[Development] Speeding up the review process (was: PostgreSQL cross compile for Pi)
edward.welbourne at qt.io
Fri Oct 13 16:25:08 CEST 2017
Sérgio Martins (13 October 2017 16:05)
> Some users have been complaining about the review process and have
> rotting patches, so I welcome brainstorming around this. Let's see if
> we can conclude improvements!
Indeed - and the remedy for that is, painfully enough, that we, as
developers, need to spend more time doing review.
Change review has been a recommended practice in our industry for
several decades; in the last decade, I've seen it actually become a
*standard* practice, too, which is most welcome. However, as an
industry, we're so used to it not being a standard practice that we
aren't yet used to allocating to it the time it actually needs. We all
just want to "get on with real work" - by which developers tend to mean:
hacking on code, making our own changes. For code review to realise its
full benefits, we have to learn to make time for it - to count it as
part of the "real work".
While refinements to process might help - as might a more modern version
of Gerrit - there is no cure for "rotting patches" other than that *we*
review those patches. That takes time - and takes it away from the
other parts of our work.
Relaxing the rules is relatively incidental: giving time to this crucial
part of our work is the necessary part of solving the problem,
More information about the Development