Re: [GNUnet-developers] Patch for network load measurement

[Igor Wronsky]

On Thu, 15 Jan 2004, Hendrik Pagenhardt wrote:

> This time a context diff is provided to help evaluating the changes by
> the core developers.

I've looked at it (despite the latest message you sent).
To be sure your "ret" calculations were correct, I would
probably have had to take up pen and paper, and I didn't
bother to do that. Maybe there is no need to be so concise
in the future. Anyway, at a glance they seem otherwise
ok'ish. The few new variables added (.*TrafficLastProc)
would probably still benefit from comments, including units
used and the idea in their calculation.

However, what I don't understand is the

  if (ret > -10/*lastRet*/) {

part. In addition to the comment there being in obscure place
(I'd prefer separate like /* line this and that */), I don't
understand what is the significance of -10? I gather that the
routines should return -1 on failure and a positive number
rest of the time. This reminds me that the previous routines
are probably handling the error situation incorrectly as well.


Igor