testtools-dev team mailing list archive

[Jonathan Lange <jml@xxxxxxxxx>]

Hello all,

tl;dr: What do you think about our balance between rapid release cycles and
quality & trustworthiness of our releases?

I'm pretty sure that every one of us who has done a testtools release has
also done a bogus release – a release that had to be followed up almost
immediately by another one to fix some bug.

Since our process is fairly fast & automated, and since we're by-and-large
OK with post-merge reviews for these problems, it's not so much of an
issue.

However, it's likely that an unusable release erodes user trust a little,
and trust is notoriously hard to regain.

What do you reckon?

jml