Murray (as AD), Re: https://www.rfc-editor.org/authors/rfc9755.html
Would you please offer guidance on whether to add the "Note that when" paragraph in Section 7 (pasted below)? My understanding is that we do not have consensus among the authors on whether to add it (specificially, the authors are 1 in favor, 1 against, and 1 OK either way). The WG chair (Bron) wrote "no strong preference either way". To review the mails during AUTH48, please see: https://mailarchive.ietf.org/arch/search/?as=1&email_list=auth48archive&end_date=&q=subject%3A%289755%29&qdr=a&start_date= Thank you. RFC Editor/ar -- > On Feb 10, 2025, at 2:34 AM, Arnt Gulbrandsen <a...@gulbrandsen.priv.no> > wrote: > > Hi, > > I have reviewed the changes you made; very nice work. Thanks. > > I would like to insert one additional paragraph, to clarify a nonobvious > consequence. In section 7, the second-to-last paragraph starts with "All IMAP > servers that support "UTF8=ACCEPT" SHOULD accept UTF-8 in mailbox names". > Please insert the following paragraph after that paragraph: > > Note that when mailbox names comply with with the Net-Unicode Definition > ([RFC5198] as described in the previous paragraph, they cannot any longer > comply with the modified UTF-7 convention described in [RFC3501]. This > implies that once UTF8=ACCEPT is enabled, neither clients nor servers may > send mailbox names using modified UTF-7, they may only send UTF-8. (The same > applies when IMAP4rev2 has been enabled. "A&-B" is an example that shows the > conflict clearly.) > Arnt > On Feb 12, 2025, at 1:58 AM, Arnt Gulbrandsen <a...@gulbrandsen.priv.no> > wrote: > > Pete Resnick <resn...@episteme.net> writes: >> This seems like a substantive change in the document and while I think it is >> probably right, I think it is sufficiently covered by the first paragraph of >> section 7 and I wouldn't want to make this change without WG approval, not >> in AUTH48. Arnt: Do you really want to hold up the document for this? > > FYI I sent mail to the WG list on 2025-01-27, two weeks before AUTH48. That > was late, so search for "late" in the subject ;) > > I did that because a maintainer had pointed out to me that my pull request > for their code might not be compliant, according to his reading of the RFC. I > read 6855 and thought he had a point. The existing text is clear enough for > all of the IETF regulars, but I can see how someone like he might wonder what > the correct behaviour is. > > When this happens I'm in a bind. I may not get my code merged with ANY > behaviour, because the maintainer won't see ANY behaviour as clearly covered > by the RFC. > > So by now I lean towards making it abundantly clear. Not just "sufficiently > covered", but abundantly. Does this make sense? > > Arnt > -- auth48archive mailing list -- auth48archive@rfc-editor.org To unsubscribe send an email to auth48archive-le...@rfc-editor.org
