On 5/12/2026 9:59 PM, Lorenzo Stoakes wrote:
> On Tue, May 12, 2026 at 03:24:24PM +0800, Fujunjie wrote:
>>
>>
>> On 5/12/2026 12:30 AM, Lorenzo Stoakes wrote:
>>> Yeah let's go with that then I think :)
>>
>>
>> Thanks Lorenzo and Jann.
>>
>> I agree that the patch as posted goes in the wrong direction by treating
>> source gaps as ranges that should be propagated to the destination. Please
>> drop this patch.
>>
>> If disallowing source gaps in this multi-VMA move path still sounds right,
>> I can look into a v2/new patch that rejects sparse source ranges up front
>> instead.
>
> Sorry I really want to think about this a little more + make sure the
> semantics are sane, so I'd rather take a look at this myself.
>
> I do really hate to take something away from newer contributors but in this
> case things are quite fiddly so I want to make sure everything's correct :)
>
>>
>> Thanks,
>> fujunjie
>>
>
> Thanks, Lorenzo
No problem at all, that makes sense :)
