Frank Büttner wrote:
I have done some like at my first package, and then to my was
spoken,
that I use local mock and not the infrastructure. But after some updates
of mock, mock it now dead. The no response from them maintainer.
Don't worry, even I get away with my intensive use of Koji. ;-) I rarely do
local mock builds (mostly only if I build stuff for
repo.calcforge.org or if
I can't use a Koji build for some reason, like not having the right stuff in
the buildroot), I just send everything straight to Koji and I maintain more
stuff than you. ;-)
FWIW, I usually only scratch build if the package is pending review.
Otherwise, I just do this routine:
* commit change
* make tag
* make build BUILD_FLAGS=--nowait
* while (build failed) {
* commit fix
* make tag TAG_OPTS=-F
* koji resubmit taskid --nowait
}
The advantage over scratch builds is: if you scratch-build, once it's
successful, you have to redo the build as a real build, so you waste one
build. With my workflow, you save that build. Saves both Koji resources and
your time. Also saves your time over local mock builds where you also have
this "one wasted build" problem.
Kevin Kofler