Alexandre,
In the light of the note Henrik posted after I started to write
this, the note that follows may be irrelevant, but possibly
worth reading anyway.
In your tour of the problems that you have discussed on-list,
you have identified a problem or three with xml2rdc that
deserves a trouble report. I hope for better, but even if the
only effect of that report is to make sure that bugs of xml2rfc
v2 do not get carried forward into the production version of
xml2rfc v3, that would be a positive step.
I trust you are not surprised that these things, especially the
file name and "suffix"/"file type" stuff have been noticed by,
and irritated lots of others. We have just (eventually) learned
what the problem is and then decided that remembering to not
cause it is less trouble than asking for fixes (and maybe being
necessary anyway).
One problem is that such a complaint must be written carefully
to avoid causing other problems. For example, if we were to
complain that, in the <rfc> element,
draft-<stuff>-nn
is accepted but
draft-<stuff>-nn.txt
is not and produces incomprehensible messages, we might get a
fix that would clearly reject not only the second case but names
that do no9 contain exactly two digits before the end of the
string. That would be a disaster for those of us who use
assorted naming tricks to keep track of intermediate versions
that are not intended for posting and that are intended to be
seen only by the author(s) and a small circle of friends until
there have been enough iteration in that group to make them fit
for posting.
You have also uncovered a great deal of confusion about things
that may be worth getting documented (or at least flagged).
For example, when you complained about not being able to submit
an XML file for posting without a text one, I replied that the
text one was required. It was at one point, but I was wrong; it
seems that the tool will now accept XML and compile it. I make
sufficiently many XML mistakes that I haven't had the nerve to
try that or find out about it.
Other comments in response to your notes have implied that the
submission tool requires that the names of the file names that
are uploaded must match the names that will be used in the
Internet-Drafts directory. That isn't true. For example, if I
have properly structured files for
draft-ietf-example-example-00.txt and
draft-ietf-example-example-00.xml
but keep them in my file system as, e.g.,
examp-example-00.txt abd
examp-example-00.xml
the system will upload those files without problems an just
proceed normally. This is a big advantage for those who use
personal file naming conventions (perhaps, as above, for
sub-version tracking and perhaps just because they are lazy
about typing) and so on.
Anyway, I cannot do more with this until mid-next-week, but, it
you would find it helpful to work with me on a coherent list of
specific issues (in either code or documentation) rather than
just making iterative complaints of the IETF list, please get
back to me.
best,
john
