Add merge-to-branch python port.

Add merge-to-branch python port.

- To ease a line-by-line review, the script is intentionally close to the former bash version
- Disambiguate the existing "-r" option for reviewer in the other scripts
- The options design will be refactored in a follow up CL

TEST=python -m unittest test_scripts.ScriptTest.testMergeToBranch
R=ulan@chromium.org

Committed: https://code.google.com/p/v8/source/detail?r=19443

[Michael Achenbach, 2014-02-17 11:29:07]

PTAL

[Michael Achenbach, 2014-02-18 08:52:55]

[+Ulan]

[ulan, 2014-02-18 11:15:08]

https://codereview.chromium.org/163183004/diff/150001/tools/push-to-trunk/mer...
File tools/push-to-trunk/merge_to_branch.py (right):

https://codereview.chromium.org/163183004/diff/150001/tools/push-to-trunk/mer...
tools/push-to-trunk/merge_to_branch.py:103: for revision in self._options.args:
How about building full_revision_list from args first and then iterating over 
full_revision_list to find port_revision?

https://codereview.chromium.org/163183004/diff/150001/tools/push-to-trunk/mer...
tools/push-to-trunk/merge_to_branch.py:118:
full_revision_list.append(svn_revision)
See comment below, appending here can be avoided.

https://codereview.chromium.org/163183004/diff/150001/tools/push-to-trunk/mer...
tools/push-to-trunk/merge_to_branch.py:135: full_revision_list =
self._options.args
It seems cleaner to append port_revision_list on 'y' instead of restoring
full_revision_list on 'n'. Also it doesn't remove duplicates anymore.

https://codereview.chromium.org/163183004/diff/150001/tools/push-to-trunk/mer...
tools/push-to-trunk/merge_to_branch.py:186: self._state["patch_commit_hashes"] =
patch_commit_hashes
Why isn't it persisted and restored like other state variables?

[Michael Achenbach, 2014-02-18 12:39:31]

https://codereview.chromium.org/163183004/diff/150001/tools/push-to-trunk/mer...
File tools/push-to-trunk/merge_to_branch.py (right):

https://codereview.chromium.org/163183004/diff/150001/tools/push-to-trunk/mer...
tools/push-to-trunk/merge_to_branch.py:103: for revision in self._options.args:
On 2014/02/18 11:15:09, ulan wrote:
> How about building full_revision_list from args first and then iterating over 
> full_revision_list to find port_revision?

Done.

https://codereview.chromium.org/163183004/diff/150001/tools/push-to-trunk/mer...
tools/push-to-trunk/merge_to_branch.py:118:
full_revision_list.append(svn_revision)
On 2014/02/18 11:15:09, ulan wrote:
> See comment below, appending here can be avoided.

Done.

https://codereview.chromium.org/163183004/diff/150001/tools/push-to-trunk/mer...
tools/push-to-trunk/merge_to_branch.py:135: full_revision_list =
self._options.args
On 2014/02/18 11:15:09, ulan wrote:
> It seems cleaner to append port_revision_list on 'y' instead of restoring
> full_revision_list on 'n'. Also it doesn't remove duplicates anymore.

Done.

https://codereview.chromium.org/163183004/diff/150001/tools/push-to-trunk/mer...
tools/push-to-trunk/merge_to_branch.py:186: self._state["patch_commit_hashes"] =
patch_commit_hashes
On 2014/02/18 11:15:09, ulan wrote:
> Why isn't it persisted and restored like other state variables?

Just to be similar with the bash script. Improved now. In a follow up CL I'll
replace the whole persist/restore logic of the scripts with one saved JSON dict.

https://codereview.chromium.org/163183004/diff/150001/tools/push-to-trunk/tes...
File tools/push-to-trunk/test_scripts.py (right):

https://codereview.chromium.org/163183004/diff/150001/tools/push-to-trunk/tes...
tools/push-to-trunk/test_scripts.py:857: TextToFile("1999-04-05: Version
3.22.4", TEST_CONFIG[CHANGELOG_FILE])
Removed also some c/p relics.

https://codereview.chromium.org/163183004/diff/210001/tools/push-to-trunk/tes...
File tools/push-to-trunk/test_scripts.py (right):

https://codereview.chromium.org/163183004/diff/210001/tools/push-to-trunk/tes...
tools/push-to-trunk/test_scripts.py:853:
TEST_CONFIG[ALREADY_MERGING_SENTINEL_FILE] = self.MakeEmptyTempFile()
Added sentinel file logic to the test. Otherwise you can't execute the test
twice in a row.

https://codereview.chromium.org/163183004/diff/210001/tools/push-to-trunk/tes...
tools/push-to-trunk/test_scripts.py:863: msg = """Merged r12345, r23456, r34567,
r45678, r56789 into trunk branch.
The changes to the "find MIPS ports" algorithm lead to a different order of the
patches.

https://codereview.chromium.org/163183004/diff/210001/tools/push-to-trunk/tes...
tools/push-to-trunk/test_scripts.py:962: options = MakeOptions(p=extra_patch,
f=True)
Deletes the sentinel file on each run.

[ulan, 2014-02-18 13:02:26]

lgtm

[Michael Achenbach, 2014-02-18 13:09:22]

Committed patchset #3 manually as r19443 (presubmit successful).