Re: [PATCH 0/7] Implement gdbarch_gdb_signal_{to,from}_target

[Doug Evans <dje@google.com>]

On Fri, Jul 26, 2013 at 1:13 PM, Sergio Durigan Junior
<sergiodj@redhat.com> wrote:
> This is the second attempt to implement gdbarch_gdb_signal_to_target,
> but this time with the additional implementation of
> gdbarch_gdb_signal_from_target (a MIPS-only method so far).
>
> The first attempt is here:
>
> <http://sourceware.org/ml/gdb-patches/2013-07/msg00002.html>
>
> The notable changes are:
>
> - linux-tdep.c now registers the generic converters itself, thus
>   eliminating the need for every target to do so.
>
> - Comment about SIGSWI on ARM (discussion here:
>   <http://sourceware.org/ml/gdb-patches/2013-07/msg00002.html>).
>
> - Rewrite MIPS implementation of gdbarch_gdb_signal_from_target.
>   Remove "enum mips_signals".
>
> Hopefully everything is right now.  If the patch is approved, I intend
> to commit it as a whole, instead of each chunk separately.
>
> Thanks.

Hi.

I reviewed all the patches.
Except for a few nits pointed out in some of the patches, the set is ok with me.
As a sanity check, did you do an enable-targets=all build?
[modulo I didn't audit the signal lists and values]
Also, you may wish to leave it a few days in case others wish to comment.

I gather there is a use for gdbarch_gdb_signal_to_target coming, so no
worries there.

In older times one couldn't use "signal" as a variable name to avoid a
possible collision with a system header defining a macro signal.  I
see record-full.c uses "signal" so perhaps we no longer have to deal
with that, but heads up.

[fwiw, and this is just a side discussion, not meant to block the
patch or anything ...
This is an example of the kind of thing that I wish existed in a
library anyone could use.
Yeah, we're converting to/from GDB_FOO, but I can imagine it being
useful to multiple tools.]