On Thursday 24 November 2011 15:25:40 Daniele Furlan wrote:
2011/11/24 Sven Eckelmann sven@narfation.org:
On Thursday 24 November 2011 14:06:01 Daniele Furlan wrote:
What do you think about some information, reasoning, examples,....
Ok, using letter a,b,c,k instead of:
c = total_count a = neigh_rq_count b = orig_eq_count k = TQ_LOCAL_BIDRECT_SEND_MINIMUM
what about l = TQ_LOCAL_BIDRECT_RECV_MINIMUM?
then the code do the following:
c = min (a,b) and then check this condition: (c<k OR a<k)
no, it compares (c<k || a<l)
[snip] Wrong proof using wrong assumptions. You just cannot say that TQ_LOCAL_BIDRECT_SEND_MINIMUM has always the same value as TQ_LOCAL_BIDRECT_RECV_MINIMUM just because it is in this particular example.
And just as reminder: Such things belong in the commit message.
Kind regards, Sven