[xml2rfc] feedback for <u>

[Julian Reschke <[e-mail]>]

Hi,

I recently spent some time on the implementation in rfc2629.xslt (not
ready yet), and here's more feedback.

1) Character Repertoire

By the way this is defined, the characters we can express are restricted
by XML. That is: no controls except HTAB, CR, and LF.

This *might* be considered a problem when documenting protocols where
controls matter, but I guess this isn't new, and the purpose of <u> is
mainly to address non-ASCII characters. But it would be interesting see
whether other people fell different.

I also noticed that the current version of xml2rfc fails in different
ways when the Unicode database doesn't have the expected entries. Filed
<https://trac.tools.ietf.org/tools/xml2rfc/trac/ticket/402> for C1
controls such as HTAB and
<https://trac.tools.ietf.org/tools/xml2rfc/trac/ticket/401> for C1 controls.

2) Format string

So we have two different format notations (see
<https://tools.ietf.org/html/draft-levkowetz-xml2rfc-v3-implementation-notes-08#appendix-A.1>).

It would be good to document how exactly a processor decides between
these two. Also, every time placeholders are used I routinely wonder
whether we might need a way to escape them...

Best regards, Julian