[libvirt] [dbus PATCH v2] build: convert to Meson/Ninja build system

Andrea Bolognani abologna at redhat.com
Mon Sep 23 09:59:01 UTC 2019


On Mon, 2019-09-23 at 10:17 +0200, Pavel Hrdina wrote:
> On Sat, Sep 21, 2019 at 09:22:38AM +0200, Fabiano Fidêncio wrote:
> > One last things here (thanks, Pavel, for pointing this out), I'd
> > prefer the 'syntax' suite being called 'syntax-check' just because
> > people are already used to the 'syntax-check' name.
> > 
> > So, please, would you mind changing it as well?
> 
> I don't mind changing it if we agree on the naming, I'll have one last
> argument for the shorter and easier to type name, coping it from the
> comment on gitlab:
> 
> To me it feels redundant to have the `check` as part of the suit label
> as you already know that you are running test suit so the check is
> somehow implied. Consistency is a nice thing if it makes sense, but we
> are completely changing the workflow so I don't see any reason to pick
> longer and redundant name just for consistency reasons.
> 
> Based on that I still prefer using simply `syntax`, let's see of others
> have some opinion about it.

For what it's worth, I also prefer 'syntax': the -check suffix is
useful for make because it highlights that you're verifying some
property of the program, and also clearly ties

  make check
  make syntax-check

together. In the case of Meson, the fact that you're verifying
something is explicit in the name of the ninja target, and also

  ninja test --suite unit
  ninja test --suite syntax

are already very obviously connected. Using 'syntax-check' for the
latter would, if anything, make the relationship unbalanced: why
didn't we call the former 'unit-check' then?

The switch to Meson is a clear-cut break from what we had before,
and we should take the opportunity to re-evaluate whether our
existing decisions are still valid in a post-autotools world: in
this specific case, I think it makes more sense to change the name
rather than stick with the historical one.

-- 
Andrea Bolognani / Red Hat / Virtualization




More information about the libvir-list mailing list