webrtc/tools/quality_tracking/README
phoglund@webrtc.org d4f0a0e2bc Refactored the dashboard in order to add new functionality and added some new functionality.
Note that all files were moved to a new directory. The diffs won't be 100% friendly because of this.

Extracted common handling for OAuth on both sides of the connection in order to add a new build status
data handler. This data handler will be used to report build status data. Don't look too closely at the
details of what data is transferred as this will change in the next patch. We will also extract data from
a different page in a slightly different way, but there won't be huge differences.

In particular, we won't look at the /one_box_per_builder page on the master but rather at the transposed
grid (/tgrid) on the build master since we also need the revision number. The regular expressions to
extract the data will be slightly more complex.

BUG=
TEST=

Review URL: https://webrtc-codereview.appspot.com/367023

git-svn-id: http://webrtc.googlecode.com/svn/trunk@1586 4adac7df-926f-26a2-2b94-8c16560cd09d
2012-02-01 10:59:23 +00:00

31 lines
1.6 KiB
Plaintext

This file describes the coverage tracking script and the coverage dashboard.
ABSTRACT:
The intention of this small tracking system is to track code coverage data
over time. Since code coverage is continuously recomputed on the build bots,
the track_coverage.py script is intended to run on the build bot as a cron job
and extract the data from there. The dashboard doesn't care how often this
script runs, but running each hour should be more than enough.
The track_coverage.py script uses OAuth to authenticate itself. In order to do
this, it needs two files: consumer.secret and access.token. The consumer secret
is known within the organization and is stored in a plain file on the bot
running the scripts (we don't want to check in this secret in the code in the
public repository). The consumer secret is a plain file with a single line
containing the secret string.
The access.token file is generated by request_oauth_permission.py. It does this
by going through the three-legged OAuth authorization process. An administrator
of the dashboard must approve the request from the script. Once that is done,
access.token will be written and track_coverage.py will be able to report
results.
HOW TO RUN LOCALLY:
Follow the following instructions:
http://code.google.com/appengine/docs/python/gettingstartedpython27/devenvironment.html
The dashboard can be started on 127.0.0.1:8080 using the dev_appserver.py script
as described in the above URL (and in the following 'hello world' page).
HOW TO DEPLOY:
Follow the following instructions:
http://code.google.com/appengine/docs/python/gettingstartedpython27/uploading.html