bug in ggc-page?

[Zack Weinberg]

Mark Mitchell <mark@codesourcery.com> writes:

>> The algorithm in use deliberately applies integer overflow on multiply
>> to perform division.
>
> OK, so you're saying Daniel isn't seeing a real problem, right?

I'm saying that, while there may well be something wrong, the
apparently overflowed values in the inverse_table for orders that came
from the extra_order_size_table (which is where all the
non-power-of-two bucket sizes come from) are not a symptom thereof.

zw