On 9/23/21 1:00 AM, Ilya Maximets wrote:

>>> I was wondering if it makes sense to factor out the common part of
>>> ovsdb_datum_added_removed() and ovsdb_datum_diff().  It seems to me that
>>> ovsdb_datum_diff(diff, old, new, type) is almost the same as
>>> ovsdb_datum_added_removed(diff, diff, old, new, type), we just need to
>>> handle the non-composite types separately and deal with the case when
>>> 'added' == 'new' if we detect a value changing for a given key.
> 
> I thought about this and it would be great to reduce some code duplication.
> But I found dealing with corner cases inside he function a bit awkward.
> Especially the part where we need to check that 'added' == 'new' and add
> only one of the values to one of the sets and not touch the other one.

I agree, it's not that pretty.

> So, I kept it as is for now.
> 

Sounds good to me, thanks!

_______________________________________________
dev mailing list
[email protected]
https://mail.openvswitch.org/mailman/listinfo/ovs-dev

Reply via email to