mismatch in hdf packages prevent upgrade from F30->F31

[Jouk Jansen]

Hi All,

When performing an upgrade from F30 to F31 on one of my machines the upgrade
process tries to install (yes an install not an upgrade) the packages
hdf.x86-64 and hdf.i686. This fails due to a mismatch in these packages of
/usr/bin/h4cc and /usr/bin/h4fc.

Does anyone has an idea which other package recently got a dependence on hdf
and triggers this problem? If I know I may can remove the i686 version from
my system.

Regards
                 Jouk

Pax, vel iniusta, utilior est quam iustissimum bellum.
    (free after Marcus Tullius Cicero (106 b.Chr.-46 b.Chr.)
     Epistularum ad Atticum 7.1.4.3)

Touch not the cat bot a glove
> ------------------------------------------------------------------------------<

Jouk Jansen
    	 
  [e-mail]

Technische Universiteit Delft        tttttttttt  uu     uu  ddddddd
  Kavli Institute of Nanoscience       tttttttttt  uu     uu  dd    dd
  Nationaal centrum voor HREM              tt      uu     uu  dd     dd
  Lorentzweg 1                             tt      uu     uu  dd     dd
  2628 CJ Delft                            tt      uu     uu  dd     dd
  Nederland                                tt      uu     uu  dd    dd
  tel. [phone]                       tt       uuuuuuu   ddddddd
> ------------------------------------------------------------------------------<

[Ed Greshko]

On 11/7/19 5:11 PM, Jouk Jansen wrote:
> When performing an upgrade from F30 to F31 on one of my machines the upgrade
> process tries to install (yes an install not an upgrade) the packages
> hdf.x86-64 and hdf.i686. This fails due to a mismatch in these packages of
> /usr/bin/h4cc and /usr/bin/h4fc.
>
> Does anyone has an idea which other package recently got a dependence on hdf
> and triggers this problem? If I know I may can remove the i686 version from
> my system.

I had no issues to upgrade several systems.

But, what you've said suggests that you may have at least one i686 installed which now needs
hdf when it didn't before.

If you do

rpm -qa | grep i686

does it return anything?

[Samuel Sieb]

On 11/7/19 1:11 AM, Jouk Jansen wrote:
> When performing an upgrade from F30 to F31 on one of my machines the upgrade
> process tries to install (yes an install not an upgrade) the packages
> hdf.x86-64 and hdf.i686. This fails due to a mismatch in these packages of
> /usr/bin/h4cc and /usr/bin/h4fc.

The problem is that one package is providing both libraries and 
executables.  It really should be split.  The libraries are parallel 
installable, but the executables are not.

There is already a bug filed:
https://bugzilla.redhat.com/show_bug.cgi?id=1769326

[Samuel Sieb]

On 11/7/19 10:28 AM, Samuel Sieb wrote:
> On 11/7/19 1:11 AM, Jouk Jansen wrote:
> > When performing an upgrade from F30 to F31 on one of my machines the 
> > upgrade
> > process tries to install (yes an install not an upgrade) the packages
> > hdf.x86-64 and hdf.i686. This fails due to a mismatch in these 
> > packages of
> > /usr/bin/h4cc and /usr/bin/h4fc.
>
> The problem is that one package is providing both libraries and 
> executables.  It really should be split.  The libraries are parallel 
> installable, but the executables are not.
>
> There is already a bug filed:
> https://bugzilla.redhat.com/show_bug.cgi?id=1769326

And now I see that you filed it.  I was looking at the assignee and 
thought the package owner had filed it.

[Ed Greshko]

On 11/8/19 2:28 AM, Samuel Sieb wrote:
> On 11/7/19 1:11 AM, Jouk Jansen wrote:
> > When performing an upgrade from F30 to F31 on one of my machines the upgrade
> > process tries to install (yes an install not an upgrade) the packages
> > hdf.x86-64 and hdf.i686. This fails due to a mismatch in these packages of
> > /usr/bin/h4cc and /usr/bin/h4fc.
>
> The problem is that one package is providing both libraries and executables.  It really should be split.  The libraries are parallel installable, but the executables are not.

I am not familiar with the software in question.  So, I wonder.  Is there any use case where on an
x86_64 system it is advisable to have both i686 and x86_64 executables?

[Samuel Sieb]

On 11/9/19 4:34 AM, Ed Greshko wrote:
> On 11/8/19 2:28 AM, Samuel Sieb wrote:
> > On 11/7/19 1:11 AM, Jouk Jansen wrote:
> > > When performing an upgrade from F30 to F31 on one of my machines the 
> > > upgrade
> > > process tries to install (yes an install not an upgrade) the packages
> > > hdf.x86-64 and hdf.i686. This fails due to a mismatch in these 
> > > packages of
> > > /usr/bin/h4cc and /usr/bin/h4fc.
> >
> > The problem is that one package is providing both libraries and 
> > executables.  It really should be split.  The libraries are parallel 
> > installable, but the executables are not.
>
> I am not familiar with the software in question.  So, I wonder.  Is 
> there any use case where on an
> x86_64 system it is advisable to have both i686 and x86_64 executables?

I can't think of any use case for that.

[Samuel Sieb]

On 11/7/19 1:11 AM, Jouk Jansen wrote:
> When performing an upgrade from F30 to F31 on one of my machines the upgrade
> process tries to install (yes an install not an upgrade) the packages
> hdf.x86-64 and hdf.i686. This fails due to a mismatch in these packages of
> /usr/bin/h4cc and /usr/bin/h4fc.
>
> Does anyone has an idea which other package recently got a dependence on hdf
> and triggers this problem? If I know I may can remove the i686 version from
> my system.

The maintainer was very responsive.  There's already an update pending 
to fix this.