Bruce Lilly wrote:
I've updated the revised grammar file, which is available from
http://users.erols.com/blilly/mparse/rfc2822grammar_simplified.txt
and
http://mysite.verizon.net/vze74u4p/mparse/rfc2822grammar_simplified.txt
Changes include:
1. aforementioned issues hopefully fixed
2. corrections to date-time syntax as discussed here
3. separated obs- and non-obs rules for clarity
4. rearranged so that most productions are defined before use, the only
exceptions being "comment" and "ccontent" (and variations), which are
mutually dependent. This should make the revised grammar easier to
read.
5. incorporated alternatives for handling encoded-words and for ignoring
RFC 2047
6. yanked out the Subject field hacks
A side effect of dealing with the FWS-at-field-end issue is that much of
the CFWS use has percolated up to higher-level constructs (further down in
the file). I think that makes it easier to understand, but maybe that's
just because I've been working on the file for much of the past week.
As Pete has said that he doesn't want to incorporate the encoded-word
syntax, if there is any discussion about that, please start a separate
thread. The encoded-word syntax is a bit complex; it is certainly possible
that there are errors or that it can be simplified. [F]or the purpose of
discussing the RFC 2822 successor, please concentrate on the alternatives
that do not address the encoded-word issues.
I would like to request help in verifying the following:
A. that the changes that we've discussed have been correctly incorporated,
including the unstructured field grammar and the FWS-at-field-end issues
B. that the revised grammar (ignoring encoded-words and related issues) is
compatible with RFC 2822 except in areas where 2822 has errors
C. that no new errors have been introduced
There are notes in the file regarding some differences from RFCs 821, 822,
and 2821, but let's defer discussion of those issues until after we've
verified that we have a grammar that is correct and consistent with things
as they stand. Then we can discuss what should be changed (that will cleanly
separate the issue of rearranging the grammar from changing the grammar per
se).
I've found a few typos and encountered another set of conflicts in the past
week, so I've once again updated the file, available at the URIs above. One
set of conflicts was related to "phrase" and resolving that moved many of the
[CFWS], [FWS] instances back into the lower-level productions. I've added
more notes on issues that have been brought up on this mailing list. Once
again I request that others look over the file to check for completeness
and lack of errors.