On 2015-09-08 01:21, Ananyev, Konstantin wrote: >> >> Thanks. I got it wrong anyway, what I wanted was to be able to handle >> the day when nb_segs changes to a 16-bit number, but then it should >> really be >> >> ... >= 1 << (sizeof(head->nb_segs) * 8) >> >> anyway. I'll fix that and also add a warning that the implementation >> will do a linear search to find the tail entry. > > Probably just me, but I can't foresee the situation when we would need to > increase nb_segs to 16 bits. > Looks like an overkill to me.
I don't think it will happen either, but with this solution, this particular piece of code will work regardless. The value is known at compile-time anyway, so it should not be a performance issue. // Simon
