Thanks for posting the tasks. And by the way thanks to everyone involved
in the eXUG mtg on Friday.
> * that we nominate and communicate a release manager as 'hub' between
> testers and developers
>
great idea.
> * testers should be on #existdb IRC channel to facilitate comms and
> ensure everyone is in sync
>
could arrange for that; on which network is the channel, btw? Haven't
used IRC since the early 2000s, but I'll manage. Anyway, what do you
mean by "in sync"? Is there a need to do the testing simultaneously? Or
do you think in terms of having all the same version of eXist and the
installer, such that things don't get mixed up with several revisions?
> * that results are captured as bugs in bug tracker (or is this too cumbersome?)
>
No, I think this is feasible. I don't know how sf.net deals with bugs,
but with bugzilla based bugtracking it is common practice to open a
release bug entry and make it depend on a bug for each test that has to
be performed. This way it is pretty simple to track when all tests are
completed, that is, when all bugs, this release-specific bug depends on,
are closed. I assume there would be some similar mechanism on sf. I like
the fact that the process and the results are then well documented. I
could imagine, that this reduces the strain on the main developers of
eXist. If there were a release manager, he could open the necessary
bugs. If not, the person doing the test, could do so. Then, only the
release specific bug-report would have to be reported.
Stefan
|
