On Mon, Oct 21, 2013 at 2:22 PM, Fred Akalin <akalin@google.com> wrote: > On Mon, Oct 21, 2013 at 2:15 PM, Martin Thomson <martin.thomson@gmail.com>wrote: > >> On 21 October 2013 14:03, Fred Akalin <akalin@google.com> wrote: >> > I'm not sure I see the problem. While decoding a varint, you have to >> keep >> > track of amount to right-shift the low 7 bits of the next octet. You can >> > then check if doing so would overflow 32 bits, and abort if so. >> >> If you want to use all the 32 bits, then you have to check what bits are >> set. >> > > I guess, but that doesn't seem too onerous? > > >> The bigger problem is the extra 2^N-1 you are required to add (255 for >> an 8-bit prefix), which will cause an overflow if you aren't careful. >> Hence the tricky little mask I used... >> > > Unless I'm grossly misunderstanding the format, the varint format is > little-endian, so you fill in the lower bits first, so this shouldn't be a > concern. > Egg on my face! I did misunderstand the format (my decoder was erroneous). I still contend that checking for overflow is straightforward, although doing it via masks seems too tricky.
Received on Monday, 21 October 2013 22:55:37 UTC