Patrick Steinhardt <ps@xxxxxx> writes: >> I suspect "%cs" is more in line with the spirit of GIT_DATE if I >> understand its purpose, i.e. "this is the time this version was >> recorded in the Git history, with the intention to give it the public" >> and better than "%as". > > While I agree with you, I'll leave this one as-is for now because this > is preexisting logic. Ah, then please disregard what I said. We are not _fixing_ anything that is not broken in this series. > But I'll rearrange a bit so that required > placeholders are only wired up once they are actually used so that it > can be seen where they come from and that the rewrite is faithful to the > original logic. Sounds very good. > Yeah, being consistent would be nice indeed. But for now I'd prefer to > keep this as-is because we'd otherwise change the version schema used by > CMake builds. In theory we can use 4 numbers here, too, where the fourth > number would correspond to the `PROJECT_VERSION_TWEAK` CMake variable. As I only write vX,Y,Z with three numbers, any fourth number a build system comes up with has no bit of information. So we could consider what we use is 3-number scheme, and supply a fixed and meaningless '0' to whoever wants the fourth number. Which is essentially what you are doing with the "split $(echo $V 0 0 0 0) with 'read A B C D' construct" ;-) I am fine with that as long as we have a shared understanding documented somewhere. Thanks.
