Peter Maydell writes:
> On 8 October 2012 14:29, Anthony Liguori wrote:
>> This is wrong.
>>
>> Container properties are added by the user. You will turn a gracefully
>> failure (during hotplug) into an abort().
>
> No, it's turning a bug into an abort -- we don't handle trying to
> create two
On 8 October 2012 14:38, Peter Crosthwaite
wrote:
> Can we just populate errp with a nice meaningful error (perhaps the
> contents of that printf), then the caller can decide if failure is
> tolerable?
I would find this approach more plausible if it wasn't that so many
places in qemu just happily
On 8 October 2012 14:29, Anthony Liguori wrote:
> This is wrong.
>
> Container properties are added by the user. You will turn a gracefully
> failure (during hotplug) into an abort().
No, it's turning a bug into an abort -- we don't handle trying to
create two identically named properties correc
On Mon, Oct 8, 2012 at 11:29 PM, Anthony Liguori wrote:
> Peter Maydell writes:
>
>> Ping!
>
> This is wrong.
>
> Container properties are added by the user. You will turn a gracefully
> failure (during hotplug) into an abort().
>
Can we just populate errp with a nice meaningful error (perhaps
Peter Maydell writes:
> Ping!
This is wrong.
Container properties are added by the user. You will turn a gracefully
failure (during hotplug) into an abort().
Please limit this to static properties as they are not added by a user.
Regards,
Anthony Liguori
>
> -- PMM
>
> On 7 September 2012
Ping!
-- PMM
On 7 September 2012 14:55, Peter Maydell wrote:
> Reject attempts to add a property to an object if one of
> that name already exists. This is always a bug in the caller;
> this is merely diagnosing it gracefully rather than behaving
> oddly later.
>
> Signed-off-by: Peter Maydell
Am 07.09.2012 15:55, schrieb Peter Maydell:
> Reject attempts to add a property to an object if one of
> that name already exists. This is always a bug in the caller;
> this is merely diagnosing it gracefully rather than behaving
> oddly later.
>
> Signed-off-by: Peter Maydell
Looks fine to me,
Reject attempts to add a property to an object if one of
that name already exists. This is always a bug in the caller;
this is merely diagnosing it gracefully rather than behaving
oddly later.
Signed-off-by: Peter Maydell
---
Changes v1->v2:
* use abort() rather than assert(0), as suggested by P