Re: SQUID-2.4.PRE-STABLE2

> > > I tagged this version and put it on the http/ftp servers.
> >
> > Just had a look at it - at the least, class 3 delay pools are broken, and the
> > patch in bugzilla is not sufficient to fix them. The original algorithm
> > should have worked, why was it changed?
> >
> > The patch in bugzilla may fix most of the breakage but it won't fix network
> > #255. I know the original code depended on obscure side-effects, but these
> > could have been documented in comments.
> >
> > I'll try and come up with a patch.
>
> Actually class 2 are also broken and will cause an assertion failure if host
> 255 is used, since the type has been changed from an unsigned char to an
> unsigned int, so 255 + 1 = 256 not 0. Apart from that, 255 looks like the
> original code.

And now I consider changing it back, I see a potential infinite loop in the
code when *all* 255 addresses are used. So that would be the bug that people
were attempting to fix, but the result was the breaking of both class2 and
class3 pools in certain situations.

David.

-- 
David Luyer                                        Phone:   +61 3 9674 7525
Engineering Projects Manager   P A C I F I C       Fax:     +61 3 9699 8693
Pacific Internet (Australia)  I N T E R N E T      Mobile:  +61 4 1111 2983
http://www.pacific.net.au/                         NASDAQ:  PCNTF