Script to archive or replay webpages, generate skp files and store both in Google Storage

Script to archive or replay webpages, generate skp files and store both in Google Storage.

The intent is for webpages to be archived and stored along with GM skp files rarely (maybe once a month or once every 3 months).
The archived webpages will be replayed and skp files generated per revision.

This changeset has been created to fix the following feature requests-
http://code.google.com/p/skia/issues/detail?id=883 : create system to capture HTML source of popular web sites for SkPicture testing.
http://code.google.com/p/skia/issues/detail?id=884 : create semi-automated system to generate SkPicture files from captured HTML source.
Committed: https://code.google.com/p/skia/source/detail?r=6204

[rmistry, 2012-10-24 11:51:31]

Updated the chromium_revision at Patch Set 4 to include
https://codereview.chromium.org/11189123

[rmistry, 2012-10-24 13:29:20]

On 2012/10/24 11:51:31, rmistry wrote:
> Updated the chromium_revision at Patch Set 4 to include
> https://codereview.chromium.org/11189123

Updating the chromium_revision caused missing dependencies and refactored code
in the skpicture_printer code. Fixed them all at Patch Set 6.

[rmistry, 2012-10-26 19:09:50]

Patch Set 7 contains a big refactoring that changed the script to support the
case of replaying archived webpages and to store the generated skp files in
Google Storage.

If --record is passed as an argument into the script then an archive of the page
set is taken and stored in Google Storage along with GM skp files.
If --record is not passed then the archive from Google Storage is downloaded,
replayed and used to generated new skp files.

[epoger, 2012-10-29 16:51:35]

Eric, I'm gonna let you review this one first.  Let me know if that's
problematic, or if there are particular parts you would like me to look at.

[EricB, 2012-10-29 17:26:51]

LGTM with comment/question.

https://codereview.appspot.com/6750057/diff/14001/slave/skia_slave_scripts/sk...
File slave/skia_slave_scripts/skpicture_playback/webpages_playback.py (right):

https://codereview.appspot.com/6750057/diff/14001/slave/skia_slave_scripts/sk...
slave/skia_slave_scripts/skpicture_playback/webpages_playback.py:85:
multipage_benchmark_runner.Main()
Modifying sys.argv and calling Main() seems like a strange way to invoke this
script.  I guess there isn't a good way around it since the script comes from
upstream?

[epoger, 2012-10-30 14:32:31]

LGTM

https://codereview.appspot.com/6750057/diff/14001/DEPS
File DEPS (right):

https://codereview.appspot.com/6750057/diff/14001/DEPS#newcode30
DEPS:30: # build/android/pylib/android_commands.py requires android_testrunner
to be in third_party.
maybe add a blank line before this one

https://codereview.appspot.com/6750057/diff/14001/DEPS#newcode33
DEPS:33: # chrome_remote_control/replay_server.py requires webpagereplay to be
in src/third_party.
maybe add a blank line before this one

[epoger, 2012-10-30 14:33:02]

On 2012/10/30 14:32:31, epoger wrote:
> LGTM

not that you need my LGTM, since you have Eric's   :-)

[rmistry, 2012-10-30 17:17:19]

https://codereview.appspot.com/6750057/diff/14001/DEPS
File DEPS (right):

https://codereview.appspot.com/6750057/diff/14001/DEPS#newcode30
DEPS:30: # build/android/pylib/android_commands.py requires android_testrunner
to be in third_party.
On 2012/10/30 14:32:31, epoger wrote:
> maybe add a blank line before this one

Done.

https://codereview.appspot.com/6750057/diff/14001/DEPS#newcode33
DEPS:33: # chrome_remote_control/replay_server.py requires webpagereplay to be
in src/third_party.
On 2012/10/30 14:32:31, epoger wrote:
> maybe add a blank line before this one

Done.

https://codereview.appspot.com/6750057/diff/14001/slave/skia_slave_scripts/sk...
File slave/skia_slave_scripts/skpicture_playback/webpages_playback.py (right):

https://codereview.appspot.com/6750057/diff/14001/slave/skia_slave_scripts/sk...
slave/skia_slave_scripts/skpicture_playback/webpages_playback.py:85:
multipage_benchmark_runner.Main()
On 2012/10/29 17:26:52, EricB wrote:
> Modifying sys.argv and calling Main() seems like a strange way to invoke this
> script.  I guess there isn't a good way around it since the script comes from
> upstream?

Yes, unfortunately this is the best way I found of invoking
multipage_benchmark_runner.py