Re: [Qemu-devel] [PATCH RFC 1/3] error: don't rely on pointer comparisons

[Eric Blake]

On 06/16/2015 08:45 AM, Eric Blake wrote:
> On 06/16/2015 06:53 AM, Michael S. Tsirkin wrote:
>> makes it possible to copy error_abort pointers,
>> not just pass them on directly.
>>
>> Signed-off-by: Michael S. Tsirkin <address@hidden>
>> ---
>>  util/error.c | 16 +++++++++++-----
>>  1 file changed, 11 insertions(+), 5 deletions(-)
> 
> Where is this patch needed?
> 
> Is the goal to allow:
> 
> Error err = error_abort;

Oh, my bad for not double checking the layers of indirection involved.
The above comment makes no sense type-wise, since error_abort is already
'Error *', and &error_abort is 'Error **'.

> 
> as a way to statically initialize err via copy to behave the same as the
> global error_abort?
> 
> But that's not how you were using it in patch 3.  There, you were
> initializing Error *, so using &error_abort is still useful there, and
> pointer equality still suffices.
> 
> I don't see a convincing use for this patch.

And I retract that, because now I do see the use.

Error *local_err = ...;

as a way to set an abort-on-error pointer requires that we have more
than just a global error_abort abort-on-error pointer, but that any
number of pointers all resolve to something with embedded properties.

Sorry for the confusion on my part, and:

Reviewed-by: Eric Blake <address@hidden>

> 

-- 
Eric Blake   eblake redhat com    +1-919-301-3266
Libvirt virtualization library http://libvirt.org

signature.asc
Description: OpenPGP digital signature