@dcallagh - @mjia contributed the original tag-release script. Any opinion on this?

Discussed this with Lukas on Bluejeans... There are some nice improvements in here. But my preference is to keep these helper scripts like tag-release.sh as simple and "dumb" as possible. Otherwise we risk reinventing yet another tito (or rpkg-util, or...)

I think the original motivation here was to prevent us accidentally tagging a release with a "Next Release" heading still present in the release notes. We could do that with a little extra check like:

if grep -q 'Next Release' docs/release-notes.rst ; then
    echo "You forgot to update the release notes"
    exit 1
fi

Simple alternative is PR#240.

I don't like doing anything extra when releasing a software because it's so error prone. Last time I remember messing up was putting wrong release date (where the hell did YYYY/dd/MM come from) into one file and one late part of my release pipeline failed. I wrote a script so this won't happen again and I have to manually update less things.

Therefore I like to automate everything and have as much stuff verified as possible.

That's totally fair.

We can look at moving Greenwave and Waiverdb to use tito, or one of the other RPM-based releasing tools, if you want. I just don't want us to reinvent our own version of them :-)