"Sibi Siddharthan via GitGitGadget" <gitgitgadget@xxxxxxxxx> writes:
> This is an attempt to build Git using CMake.
What is your longer-term plan on this one? Would the ultimate goal
be
(1) to maintain both CMake and Makefile and keep them in sync?
(2) force all the Git developers (not users or binary packagers)
start using cmake and abandon Makefile, and let CMake generate
Makefile?
(3) something else?
Given that bootstrapping CMake typically takes a working "make", by
definition, if you drop Makefile from our shipped source tree, the
audience you can support on minority platforms would be smaller than
what we have. It is not negotiable that we keep Makefile as a valid
way to build Git, and the Makefile must be tweakable for those on a
minority platforms when things do not go the way they want to (the
tweakability is most needed on minority/exotic platforms).
W ith either (1) or (2) above, I have a feeling that engineering
burden to divert resources to add cmake support (with sustainable
plan to keep "fixable/tweakable" Makefile that is in sync) alone
would already be high enough. It is unclear the reason why we want
to pay all that cost---to gain what? Other than to make those who
like CMake feel happier, that is.
I do not want to sound overly negative in my first response, but you
might thank whoever says "no" early before you invest too much time
on this topic, so I'll send the above before even reading these 8
patches.
Thanks for your interest in the project.
