On Wed, 2020-09-30 at 13:56:47 +0200, Mattia Rizzolo wrote: > On Wed, Sep 30, 2020 at 11:23:43AM +0200, Christian Kastner wrote: > > To be honest, as a reader, I found that to be the opposite. The "If > > [epoch] is omitted" makes it sound as if there were an alternative > > handling if it's not omitted. > > > > So the text > > > > If it is omitted then the upstream_version may not contain any colons > > > > actually means > > > > The upstream_version may not contain any colons > > If my memory serves correctly¹, this is just a historic remnant, as > colons used to be allowed if the epoch was present (i.e., a version > string "1:2.3:abc" used to be valid). > > > ¹ and I think it does: > https://salsa.debian.org/dbnpolicy/policy/-/commit/918cac858424739a5af269d993e4cadfab285b29 > > > So, yes. I think it would be good to make the wording just clearer, > instead of carrying over some previous syntax from when the rules were > different.
Yes, more so given that this is a Debian policy specific restriction, whereas dpkg does still treat such versions as valid. Thanks, Guillem
