This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH] Eliminate UNSUPPORTED_ERROR.
- From: Doug Evans <xdje42 at gmail dot com>
- To: Pedro Alves <palves at redhat dot com>
- Cc: Hui Zhu <hui_zhu at mentor dot com>, gdb-patches ml <gdb-patches at sourceware dot org>
- Date: Wed, 11 Dec 2013 20:22:55 -0800
- Subject: Re: [PATCH] Eliminate UNSUPPORTED_ERROR.
- Authentication-results: sourceware.org; auth=none
- References: <5265022F dot 8060203 at mentor dot com> <52654A2C dot 9010202 at redhat dot com> <529707C7 dot 4040504 at mentor dot com> <5298AE7C dot 6020607 at redhat dot com> <529C80D2 dot 2080608 at mentor dot com> <529C9B42 dot 20600 at redhat dot com> <529D62F7 dot 80701 at mentor dot com> <52A22582 dot 8040509 at redhat dot com> <52A40015 dot 207 at mentor dot com> <52A61E86 dot 3020005 at redhat dot com> <m31u1lvimb dot fsf at sspiff dot org> <52A750AA dot 1080807 at redhat dot com> <52A75A05 dot 60006 at redhat dot com> <CAP9bCMRyMQESGWQFSsB-+OJ=eSgtcnNcC6J4hDxq5n6kSY=c6Q at mail dot gmail dot com> <52A8BA3D dot 9020403 at redhat dot com>
On Wed, Dec 11, 2013 at 11:17 AM, Pedro Alves <palves@redhat.com> wrote:
> On 12/11/2013 04:33 PM, Doug Evans wrote:
>>> > - /* Should we fallback to ye olde GDB script mode? */
>>> > - if (script_ext_mode == script_ext_soft
>>> > - && e.reason == RETURN_ERROR && e.error == UNSUPPORTED_ERROR)
>>> > - {
>>> > - fseek (stream, 0, SEEK_SET);
>>> > - script_from_file (stream, (char*) file);
>>> > - }
>>> > - else
>>> > - {
>>> > - /* Nope, just punt. */
>>> > - throw_exception (e);
>>> > - }
>>> > + /* Fallback to GDB script mode. */
>>> > + fseek (stream, 0, SEEK_SET);
>>> > + script_from_file (stream, (char*) file);
>> Remove the fseek and cast.
>
> Hmm, indeed the fseek doesn't look necessary. Will do. However,
> that makes me wonder whether I'm missing something, as it doesn't
> look necessary before my patch either. Was it ever really
> needed?
I think I added it to be extra conservative in case
source_python_script happened to read from the file.
[I know it normally wouldn't read and then throw an UNSUPPORTED_ERROR,
but I'm guessing I didn't want this code to assume that.]
Now that we're not even calling source_python_script the need is gone.
I see you added it (and the unnecessary cast too ;-) ) in:
> https://www.sourceware.org/ml/gdb-patches/2010-04/msg00110.html
> but I can't tell why.
script_from_file was changed to take a const char *arg the prior week.
Presumably I just didn't update that part of the patch.