[PROJ] PROJ 6.0.0RC4

[Elliott Sales de Andrade]

On Wed, 27 Feb 2019 at 05:00, Even Rouault <even.rouault at spatialys.com> wrote:
>
> On mercredi 27 février 2019 04:43:35 CET Elliott Sales de Andrade wrote:
> > On Tue, 26 Feb 2019 at 16:19, Kristian Evers <kreve at sdfe.dk> wrote:
> > > Following the issues reported with NetBSD builds I have just uploaded PROJ
> > > 6.0.0RC4. Thanks to Even for cracking this one so quickly!
> > >
> > > Links are here:
> > >
> > > https://download.osgeo.org/proj/proj-6.0.0RC4.tar.gz
> > > (https://download.osgeo.org/proj/proj-6.0.0RC4.tar.gz.md5)
> > > https://download.osgeo.org/proj/proj-6.0.0RC4.zip
> > > (https://download.osgeo.org/proj/proj-6.0.0RC4.zip.md5)
> > With these versions, all the build failures I saw are fixed, and it
> > scratch builds on all Fedora systems:
> > https://koji.fedoraproject.org/koji/taskinfo?taskID=33077369
> >
> > Locally, if I install all released regional datumgrid tarballs, then
> > the build is fine. However, if I install
> > proj-datumgrid-north-america-1.2RC2, then I see the following failure
> > (it is okay with 1.1, maybe because the required file is not in
> > there):
>
> Yes, that's "expected". testntv2 requires base proj-datumgrid (all CI setups
> install proj-datumgrid only), but is going to react differently if other extra
> grids are installed.
>

But testntv2 is only run if ntv2_0.gsb exists, which means it could
only possibly be run with proj-datumgrid-north-america installed. In
other words, it's not reacting *differently*, it's just not reacting
at all. So I'm not saying it's wrong, but CI passing is not proof that
it's right.


--
Elliott