<div dir="ltr"><span style="font-family:monospace"><span style="color:rgb(0,0,0)">You might want to check the output of package_do_shlibs function in </span></span><span style="font-family:monospace"><span style="color:rgb(0,0,0)">meta/classes/package.bbclass in case it's already missing there.</span><br></span></div><div class="gmail_extra"><br><div class="gmail_quote">On Thu, May 4, 2017 at 1:30 PM, Peter Kjellerstedt <span dir="ltr"><<a href="mailto:peter.kjellerstedt@axis.com" target="_blank">peter.kjellerstedt@axis.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">> -----Original Message-----<br>
> From: Alexander Kanavin [mailto:<a href="mailto:alexander.kanavin@linux.intel.com">alexander.kanavin@<wbr>linux.intel.com</a>]<br>
> Sent: den 4 maj 2017 11:54<br>
> To: Peter Kjellerstedt <<a href="mailto:peter.kjellerstedt@axis.com">peter.kjellerstedt@axis.com</a>><br>
> Cc: OE Core (<a href="mailto:openembedded-core@lists.openembedded.org">openembedded-core@lists.<wbr>openembedded.org</a>) <openembedded-<br>
> <a href="mailto:core@lists.openembedded.org">core@lists.openembedded.org</a>><br>
> Subject: Re: Regression with RPM4 and unversioned DSOs<br>
><br>
> On 05/04/2017 11:46 AM, Peter Kjellerstedt wrote:<br>
> > Since I would very much like to avoid having to add explicit runtime<br>
> > dependencies for all those dependencies that were previously<br>
> automatically<br>
> > detected, which would be a nightmare, I am now looking for any help<br>
> as to<br>
> > where to find the code that adds the dependencies on the packages<br>
> providing<br>
> > DSOs and to see if it is possible to change it to also detect<br>
> dependencies<br>
> > on unversioned DSOs.<br>
><br>
> I'm not sure such implicit dependency detection at rpm generation point<br>
> is guaranteed to work in the first place. Don't we have a package_qa<br>
> check for this that should be failing? Can you publish a minimal layer<br>
> so I can reproduce locally?<br>
><br>
> If you want to poke in rpm4 source, I think grepping for<br>
> use_internal_dependency_<wbr>generator is a good starting point.<br>
<br>
I was more thinking of the code in OE-core that generates the list of<br>
Requires that goes into the spec file. The matching sections from the<br>
spec files for the apta-utility package that I used as example in my<br>
first mail are (for Pyro):<br>
<br>
Requires: libc6 >= 2.25<br>
<br>
and (for Morty):<br>
<br>
Requires: ld.so.1<br>
Requires: ld.so.1(GLIBC_2.4)<br>
Requires: libapta.so<br>
Requires: libc.so.6<br>
Requires: libc.so.6(GLIBC_2.0)<br>
Requires: libc.so.6(GLIBC_2.2)<br>
Requires: libc.so.6(GLIBC_2.3)<br>
Requires: libc.so.6(GLIBC_2.3.4)<br>
Requires: libc.so.6(GLIBC_2.4)<br>
Requires: libc.so.6(GLIBC_2.7)<br>
Requires: libc6 >= 2.24<br>
Requires: libubihelper.so<br>
Provides: elf(buildid) = 097a4862c4d6275838e8eb57595016<wbr>30b7913485<br>
<br>
So I think it is quite obvious that something has changed in OE-core in<br>
how these lists are generated... It is also apparent that the problem<br>
with lack of dependencies on packages with unversioned DSOs was already<br>
present in Morty. However, since it was hidden by the direct dependencies<br>
on the DSOs it was not apparent before.<br>
<br>
Does RPM4 support dependencies on DSOs, or was that an RPM5 feature?<br>
If RPM4 supports it, why was it removed when dnf was introduced?<br>
<br>
> Alex<br>
<span class="HOEnZb"><font color="#888888"><br>
//Peter<br>
<br>
--<br>
______________________________<wbr>_________________<br>
Openembedded-core mailing list<br>
<a href="mailto:Openembedded-core@lists.openembedded.org">Openembedded-core@lists.<wbr>openembedded.org</a><br>
<a href="http://lists.openembedded.org/mailman/listinfo/openembedded-core" rel="noreferrer" target="_blank">http://lists.openembedded.org/<wbr>mailman/listinfo/openembedded-<wbr>core</a><br>
</font></span></blockquote></div><br></div>