Re: No more static `libgnat` with FSF GNAT?

["Yannick Duchêne (Hibou57)" <yannick_duchene@yahoo.fr>]

On Sun, 14 Apr 2013 15:19:11 +0200, Pascal Obry <pascal@obry.net> wrote:

Hi Pascal, nice to see you


> Yannick,
>> Is this Ubuntu specific or is this FSF GNAT? Is there a technical  
> reason
> > for not including any‑more a static version of `libgnat` or was it just
> > forgotten?
> > > Note: using libgnat-4.6.3

> I still have the static libgnat in gnat-4.6 package on GNU/Debian:
>$ apt-file list gnat-4.6 | grep libgnat
> […]

So that's in the GNAT package. I did not look here, and looked at the  
libgnat-4.6 package instead.

Here is about what I did:

     $ dpkg-query --listfiles libgnat-4.6 | grep libgnat

     /usr/share/lintian/overrides/libgnat-4.6
     /usr/lib/i386-linux-gnu/libgnat-4.6.so.1
     /usr/share/doc/libgnat-4.6

Thus I did not find it.

Using a command similar to the one I used, here is what you did:

     $ dpkg-query --listfiles gnat-4.6 | grep libgnat

     /usr/lib/gcc/i686-linux-gnu/4.6/rts-native/adalib/libgnat.a
     /usr/lib/gcc/i686-linux-gnu/4.6/rts-native/adalib/libgnat.so
     /usr/lib/i386-linux-gnu/libgnat.so
     /usr/lib/i386-linux-gnu/libgnat-4.6.so

And here the *both* a shared and static version appears.

Something I don't understand now then, is: if the GNAT package provides  
the library, what is the purpose of this separate shared library package?  
I though it may be a dependency package (without installing GNAT) for  
applications linked to the shared GNAT library, but if I try to remove it,  
Synaptic tells me I have to remove GNAT as a whole too. Strange…

I could build adding this to the Linker package option in the GPR file:

     Common_Switches :=
       ([…],
        "/usr/lib/gcc/i686-linux-gnu/4.6/rts-native/adalib/libgnat.a");


That works, but I don't understand the rational behind this, and I don't  
like it (to not understand it). I will need to have a deeper review of  
this topic.


--
Yannick Duchêne
-- 
“Syntactic sugar causes cancer of the semi-colons.” [1]
“Structured Programming supports the law of the excluded muddle.” [1]
[1]: Epigrams on Programming — Alan J. — P. Yale University