diff options
| author | Chris Sosa <sosa@chromium.org> | 2015-02-25 13:29:44 -0800 |
|---|---|---|
| committer | ChromeOS Commit Bot <chromeos-commit-bot@chromium.org> | 2015-02-26 05:12:13 +0000 |
| commit | 75900185ecf0eeae21a524e32944289137f787f0 (patch) | |
| tree | 9eb287641171a56b664e1043b9ae7238a78c6fea /cbuildbot/stages/sync_stages_unittest.py | |
| parent | 5542d1ae69528b6f6bb175b02c161b9ecedc0836 (diff) | |
| download | chromite-75900185ecf0eeae21a524e32944289137f787f0.tar.gz | |
cros flash: fix 3-5 minute regression when cros flashing a device.
As we stage update payloads with devices, there's no need to "create"
a update payload when we already have it. Looks like when we libraryized
the xbuddy method, we regressed functionality such that we always download
the image and generate a payload regardless of whether we have it or not.
This increases the default time to flash a pre-built image by a 3-5 min.
Note, this is caused by just a small mis-use of the XBuddy API.
XBuddy.Get both downloads the artifact while also translating the Path.
XBuddy.Translate just does the latter.
BUG=chromium:461623
TEST=Manual + Unittests. Any suggestions for better testing?
Change-Id: Ie979ae3ae15a636258ceb590bc6d05d4221b13aa
Reviewed-on: https://chromium-review.googlesource.com/253631
Reviewed-by: Chris Sosa <sosa@chromium.org>
Commit-Queue: Chris Sosa <sosa@chromium.org>
Trybot-Ready: Chris Sosa <sosa@chromium.org>
Tested-by: Chris Sosa <sosa@chromium.org>
Diffstat (limited to 'cbuildbot/stages/sync_stages_unittest.py')
0 files changed, 0 insertions, 0 deletions