Michele Calgaro via tde-devels wrote:
on top of Slavek's comment, ninja-build is just
another tool that is
available out there. We looked at pros/cons and in the end it provide a
decent improvement on build times. So if a tool is good to make TDE
better, it is definitely worth considering it. The same was for gitea,
weblate, cmake and so on.
To be honest, I was a bit surprised to see such strong response on the
mailing list for a relatively small change. End users won't see much of a
different, packages are the same. And the build logic is done by cmake in
the same way. ninja-build is just a worker doing what cmake tell it to do,
just it does the job faster than make and with better progress feedback.
And even at 20%, it is a considerable time gain when you have to build for
40-60 distro/arch pais (I don't even know the exact number now, Slavek can
confirm that is needed - but it is a lot of building for sure :-) ).
"strong response" is a bit strong
In the end, we want to do what we believe is best for
TDE while
maintaining the TDE spirit and philosophy. I don't see "moving to
ninja-build" breaking either of them. Am I wrong?
Of course not - as mentioned my question is what is motivation behind and
numbers. The numbers match my observations. The thing is that we still have
to learn, understand and get used to the changes anyway.