The spec considers one octet to be one octet in length. It specifies limits in terms of bytes on the wire, not in terms of characters, code points, or glyphs.
On Sun, Dec 29, 2019 at 2:28 PM Oleg Kalnichevski <[email protected]> wrote: > On Sun, 2019-12-29 at 15:32 -0500, Gary Gregory wrote: > > There was a PR flying around about string length and if × 2 was right > > or > > not. But with Unicode the string length might not be the same as the > > codePointCount... > > so if want want to count what to put on the wire using code points > > might be > > correct. > > > > Gary > > It got dropped in favor of simpler approach based on assumption that > HTTP/2 spec considers one char to be one octet in length. > > Oleg > > > --------------------------------------------------------------------- > To unsubscribe, e-mail: [email protected] > For additional commands, e-mail: [email protected] > >
