Re: [lwip-users] Bug in inet_chksum_pseudo() produces bad checksum (Paradigm C++)

[Dave Lyneham]

Hi David,

I tend to agree.  I don't think the code should be distorted to
compensate for a bug in a particular compiler.  IMHO best solution
would be for the compiler to be fixed and I am happy to patch my code
until that happens.

Cheers
Dave


> I think you meant "should be negligible".
>
>
> Performance issues aside, if I was reading code which did a "greater than 
> zero" test I
> would  immediately be wondering what was special about the value being 
> negative, then
> discover the 'acc' variable was unsigned, then wonder why someone had written 
> it using
> greater-than instead of not-equal.
>
> My opinion (for what it is worth) is that LWIP should stick with either the 
> original form
> (number 1) or the more explicit form (number 2) which has already been 
> checked in, but
> Dave should patch his own copy to work around the bug in his compiler.