On Mon, Mar 27, 2023, at 2:30 PM, Carlos O'Donell wrote: > On 3/27/23 13:45, Sam James wrote: >> "Zack Weinberg" <zack@xxxxxxxxxxxx> writes: >>> On Mon, Mar 27, 2023, at 11:38 AM, Jim Meyering wrote: >>>> We're overdue for a new release, so here's a snapshot in >>>> preparation for that, which I want to call 2.73 (skipping 2.72). >>>> There has never been an autoconf-2.72 release, yet `git describe` >>>> now prints 2.72c and has been printing strings like >>>> v2.72a-92-g8db00aa8 for years. >>> >>> Just as a note, I thought this version numbering scheme was weird >>> too the first time I encountered it, but the historical practice has >>> been that 2.72a, 2.72b, 2.72c, etc. are beta releases of 2.72. >> >> FWIW, the historical practice doesn't work very well for at least >> Gentoo's package manager, and I believe this is true for other >> distributions too. > > It doesn't work for Fedora either, where we often want to use git > describe. '2.72a' is also greater than '2.72' according to dpkg --compare-versions and glibc's strverscmp(). I fully support a change to a more easily machine-sortable version numbering scheme, but let's clearly announce that we are changing it, what we are changing it to, and that this is the reason we are skipping version 2.72. And write it down in HACKING and maybe also the manual proper. zw
