Re: [rfc] [02/05] Get rid of current_gdbarch in hppa-linux-nat.c

[Markus Deuling <deuling at de dot ibm dot com>]

Mark Kettenis schrieb:
> > Date: Wed, 07 Nov 2007 12:11:02 +0100
> > From: Markus Deuling <deuling@de.ibm.com>
> >
> > Hi,
> >
> > this patch adds gdbarch as a parameter to hppa_linux_register_addr. 
> >
> > Tested with gdb_mbuild. Ok to commit ?
>
> Actually, this is getting a bit silly.  That gdbarch is only needed
> for a sanaity check, and obviously the gdbarch_num_regs call can just
> be replaced with an appropriate bounds check on the u_offsets array.
>
>    if (regno < 0 || regno >= ARRAY_SIZE(u_offsets))
>
> should do the trick.

Sure it would. But what for do we have gdbarch_num_regs? I dont think its a good idea
to either use gdbarch_num_regs or ARRAY_SIZE(whatever) at will. This is redundant and error-prone.
Btw, there are two further uses of gdbarch_num_regs in that file.

For my opinion gdbarch should be used to describe an architecture.


--
Markus Deuling
GNU Toolchain for Linux on Cell BE
deuling@de.ibm.com