This is the mail archive of the
gdb-patches@sourceware.org
mailing list for the GDB project.
Re: [PATCH][gdb/testsuite] Rewrite catch-follow-exec.exp
On 10/14/18 12:18 AM, Simon Marchi wrote:
> On 2018-10-11 04:33, Gary Benson wrote:
>> Tom de Vries wrote:
>>> On Wed, Oct 10, 2018 at 02:44:24PM +0100, Gary Benson wrote:
>>> > Oh! Ok, so [target_info gdb_protocol] != ""] maybe?
>>>
>>> Attached patch uses this this.
>>>
>>> OK for trunk?
>>
>> Please reorder the checks at the start like this, to minimize the
>> work done before bailing:
>>
>> 1. check gdb_protocol native
>> 2. check /bin/ls exists on target
>> 3. build_executable
>>
>> The patch is ok by me with these changes, but please wait for a
>> maintainer to give the final approval, I don't have those powers :)
>> And thanks for doing the work Tom!
>>
>> Cheers,
>> Gary
>
> Just wondering. Would it make life easier if we fixed PR 23368, which
> is the reason we have to do the test in an unnatural way?
Yes.
> I don't know
> if the fix I proposed here is totally correct:
>
> https://sourceware.org/bugzilla/show_bug.cgi?id=23368#c4
>
> but I think it would be an improvement compared to what we have now. I
> ran it through the buildbot and there are no regressions. If you want I
> can submit a proper patch for that, and then we can rewrite this test in
> a more normal way.
>
That makes sense, but also I propose to commit the patch I've submitted
(with the early-out checks reorder as Gary suggested above), since
that's also an improvement on the current situation.
OK?
Thanks,
- Tom