[Buildroot] [PATCH] package/e2fsprogs: backport build fix for rhel5 due to missing magic

[Thomas Petazzoni]

Hello,

On Wed, 9 Nov 2016 10:24:35 +0200, Baruch Siach wrote:

> On Wed, Nov 09, 2016 at 12:18:07AM -0800, Max Filippov wrote:
> > On Wed, Nov 9, 2016 at 12:03 AM, Baruch Siach <baruch at tkos.co.il> wrote:  
> > > On Tue, Nov 08, 2016 at 11:39:02PM -0800, Max Filippov wrote:  
> > >> RHEL 5.x does have magic.h, but it does not define all expected symbols. In
> > >> particular, the NO_CHECK symbols were only added in file 4.20 and RHEL 5.x
> > >> is using 4.17.  
> > >
> > > Instead of relying on host libmagic that may or may not be installed, why not
> > > add host-file to HOST_E2FSPROGS_DEPENDENCIES?  
> > 
> > My first reaction was to look if the issue is addressed in the package mainline,
> > and that's how e2fsprogs mainline decided to deal with that. So the next
> > version bump of e2fsprogs will make host-libmagic not necessary, there's
> > probably no reason to introduce that dependency?
> > 
> > "libmagic installed" vs. "not installed" is handled well by e2fsprogs, it's
> > "installed, but too old" case that fails.  
> 
> We would generally like to avoid dependency on host libraries as much as 
> possible since it reduces the build reproducibility. But as this is just a 
> host tool I'm not sure we care.

We do care.

So in this case, it would be better (if possible) to completely disable
the use of libmagic by host-e2fsprogs. This way, the behavior is
consistent on both machines that have libmagic installed and those who
don't.

Best regards,

Thomas
-- 
Thomas Petazzoni, CTO, Free Electrons
Embedded Linux and Kernel engineering
http://free-electrons.com