On Sun, 11 Nov 2012 19:20:21 +0100, you wrote:

>Hi guys,

>then I gave the target=i586-mingw32msvc option a go. Turns out that
>theory didn't work so far. Actually I had to comment this in
>gcc/vhdl/Make-lang.in:
>
>vhdl.all.cross:
>        # @echo "No support for building vhdl cross-compiler"
>        # exit 1

That must have been there for a reason, but the reason may have been to reduce
the permutations and testing required...

>With some more library fiddling, I did eventually get a ghdl and ghdl1
>executable, but it turns out on installation (when trying to compile the
>standard libraries for the target), that ghdl wants to use the _native_
>assembler, not the cross one. So now I don't know if I'm
>just one step from the goal or on the completely wrong track. 

Even reporting that may turn out to be useful. I vaguely recall in my travels
through the build system, seeing some option for (I didn't pay enough attention)
to use a named assembler instead of the default... 

Alternatively, during an initial configuration problem I experimented with NOT
passing the -s flag to ghdl1 which does the work, and it *appeared* to build a
reasonable-looking .o directly. (at the time I had no way to test the resulting
.o) That might give you another avenue (or dead end!) to explore...

- Brian

_______________________________________________
Ghdl-discuss mailing list
[email protected]
https://mail.gna.org/listinfo/ghdl-discuss

Reply via email to