
is probably doing the right thing with the input we're giving it (or it not, we
should fix it).

[commit-bot: I haz the power, 2013-08-06 08:42:02]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/jiangj@opera.com/18309008/19001

[commit-bot: I haz the power, 2013-08-06 13:30:28]

Change committed as 215882