This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: PATCH for PR 14699
Neil Booth <neil@daikokuya.co.uk> writes:
> Serge Belyshev wrote:-
>
>> >No. I believe the whole idea is to avoid fp!
>>
>> For what reason?
>
> I can't remember - you should ask Zack as he wrote it.
No, the point is not to avoid fp (that would be silly). The point of
approx_sqrt is to avoid having to figure out whether or not it's
necessary to stick -lm on the link line.
I don't see anything wrong with changing from integer to floating
point arithmetic in here, but I ask that you explain precisely what
you mean by "it calculates standar[d] deviation of entry size using
wrong formula" (the formula used may not look like the usual one, but
it is correct to the best of my knowledge) and also why this causes an
abort instead of garbage results. (Presumably the issue is integer
overflow?)
N.B. I am traveling for the next week and a bit and will not be
responding to email promptly.
zw