Re: [PATCH 3/4] 'catch syscall' feature -- XML support part

[Tom Tromey <tromey@redhat.com>]

>>>>> "Sérgio" == Sérgio Durigan Júnior <sergiodj@linux.vnet.ibm.com> writes:

Tom> This will make an entry with an empty name given the above XML
Tom> snippet.  But that doesn't seem helpful... if this triggers, won't it
Tom> print as an unnamed syscall?  It seems to me that it would be
Tom> preferable to skip nameless ones and just report them numerically, if
Tom> they occur.

Sérgio> I'm sorry, but I think I didn't understand completely what you
Sérgio> suggested. You mean that I should set the syscall name to NULL when
Sérgio> there's no name for it in the XML file?

I think nameless syscalls should simply not be in the table at all.

Sérgio> +  if (gdb_xml_parse (parser, document) == 0)
Sérgio> +    {
Sérgio> +      /* Parsed successfully.  */
Sérgio> +      discard_cleanups (result_cleanup);
Sérgio> +      do_cleanups (back_to);
Sérgio> +      return data.sysinfo;
Sérgio> +    }
Sérgio> +  else
Sérgio> +    {
Sérgio> +      warning (_("Could not load XML syscalls info; ignoring"));
Sérgio> +      do_cleanups (back_to);
Sérgio> +      return NULL;
Tom> 
Tom> ... you can either just do_ or discard_ result_cleanup instead.

Sérgio> I'm sorry, instead of what? do_cleanups? :-)

I think instead of the above you can simply:

if (...)
  ...
  discard_cleanups (result_cleanup);
else
  ...
  do_cleanups (result_cleanup);

Tom