To clarify what I said there: 22-chars (128 bits) seems like way more than enough for a lower limit when using the "plain" challenge method. When using the "S256" challenge method, exactly 43 char (256 bits) should probably always be used.
On Thu, Feb 5, 2015 at 11:09 AM, Brian Campbell <[email protected]> wrote: > 22-chars (128 bits) as a lower limit seems just fine for this case. > > "ccm" works for me but I don't feel strongly about it either way. > > > > On Thu, Feb 5, 2015 at 9:49 AM, John Bradley <[email protected]> wrote: > >> Inline >> >> >> > On Feb 4, 2015, at 10:43 PM, Manger, James < >> [email protected]> wrote: >> > >> >> Title : Proof Key for Code Exchange by OAuth Public >> Clients >> >> Filename : draft-ietf-oauth-spop-09.txt >> >> https://tools.ietf.org/html/draft-ietf-oauth-spop-09 >> > >> > >> > Some nits on this draft: >> > >> > 1. 42 chars. >> > The lower limit of 42 chars for code_verifier: is not mentioned in >> prose (just the upper limit); is too high (128-bits=22-chars is >> sufficient); and doesn't correspond to 256-bits (BASE64URL-ENCODE(32 bytes) >> gives 43 chars, not 42). >> >> In my editors draft I fixed the 43 octet base64url encoding of 32bytes. >> I originally had 43 but it got changed at some point >> >> Is there working group feedback on making the lower limit clear in the >> prose and if so what should it be? 22-chars (128 bits) or 43 char (256 >> bits)? >> >> >> > >> > 2. >> > Quotes around "code_verifier" and "code_challenge" in prose are okay, >> though not really necessary as the underscore is enough to distinguish them >> as technical labels. Quotes around these terms in formula is bad as it >> looks like the formula applies to the 13 or 14 chars of the label. The >> quoting is also used inconsistently. >> > Suggestion: remove all quotes around "code_verifier" and >> "code_challenge" in prose and formula. >> > For example, change ASCII("code_verifier") to ASCII(code_verifier). >> > >> >> I am going to leave this for a later formatting cleanup at the moment, I >> need to find a good style compromise that works with rfcmarkup. >> >> > 3. >> > Two ways to check code_verifier are given in appendix B, whereas only >> one of these is mentioned in section 4.6. >> > SHA256(verifier) === B64-DECODE(challenge) >> > B64-ENCODE(SHA256(verifier)) === challenge >> > >> > I suggest only mentioning the 2nd (change 4.6 to use the 2nd, and drop >> the 1st from appendix B). It is simpler to mention only one. It also means >> base64url-decoding is never done, and doesn't need to be mentioned in the >> spec. >> > >> Yes when I added the example I realized that the normative text was the >> more complicated way to do the comparison. >> >> I will go back and refactor the main text to talk about the simpler >> comparison and drop the base64url-decode references. >> > >> > 4. >> > Expand "MTI" to "mandatory to implement". >> >> Done in editors draft. >> > >> > P.S. Suggesting code challenge method names not exceed 8 chars to be >> compact is a bit perverse given the field holding these values has the long >> name "code_challenge_method" ;) >> >> On the topic of the parameter name "code_challange_method", James >> has a point in that it is a bit long. >> >> We could shorten it to "ccm". If we want to change the name sooner is >> better than later. >> >> It is that balance between compactness and clear parameter names for >> developers, that we keep running into. >> >> I don't know that encouraging longer parameter values is the best >> direction. >> >> Feedback please >> >> John B. >> > >> > -- >> > James Manger >> > >> > _______________________________________________ >> > OAuth mailing list >> > [email protected] >> > https://www.ietf.org/mailman/listinfo/oauth >> >> >> _______________________________________________ >> OAuth mailing list >> [email protected] >> https://www.ietf.org/mailman/listinfo/oauth >> >> >
_______________________________________________ OAuth mailing list [email protected] https://www.ietf.org/mailman/listinfo/oauth
