In <4272A3FE(_dot_)740A(_at_)xyzzy(_dot_)claranet(_dot_)de> Frank Ellermann
<nobody(_at_)xyzzy(_dot_)claranet(_dot_)de> writes:
wayne wrote:
*YOU KNEW ABOUT IT*! grumble, grumble, grumble.
In hindsight, I hope it was clear that I was being mostly sarcastic
about the xml2rfc changes.
I've never tested xml2rfc on your xml. The only problem I'm
aware of are obscure symbolic character references like hellip
or for umlauted characters.
And an oddity with <vspace BlankLines="1" /> in tables, but as
far as I know that's no new 1.29 feature. So what problem did
you have with xml2rfc 1.29, the ToC ?
The two main problems with the newer xml2rfc are:
1) It is breaking lines in places I don't want it to. For example, it
in one case it is breaking the string "mx:example.com" across two
lines with one line ending in ``"mx:'' and the next line starting
with ``example.com"''. There are several places that it breaks a
hyphenated ABNF token like "target-name" across a line which I
think is very confusing.
2) The newer xml2rfc adds a lot more blank lines in <list>s, which
looks very ugly if it is just a list of short words. Also, Mark
Lentczner's original XML-ification of Meng's flat ascii file
contained a lot of strange <list> constructs that produced nice
looking output, but were making assumptions that were not
guaranteed to hold.
Actually, MarkL's xml produced nice results for the ascii text
files, but the html output had some rough edges. I'm working on
fixing up both cases.
Anyway, so far the only comments I have received are these minor
editorial fixes. Maybe the spec really is close to perfect, but I
suspect not. ;-)
-wayne