Housekeeper should not fail when skia-autogen is reset

Housekeeper should not fail when skia-autogen is reset. 

Created to fix https://code.google.com/p/skia/issues/detail?id=1269 : UpdateDoxygen and UploadDoxygen steps fail when skia-autogen repository is reset. 

(SkipBuildbotRuns)

R=borenet@google.com

Committed: https://code.google.com/p/skia/source/detail?r=8935

[borenet, 2013-05-01 12:10:38]

https://codereview.chromium.org/14772003/diff/2001/tools/update-doxygen.sh
File tools/update-doxygen.sh (right):

https://codereview.chromium.org/14772003/diff/2001/tools/update-doxygen.sh#ne...
tools/update-doxygen.sh:22: mkdir docs
Should you "svn add" docs here?  I would think that the "svn status" on line 68
will fail since docs is not yet a working copy.

[rmistry, 2013-05-01 12:15:22]

https://codereview.chromium.org/14772003/diff/2001/tools/update-doxygen.sh
File tools/update-doxygen.sh (right):

https://codereview.chromium.org/14772003/diff/2001/tools/update-doxygen.sh#ne...
tools/update-doxygen.sh:22: mkdir docs
On 2013/05/01 12:10:39, borenet wrote:
> Should you "svn add" docs here?  I would think that the "svn status" on line
68
> will fail since docs is not yet a working copy.

Yes svn status will fail since its not yet a working copy but thats ok because
the changes are actually submitted by merge_into_svn.py, I did not want to break
backwards compatibility of this file if it is used other than the housekeeper.

svn add here will not work as well because the parent directory is not a working
copy.

[borenet, 2013-05-01 13:45:34]

Per our live discussion, I don't like that the behavior is different between
running the script manually and on the bots, but since the future of
skia-autogen is unknown, it probably isn't worth spending time to rewrite this
script.  LGTM.

https://codereview.chromium.org/14772003/diff/2001/tools/update-doxygen.sh
File tools/update-doxygen.sh (right):

https://codereview.chromium.org/14772003/diff/2001/tools/update-doxygen.sh#ne...
tools/update-doxygen.sh:68: NEWFILES=$(svn status | grep ^\? | awk '{print $2}')
Here's what worries me.  If we had to run "mkdir docs" earlier, this command
will fail.

https://codereview.chromium.org/14772003/diff/2001/tools/update-doxygen.sh#ne...
tools/update-doxygen.sh:70: svn add $NEWFILES
Which will cause this command to fail. Now, *none* of the files have been
successfully "svn add"ed, and so none will get committed.  The commit itself
will probably fail with a broken working copy or something.

[rmistry, 2013-05-01 13:49:35]

On 2013/05/01 13:45:34, borenet wrote:
> Per our live discussion, I don't like that the behavior is different between
> running the script manually and on the bots, but since the future of
> skia-autogen is unknown, it probably isn't worth spending time to rewrite this
> script.  LGTM.

Agreed.

> 
> https://codereview.chromium.org/14772003/diff/2001/tools/update-doxygen.sh
> File tools/update-doxygen.sh (right):
> 
>
https://codereview.chromium.org/14772003/diff/2001/tools/update-doxygen.sh#ne...
> tools/update-doxygen.sh:68: NEWFILES=$(svn status | grep ^\? | awk '{print
$2}')
> Here's what worries me.  If we had to run "mkdir docs" earlier, this command
> will fail.
> 
>
https://codereview.chromium.org/14772003/diff/2001/tools/update-doxygen.sh#ne...
> tools/update-doxygen.sh:70: svn add $NEWFILES
> Which will cause this command to fail. Now, *none* of the files have been
> successfully "svn add"ed, and so none will get committed.  The commit itself
> will probably fail with a broken working copy or something.

[rmistry, 2013-05-01 13:49:39]

https://codereview.chromium.org/14772003/diff/2001/tools/update-doxygen.sh
File tools/update-doxygen.sh (right):

https://codereview.chromium.org/14772003/diff/2001/tools/update-doxygen.sh#ne...
tools/update-doxygen.sh:68: NEWFILES=$(svn status | grep ^\? | awk '{print $2}')
On 2013/05/01 13:45:34, borenet wrote:
> Here's what worries me.  If we had to run "mkdir docs" earlier, this command
> will fail.

This command just returns with 'not a working copy'

https://codereview.chromium.org/14772003/diff/2001/tools/update-doxygen.sh#ne...
tools/update-doxygen.sh:70: svn add $NEWFILES
On 2013/05/01 13:45:34, borenet wrote:
> Which will cause this command to fail. Now, *none* of the files have been
> successfully "svn add"ed, and so none will get committed.  The commit itself
> will probably fail with a broken working copy or something.

This also outputs 'not a working copy'.

[rmistry, 2013-05-01 14:47:05]

Committed patchset #2 manually as r8935 (presubmit successful).