[PATCH/commit] Handle errors in tracepoint target agent

[Stan Shebs]

Pedro Alves wrote:
> On Friday 26 March 2010 01:55:50, Stan Shebs wrote:
>   
>> For the record, here's what I ended up committing.  
>>     
>
> (For the record, if a patch ends up different to how it was
> being discussed, a chance for commenting before checking in
> would be really appreciated.)
>   
It seemed uncontroversial, but I guess not. :-)
> +	      for (p = p1; p < p2; p++)
> +		if (!((*p >= '0' && *p <= '9')
> +		      || (*p >= 'a' && *p <= 'f')
> +		      || (*p >= 'A' && *p <= 'F')))
> +		  break;
>   
Oh triple yuck.  Let's just make a new utility function - an ishex-type 
test shouldn't be coded more than once, or maybe twice, in a program.
> If you really insist in handling this in FSF gdb as well,
> then I'd like to merge this patch above to head, otherwise, 
> I'd rather just remove all the plain string handling from
> FSF gdb, and put that patch in our tree only.  (I don't
> really see the point in carrying that workaround forever in
> FSF gdb).
>   
Alright alright.  My interest in the subject is now totally exhausted, 
we'll just do the hex strings here.

Stan