On 2015-08-04 09:48, Rob Austein wrote: > At Tue, 4 Aug 2015 07:06:22 -0400, Sandra Murphy wrote: >> Preferences for Richard?s option #2 (allow but do not mandate line >> breaks) or for Richard?s option #3 (mandate line breaks)? Note that >> option #3 means we have to settle on a max line length. > > Option #2 matches the running code. > > Absent some pressing need to make TALs fit on punch cards, the only > benefit that option #3 brings is the opportunity to declare running > code retroactively out of spec for trivial reasons.
I'm confused by your "out of spec" comment. In what way would option #3 render existing code out of spec where option #2 wouldn't? The advantage to option #3 is compatibility with OpenSSL's lame base64 implementation. My stated preference for option #3 should not be taken as a strong preference. I am OK with option #2: it's not much harder for people to implement (working around OpenSSL's problems is annoying but not difficult), it puts a tiny bit of pressure on OpenSSL to improve their implementation, and it avoids the need to decide on a specific maximum line length. -Richard _______________________________________________ sidr mailing list sidr@ietf.org https://www.ietf.org/mailman/listinfo/sidr
