On 02/09/2012 09:31 AM, Pedro Alves wrote:
> On 02/09/2012 11:04 AM, Luis Gustavo wrote:
>> On 02/09/2012 06:22 AM, Jan Kratochvil wrote:
>>> On Thu, 09 Feb 2012 00:39:42 +0100, Luis Gustavo wrote:
>>>> +  /* Sort by pspace.  This effectively sorts locations by inferior in
>>>> +     a multi-inferior environment.  */
>>>> +
>>>> +  if (a->pspace != b->pspace)
>>>> +    return (a->pspace>   b->pspace) - (a->pspace<   b->pspace);
>>>
>>> This does not follow the comment I made:
>>>
>>>     /* Make the internal GDB representation stable across GDB runs
>>>        where A and B memory inside GDB can differ.
>>>
>>> I would prefer there (and it will not fit on a single line :-) ):
>>>
>>>          return (a->pspace->num>   b->pspace->num) - (a->pspace->num<   b->pspace->num);
>>>
>>> NUM should be the same on each GDB run for the same inferior / command file.
>>
>> I would still like to keep an appropriate comment regarding multi-inferiors next to the pspace comparison, like the following... maybe it would've been best to keep the change above your new comment. Here's a new version.
>>
>> What do you think?
>
> We need to sort by pspace even before that, before:
>
>    /* Sort permanent breakpoints first.  */
>    if (a_perm != b_perm)
>      return (a_perm<  b_perm) - (a_perm>  b_perm);
>
> So that you don't get:
>
>   #1 PSPACE1 ADDR1 PERM
>   #2 PSPACE2 ADDR1 PERM
>   #3 PSPACE1 ADDR1
>   #4 PSPACE2 ADDR1
>
> But instead:
>
>   #1 PSPACE1 ADDR1 PERM
>   #2 PSPACE1 ADDR1
>   #3 PSPACE2 ADDR1 PERM
>   #4 PSPACE2 ADDR1
>

I've checked the following in...

Luis