Time for a HP/PA hackathon?

[Andrew Cagney <cagney@gnu.org>]

Joel, Michael,

Would either of you be able to do a HP/PA build with the "0" here:

>   /* Frame unwind methods.  */
> #if 0

     // add this
     set_gdbarch_unwind_dummy_id (gdbarch, "hack function that returns 
null_frame_id for the dummy id");

> #else
>   set_gdbarch_deprecated_saved_pc_after_call (gdbarch, hppa_saved_pc_after_call

changed to a "1" and add an extra line, and then try that GDB to see 
what happens?

- Does it start?
There's a good chance GDB will panic during initialization ...

- If it starts, the output from:
(gdb) maint print architecture architecture
(gdb) shell grep -i deprecated architecture
I'm looking to see if/why legacy_frame_p() returns true/false (did that 
#if 1 really work?)

- what happens when:
./gdb .../advance
(gdb) break main
(gdb) run
(gdb) next ...
is executed.  I'm looking to see how/where GDB panics when stepping - 
this is testing the code's ability to unwind the PC register.

- what happens when
./gdb .../advance
(gdb) break func
(gdb) run
(gdb) advance func3
(it should end up in main)
I'm looking to see if/how GDB is correctly unwinding frame IDs - this is 
testing the code's ability to keep a frame ID constant across a function.

Andrew