Re: [PATCH 0/3] meson: Drop userfaultfd_sysctl option
On Thu, Feb 15, 2024 at 09:21:15AM +0100, Jiri Denemark wrote:
On Wed, Feb 14, 2024 at 20:53:21 +0100, Andrea Bolognani wrote:
> Recently introduced as part of [1].
>
> We can skip the meson part for this one, and deal with everything
> in the spec file only.
>
> Test pipeline: https://gitlab.com/abologna/libvirt/-/pipelines/1176890275
>
> [1]
https://lists.libvirt.org/archives/list/devel@lists.libvirt.org/thread/DK...
Except for 1/3 (which I acked) I think this series is a little bit too
spec-centric. We can do a lot of things just in the spec file, but
having meson options still makes sense for people not building rpms.
Whether we use that option in our spec file or handle it differently is
another question, but I would vote against removing the meson option.
It's not always clear-cut.
A meson option is definitely necessary when it affects how the code
is built, or when we want something to be disabled by default while
giving user a convenient way to enable it. Having some non-trivial
logic deciding whether or not it should be enabled is also a hint
that something should be a meson option.
In this case, it's enabled by default and the way to undo its effects
after the file is simply to delete a single file. Doesn't quite
justify introducing yet another meson option just for it IMO.
Anyway, I just put this out there to see how people reacted to it. If
the option remains, so be it. I just think we could do without :)
--
Andrea Bolognani / Red Hat / Virtualization