[libvirt PATCH 008/351] meson: src/util/virfile: rewrite virFileActivateDirOverrideForProg

[Pavel Hrdina]

On Wed, Jul 22, 2020 at 12:13:18PM +0200, Pavel Hrdina wrote:
> On Wed, Jul 22, 2020 at 11:43:09AM +0200, Peter Krempa wrote:
> > On Wed, Jul 22, 2020 at 11:32:07 +0200, Ján Tomko wrote:
> > > On a Wednesday in 2020, Peter Krempa wrote:
> > > > On Wed, Jul 22, 2020 at 10:33:58 +0200, Ján Tomko wrote:
> > > > > On a Wednesday in 2020, Peter Krempa wrote:
> > > > > > On Thu, Jul 16, 2020 at 11:54:04 +0200, Pavel Hrdina wrote:
> > > > > > > With meson we no longer have .libs directory with the actual binary so
> > > > > > > we have to take a different approach to detect if running from build
> > > > > > > directory.
> > > > > > >
> > > > > > > This is not as robust as for autotools because if you select --prefix
> > > > > > > in the build directory it will incorrectly enable the override as well
> > > > > > > but nobody should do that.
> > > > > >
> > > > > > This wouldn't be that much of a problem as it would end up pointing to
> > > > > > the same files.
> > > > > >
> > > > > > More of a problem is if we falsely assume that the override is not
> > > > > > necessary.
> > > > > >
> > > > > 
> > > > > That's why I'd rather drop this "override-by-binary-path" approach
> > > > > and use the env variable in the "run" script for this.
> > > > 
> > > > That would mandate using ./run even in situations where it was not
> > > > required before.
> > > > 
> > > 
> > > Yes, that's what I meant.
> > 
> > I strongly disagree with doing that. Meson tries hard to make the files
> > in the build directory work out of the box. This would make it partially
> > and magically fail in many cases which is imo very bad.
> 
> I agree with Peter here, since meson already does some magic to
> correctly point to compiled libraries within the project it would be
> nice to keep this functionality.
> 
> I'll try the sentinel file approach that Peter suggested to see if it
> works and if it's not too ugly.

When I was looking into using sentinel file I realized that it is
pointless to use it.

The argument about NFS or mounting it into container is a good point but
irrelevant in our case. If we detect somehow that we need to override
the path our code has hard-coded abs_top_srcdir or abs_top_builddir as
part of the new path so it would not work in these cases even with the
sentinel file. In addition not only our code hard-codes the absolute
path, meson does the same for libraries.

So I'll keep the detection based on canonicalized path and
abs_top_builddir prefix match including some cleanup and not using PWD
directly.

Pavel
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <http://listman.redhat.com/archives/libvir-list/attachments/20200727/e8a92e0b/attachment-0001.sig>