So far all of our releases are managed in Jira and the benefit of it is
that it is very easy to see what changed in any version of any module we
released.
(Now, the latest SCR release under vote is the first very unfortunate
exception but that can be fixed in three minutes).
You don't get that by looking at the commit history and even a diff of
two tags is problematic as we have a multi module setup. But then a diff
of code is useless for most users anyway.
The insight of a changelog has a huge benefit for our users as once you
hit a problem you can have a quick assessment whether this might be a
regression or whether this might be already fixed in a newer release.
There are other ways of getting the same, like managing a changelog in
code (which of course can be automated with github actions) - but this
is the most important feature that we need in one way or the other.
Staying with Jira is zero effort (we might need to create a jira ticket
for a PR when there is none - but with one PR per month this is doable).
I am worried that we a) underestimate the effort it needs to get at
least parity on the required features (mainly changelog) and b) the
effort dies somewhere in the middle and we end up with a broken solution
(for whatever reason).
We experienced this in the past several times with such things already.
I think the much bigger issue is that we get PRs and issues - and no one
looks at them - or looks at them once but never follows up. Spending
effort into that has a higher benefit. The best way to report feedback
is worth nothing, if no one acts on the feedback.
Regards
Carsten
--
Carsten Ziegeler
Adobe
cziege...@apache.org
