Kevin Daudt <me@xxxxxxxxx> writes: > On Sun, Feb 12, 2017 at 01:15:22PM +0100, Istvan Pato wrote: > >> (master) [1] % git show-ref --tag >> 76c634390... refs/tags/1.0.0 >> b77c7cd17... refs/tags/1.1.0 >> b77c7cd17... refs/tags/1.2.0 >> >> (master) % git describe --tags --always >> 1.1.0-1-ge9e9ced >> >> ### Expected: 1.2.0 >> ... > > Are these lightweight tags? Only annotated tags have a date associated > to them, which is where the rel-notes refers to. Good eyes. The fact that the two points at the same object means that even if they were both annotated tags, they have the same tagger dates. If the code that compares the candidates and picks better tag to describe the object with knows the refnames of these "tags", I'd imagine that we could use the versioncmp() logic as the final tie breaker, but I do not offhand remember if the original refname we took the tag (or commit) from is propagated that deep down the callchain. It probably does not, so some code refactoring may be needed if somebody wants to go in that direction.