> Ah, I ran it again with my full spectrum of compilers.
> 
>   gcc 2.95.3
>   gcc 3.2-7-rh
>   gcc 3.3.2
>   gcc gcc-3_3-branch
>   gcc HEAD
> 
> With dwarf-2:
>   gcc 3.3.2 and gcc gcc-3_3-branch gave the four extra FAILs.
>   gcc 2.95.3, gcc 3.2-7-rh, and gcc HEAD gave pure PASS+KFAIL.
> 
> With stabs+
>   All compilers gave pure PASS+KFAIL.
> 
> Since it PASSes with gcc HEAD I am more sanguine that the
> questionable tests (the tests that FAIL with some gcc's)
> are good tests.

These?

>   # gcc 3.3.2, binutils 2.14, -gdwarf-2
> 
>   p/c fun2()
>   $1 = {a = 97 'a', b = 0 '\0'}
>   (gdb) FAIL: gdb.base/structs.exp: p/c fun2(); call 2 structs-tc-tll
> 
>   p/c L2
>   $2 = {a = 97 'a', b = 0 '\0'}
>   (gdb) FAIL: gdb.base/structs.exp: p/c L2; call 2 structs-tc-tll
> 
>   p/c fun2()
>   $1 = {a = 97 'a', b = 0 '\0'}
>   (gdb) FAIL: gdb.base/structs.exp: p/c fun2(); call 2 structs-tc-td
> 
>   p/c L2
>   $2 = {a = 97 'a', b = 0 '\0'}
>   (gdb) FAIL: gdb.base/structs.exp: p/c L2; call 2 structs-tc-td

They should _never_ fail.  Failing indicates a bug in GCC, GDB, or both, 
and that the testsuite is doing a good job.  All those KFAILs are 
similar (only I've gaged them :-/).

> Tarball with executable files, gdb.log, et cetera at
> 
>   ftp://ftp.shout.net/~mec/gdb/2003-11-10-structs.tar.gz
> 
> Michael C

And here's todays version.  Change is to report this problem:

KFAIL: gdb.base/structs.exp: return and finish consistent; return 1 
structs-tc (PRMS: gdb/1444)

explicitly.  It makes understanding the results easier.

that's definitly it!

Andrew