On Tue, Sep 7, 2021 at 8:07 AM Ruediger Pluem <rpl...@apache.org> wrote: > > On 9/7/21 3:52 AM, yla...@apache.org wrote: > > > > - /* Set counter in network byte order for the uuencoded unique id. */ > > - new_unique_id.counter = htons(new_unique_id.counter); > > + /* The counter is two bytes for the uuencoded unique id, in network > > + * byte order. > > + */ > > + new_unique_id.counter = htons(counter % APR_UINT16_MAX); > > Isn't '&' doing the same as '%' here and is more performant?
With a *literal* ^2-1 modulus, the compiler generates the exact same code (a '&'). The modulo is possibly more readable.. But yeah, if the modulus were a variable, the '&' is clearly more performant. Regards; Yann.
