>>>>> "Leif" == Leif Johansson <[email protected]> writes:
Leif> The only question I have is that since this is a pretty scarce
Leif> resource we might want specification required + expert review
Leif> and give the expert pool explicit review instructions that we
Leif> are only looking for a "is not crazy"-check.
I don't think it is particularly scarce actually.
By requiring implementations to deal with error codes greater than 255,
here's what I'd do if I were using com_err:
1) have all mapping from protocol error to com_err code in one function
2) have the inverse in another function
3) Until we use up the first 3255 you can just add or subtract the
error table base.
4) If we get beyond tht then have a set of bases based on what IANA has
currently assigned.
So, I don't think it is particularly scarce and thus I think
specification required is good enough.
However I'd be equally happy with expert review.
_______________________________________________
abfab mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/abfab
