On 2025-10-27, Adrian Bunk wrote: > On Sun, Oct 26, 2025 at 10:49:44AM -0700, Vagrant Cascadian wrote: >> On 2025-10-26, Adrian Bunk wrote: >> > I've prepared an NMU for m4api (versioned as 0.3~0.9646fd-3.1) and >> > uploaded it to DELAYED/2. Please feel free to tell me if I should >> > cancel it. >> >> Thanks! >> >> Go ahead and upload directly, if you wish, > > Thanks, done. > >> and please also commit to >> git, although the patch is so small it shouldn't be hard to do so after >> the fact if that is more than you want to bother with. > > There is already an unrelated commit in git, and different maintainers > have different preferences how to deal with that when integrating an > NMU into git. > > Handling this kind of workflow differences is what I do not want to > bother with. > > And then there's all kind of other "fun" you run into when looking at > maintainer git trees, like when the contents of a tag for the previous > upload does not match what is in the archive.
Fair! I have merged your changes into git in the with my (hopefully not terribly) idiosyncratic workflow. :) >> My mind baffles a little at what "minimum supported version" even means >> if you need to bump it with no other changes, but that is a question for >> cmake... :) > > Similar to for example debhelper compat levels or "use" in perl, this is > a potentially breaking change. > > There were some compatibility policies with older CMake versions where > the old behaviour is no longer supported. > > It seems to rarely break in practice, but I do see the point of it. Sure. It has probably been spitting out all sorts of deprecation warnings for ages that have gone largely ignored, but I am still surprised at the hard failure! :) live well, vagrant
signature.asc
Description: PGP signature
