[dpdk-dev] [PATCH] ci: pin meson to 0.52.0

[Bruce Richardson]

On Wed, Jan 08, 2020 at 11:09:59AM -0500, Aaron Conole wrote:
> Bruce Richardson <bruce.richardson at intel.com> writes:
> 
> > On Wed, Jan 08, 2020 at 10:10:00AM -0500, Aaron Conole wrote:
> >> David Marchand <david.marchand at redhat.com> writes:
> >> 
> >> > On Wed, Jan 8, 2020 at 1:10 PM Bruce Richardson
> >> > <bruce.richardson at intel.com> wrote:
> >> >>
> >> >> On Wed, Jan 08, 2020 at 12:59:35PM +0100, David Marchand wrote:
> >> >> > On Wed, Jan 8, 2020 at 12:29 PM Luca Boccassi <bluca at debian.org> wrote:
> >> >> > >
> >> >> > > On Wed, 2020-01-08 at 12:02 +0100, David Marchand wrote:
> >> >> > > > meson 0.53.0 has a compatibility issue [1] with the python 3.5.2 that
> >> >> > > > comes
> >> >> > > > in Ubuntu 16.04.
> >> >> > > > Let's pin meson to 0.52.0 while the fix is being prepared in meson.
> >> >> > > >
> >> >> > > > 1:
> >> >> > > > https://github.com/mesonbuild/meson/issues/6427
> >> >> > > >
> >> >> > > >
> >> >> > > > Signed-off-by: David Marchand <
> >> >> > > > david.marchand at redhat.com
> >> >> > > > >
> >> >> > > > ---
> >> >> > > >  .ci/linux-setup.sh | 2 +-
> >> >> > > >  1 file changed, 1 insertion(+), 1 deletion(-)
> >> >> > >
> >> >> > > Acked-by: Luca Boccassi <bluca at debian.org>
> >> >> >
> >> >> > There is a 0.52.1 version available, so I suppose we can blacklist
> >> >> > meson < 0.53 instead.
> >> >> > Thought?
> >> >> >
> >> >> > If noone objects, I will apply a fix by the end of the day.
> >> >> >
> >> >> Wondering if there is value in using 0.47.1, the minimum version we
> >> >> support, to catch potential issues with someone using features from newer
> >> >> versions? I suspect there are more people using the latest releases of
> >> >> meson than the baseline supported version?
> >> >
> >> > Testing with a fixed version seems better in a CI, and since we
> >> > announce this minimum version, then yes, it makes sense.
> >> > I will post a v2.
> >> 
> >> Why is 0.47.1 still the minimum?  Don't we require features that are
> >> introduced as of 0.50?
> >> 
> > No, it should still work fine, and a quick sanity check tested with 0.47.1
> > on my system shows no issues, so I think we are good.
> >
> > There are some warnings printed about future features when you use a later
> > version, but in all cases the extra parameters added are just ignored by
> > the older versions, so compatiblity is maintained. Adding 0.47.1 to the CI
> > will also help avoid any inadvertent new version requirements from sneaking
> > in.
> 
> Makes sense to me.  Is there a way to make it whitelist the warnings we
> know about already?  It would be nice to have the CI environment be
> warning-free (but I don't object to 0.47.1 being a minimum version or
> anything).
> 
I am currently working on reducing the number of meson warnings we have.
Patchset I'd hope to get out tomorrow, with meson 0.52 shows only a single
warning with doc builds disabled, and 2 with docs enabled. [That's reduced
from 5 in both cases as it is now].

I'd hope to get the extra doc build warning removed if I have time in DPDK
20.02, but the final warning is about using meson's built-in warning
options rather than our own flags, so I'm not even sure if we want to get
rid of it.

/Bruce