On Wed, Oct 26, 2016 at 4:25 PM, Matthew Miller <mattdm@xxxxxxxxxxxxxxxxx> wrote: > On Tue, Oct 25, 2016 at 07:37:32PM -0600, Kevin Fenzi wrote: >> Nope. We have talked about having some kind of fast track, but IMHO, we >> should just get the normal process faster. > > Getting the normal process faster would help in a large number of other > areas. Right now, when we have issues in a spin or other deliverable, > we have to first update the RPM, wait for it to get pushed to testing, > wait for it to get pushed to stable, and then wait for a new release > artifact (iso, qcow, whatever) with that RPM included. This means it > can take several days from fix to testing the fix, and if it takes > several tries to get right, it can easily take a week to address a > simpel problem. It'd be nice if we could reduce that turn-around time > to hours, if not minutes. If it takes several goes to get right it's clearly not a simple problem! And for RCs rel-eng and QA have a process that circumvents all of the testing phase to get it straight into the compose so most of what you outline there for delays is actually garbage. We don't use this process for nightlies because they're not meant to be fire drills so should follow the right process. Peter _______________________________________________ devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx
