Gary Buhrmaster wrote:
> On Sat, Aug 1, 2020 at 9:23 PM Kevin Kofler <kevin.kofler@xxxxxxxxx>
> wrote:
>> Especially the CMake one was completely pointless.
>
> The goal was not pointless, but I will assert that the
> implementation was flawed in practice.
The goal was to make more packages do out-of-source builds, which is a
build-time-only implementation detail that should theoretically have no
effect whatsoever on the produced packages (though, in practice, at least
the directory structure of the -debugsource packages will be different).
Hence, an end user will probably not notice any difference whatsoever, which
is why I call the change pointless.
I also do not believe that the change really makes things easier for
packagers, just different. And different is bad when it means that dozens of
specfiles have to be adapted to such an incompatible change. The boilerplate
to do an out-of-source build (even with ancient versions of CMake that did
not have the -S and -B options) was very simple and widely used.
I also think that it is more valuable to be backwards-compatible with older
versions of Fedora (and where possible, also with RHEL/CentOS and EPEL) than
to be compatible with other, unrelated distributions that happen to also use
RPM. The CMake change, on the other hand, improved cross-compatibility with
other distributions at the expense of backwards compatibility with our own
distribution.
LTO at least promises tangible benefits for the end user. But maybe it
should have been opt-in (via the specfile) at first?
Kevin Kofler
_______________________________________________
devel mailing list -- devel@xxxxxxxxxxxxxxxxxxxxxxx
To unsubscribe send an email to devel-leave@xxxxxxxxxxxxxxxxxxxxxxx
Fedora Code of Conduct: https://docs.fedoraproject.org/en-US/project/code-of-conduct/
List Guidelines: https://fedoraproject.org/wiki/Mailing_list_guidelines
List Archives: https://lists.fedoraproject.org/archives/list/devel@xxxxxxxxxxxxxxxxxxxxxxx
