Just a quick announcement that as of today, Treeherder no longer uses
Datasource (
https://github.com/jeads/datasource/). This marks the end of
the per-project databases: all Treeherder data can now be retrieved
through the Django ORM.
Along with this change, I have completely removed Treeherder's opaque
and hard-to-understand "job model": the last remaining use of it
(ingesting data) is now done through some relatively easy-to-follow
functions in `treeherder/etl/jobs.py` and `treeherder/etl/resultset.py`.
Getting here was quite the long road, spanning well over a year of
effort. I might talk more about what motivated this change in the
future, but for now please just enjoy what I hope is much a simpler and
better-documented Treeherder development experience. :)
Will