[Bug 693798] Review Request: octave-image - Image processing for Octave

[bugzilla at redhat.com]

Please do not reply directly to this email. All additional
comments should be made in the comments box of this bug.


https://bugzilla.redhat.com/show_bug.cgi?id=693798

--- Comment #2 from Orion Poplawski <orion at cora.nwra.com> 2011-04-08 12:09:33 EDT ---
(In reply to comment #1)
> Issues:
> [!] : MUST - Rpmlint output is silent.
> 
>         octave-image-debuginfo.i686: W: spurious-executable-perm
> /usr/src/debug/image-1.0.13/src/edtfunc.c

Fixed.


>         octave-image.i686: W: obsolete-not-provided octave-forge
>         octave-image.i686: W: hidden-file-or-dir
> /usr/share/octave/packages/image-1.0.13/packinfo/.autoload
>         octave-image.i686: E: zero-length
> /usr/share/octave/packages/image-1.0.13/packinfo/.autoload
>         octave-image.i686: W: dangerous-command-in-%preun rm


> [!] : SHOULD - Final provides and requires are sane (rpm -q --provides and rpm
> -q --requires).
>  Provides: octave(api) = api-v47+}

Ah, typo in the macros.  Will get a new octave package out asap. 
Interestingly, rpm/yum appears to ignore the }..  I could install via yum just
fine.

>  Requires: (strange, are this used?)
>  __bilateral__.oct()(64bit)  
>  __bwdist.oct()(64bit)  
>  bwfill.oct()(64bit)  
>  bwlabel.oct()(64bit)  
>  __custom_gaussian_smoothing__.oct()(64bit)  
>  deriche.oct()(64bit)  
>  graycomatrix.oct()(64bit)  
>  hough_line.oct()(64bit)  
>  __imboundary__.oct()(64bit)  
>  nonmax_supress.oct()(64bit)  
>  rotate_scale.oct()(64bit)  
>  __spatial_filtering__.oct()(64bit)  

Well, these are the octave interfaces that are installed.  I think we can leave
them.  They already have their own namespace of a sort (.oct).

> Regarding the rpmlint warnings:
> 
> Is the obsolete really necessary? (It is a genuine question and I accept your
> answer). It just seems strange to see this for a package that has almost two
> years.

yes, to provide an upgrade path from octave-forge in F14.

> With respect to the "dangerous-command-in-%preun" I think that this is a false
> positive so it can be ignored (unless I am missing something obvious).

Well, there is an rm in the %octave_pkg_preun macro:

%octave_pkg_preun \
rm %{octpkgdir}/packinfo/on_uninstall.m \
if [ -e %{octpkgdir}/packinfo/on_uninstall.m.orig ]; then \
  mv %{octpkgdir}/packinfo/on_uninstall.m.orig
%{octpkgdir}/packinfo/on_uninstall.m \
  cd %{octpkgdir}/packinfo \
  %octave_cmd
l=pkg('list');on_uninstall(l{cellfun(@(x)strcmp(x.name,'%{octpkg}'),l)}); \
fi \
%{nil}

Looking closer, I think I can replace the rm with a mv -f.  That seems to
remove the rpmlint warning.

> What should we do with the final provides? Probably we need to rework the
> provides filter for octave packages. What do you think?

I think it is fine as is.  May actually help with octave package dependencies. 
Filtering leads to other rpm issues too.

-- 
Configure bugmail: https://bugzilla.redhat.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are on the CC list for the bug.