On Tue, 17 Nov 2015, Donald Sharp wrote:
You do realize that most systems have at most 100 nexthops(and that is
being generous)?
The comparison code becomes a lot simpler if the lists are sorted. It has
a definite "Uh, what?" factor at present, which is exactly why it caught
my eye. Insertion sort on nexthop_add is trivial - lib/linklist already
has a helper for that, if you'd use that. The comparison logic becomes a
trivial single scan.
And the code will then also work without surprises for those systems
outside the 'most' set, and in the future.
I have major concerns about making a entire test harness a pre-requisite
for a CR which is in progress.
Welcome to free software. Reviewers abuse your desire to get stuff in to
get further work done. You get grumpy, the software gets better.
Sure seems like we are moving the goal posts.
Nope, see other link. Same bar as for me.
This is not-entirely-trivial code, and it's super-central to the
functioning of Quagga. The lack of unit tests has been an issue in the
past, an obstacle to getting fixes and cleanups in.
regards,
--
Paul Jakma [email protected] @pjakma Key ID: 64A2FF6A
Fortune:
Let your conscience be your guide.
-- Pope
_______________________________________________
Quagga-dev mailing list
[email protected]
https://lists.quagga.net/mailman/listinfo/quagga-dev
