Re: [2/2] Inspect extra signal information

[Pedro Alves <pedro at codesourcery dot com>]

On Monday 09 February 2009 06:23:36, Paul Pluzhnikov wrote:
> On Fri, Feb 6, 2009 at 3:35 PM, Pedro Alves <pedro@codesourcery.com> wrote:
> > Here's the version of this patch that went in.
> 
> AFAICT, this broke 'gdb.base/default.exp' show convenience,
> which expects "No debugger convenience variables ..."
> 

Sigh. I don't know how, but I somehow missed this one all
along.  I suspect I was comparing testresults with a log that
already had the problem.  I'll try to make sure it doesn't
happen again.

> Instead, it gets:
> 
> $ ./gdb -nx
> GNU gdb (GDB) 6.8.50.20090207-cvs
> ...
> This GDB was configured as "x86_64-unknown-linux-gnu".
> For bug reporting instructions, please see:
> <http://www.gnu.org/software/gdb/bugs/>.
> (gdb) show convenience
> $_siginfo = void
> (gdb)
> 
> I find it weird to have a "void" $_siginfo ...
> 

That's what you get for all other convenience variables when
they are empty:

E.g.:

 >gdb-6.8
 (gdb) p $_siginfo
 $1 = void

Or:

 >gdb-6.8
 (gdb) p $randomvar
 $1 = void
 (gdb) show convenience
 $randomvar = void

I'm not 100% sure if we want to fix this, or if we want to
adjust the testsuite instead, and making clear in the docs
that the variable can be empty.

We could try delaying creating the convenience variable a bit,
but, how much?  First stop? --- there's targets that could in
principle show extra signal information without a stop (opening
a core, or, opening a remote-ish target (ignore start_remote, please)).
First target push (not the exec target though) ? --- sounds like
a bit of a hack.

Grumble.

-- 
Pedro Alves