> Joel>         * gdb.ada/info_exc.exp: Start inferior before starting
> Joel>         the "info exceptions" tests.
> Joel>         * gdb.ada/mi_exc_info.exp: Start inferior before starting
> Joel>         the "-info-ada-exceptions" tests.
> 
> Thanks Joel.  This still fails, but in a new way, see appended.

Humpf :-(. At least that's progress.

This one is pretty easy to understand: Because you have debugging
info for *all* runtime units, "info exceptions" gets to list them
as well. The output you're getting is legit, so we have to accept it.
That's a bit unfortunate, as this forces us to introduce wildcards
in the expected output. Oh well...

I've pushed the patch you tested, and then pushed also the attached
patch.

gdb/testsuite/ChangeLog:

        * gdb.ada/info_exc.exp: Allow other global exceptions to be
        listed in the output of "info exceptions".
        * gdb.ada/mi_exc_info.exp: Allow other global exceptions to be
        listed in the output of "-info-ada-exceptions".

I did not test it on Fedora, since I haven't installed much on
the VM I just created. But I tested on my x86_64-linux laptop,
by reproducing another form of the same problem, involving the
use of "gnatmake -f -a -g" instead of plain "gnatmake -g" to build
the program. The added options force the GNAT runtime to be recompiled,
thus putting me in the same situation as you. The testsuite passes
for me in both situations (no debug info, with debug info).

Hopefully that'll solve your issue as well.

-- 
Joel