Re: RFA: implement ambiguous linespec proposal

[Tom Tromey <tromey at redhat dot com>]

Yao> I'll debug this problem tomorrow, but if you have any clues, that will
Yao> be helpful.

Tom> What is mysterious is that I do not see this failure.
Tom> I actually think this test should fail, because it is applying a
Tom> condition in a scope where there is no variable 'a'.

Tom> I will debug it today.

Ok, I found the problem.  "a" is defined in a system library for which I
have debuginfo installed:

    (top-gdb) p *sym.symtab
    $22 = {
    [...]
      filename = 0x26ba950 "../sysdeps/ieee754/dbl-64/atnat.h", 
      dirname = 0x26c4f10 "/usr/src/debug/glibc-2.14-34-ge2a3090/math", 

So the condition erroneously parses.

Changing the names to "a_param" instead makes it fail again.

I think the test is bogus, though, so I changed it in a way that will
make it succeed.

I can send a new revision of everything if you want.

Tom