[gtkada] [Fwd: patches: missing shared-libgcc + DESTDIR support]

[John Marino]

Arnaud Charlet wrote:
> 
> That's actually not a correct diagnostic. gnatmake does not ignore
> -static-libgcc. I believe the issue is rather than libgtkada.so depends on
> libgcc_s, and testgtk depends on libgtkada.so.

I'm sure you are correct.  I was also getting confused with an issue 
with similar symptoms that was seen on AWS.  The bottom line is that 
without -shared-libgcc, /usr/lib/libgcc_s.so is linked in, and with 
-shared_libgcc, /usr/pkg/lib/libgcc_s.so or /usr/local/lib/libgcc_s.so 
is linked instead (as seen with FreeBSD / DragonFlyBSD, and the 
successful compile on NetBSD)

> 
> Can you explain how the DESTDIR change works and is intended to be used?
> If you could also update the Windows sections, that would make it easier to
> integrate the patch, since integrating a partial/inconsistent patch is much
> trickier.

This is also the first time I've heard of DESTDIR.  I found this 
interesting link:
http://www.gnu.org/prep/standards/html_node/DESTDIR.html

particularly this quote at the bottom:
"DESTDIR support is commonly used in package creation. It is also 
helpful to users who want to understand what a given package will 
install where, and to allow users who don’t normally have permissions to 
install into protected areas to build and install before gaining those 
permissions. Finally, it can be useful with tools such as stow, where 
code is installed in one place but made to appear to be installed 
somewhere else using symbolic links or special mount operations. So, we 
strongly recommend GNU packages support DESTDIR, though it is not an 
absolute requirement."

So to be clear, Matthias is the one that wrote the provided patches 
(patch-b*) and tested them, and of course gave his blessing to push them 
back upstream.  I could ask him if he's willing to complete the set, 
although I got the idea he was hoping somebody would take that over.  I 
also could take a stab at it, but I don't build on windows either so I 
can't really test any patches I create.

> 
>>             $(MAKE) -C glade install; \
>>             $(MAKE) -C gnome install; \
>>             $(MAKE) -C tools install; \
>> -	   $(CHMOD) ugo-w $(alidir)/static/*.ali; \
>> -	   if [ $(BUILD_SHARED) = yes ]; then \
>> -	      $(CHMOD) ugo-w $(alidir)/relocatable/*.ali; \
>> -	   fi; \
>>  	fi
> 
> The above hunk on src/Makefile.in looks wrong/unjustified though, can you
> clarify?
> 
> Arno

That looks strange to me as well.  I'll get clarification from Matthias.

John