[ On Thursday, May 6, 2004 at 16:37:06 (+0100), Philip Hazel wrote: ]
> Subject: Re: [Exim] verify=header_syntax Buffer overflow (CAN-2004-0400)
>
> > My suspicion is that Some Other MTA (tm) allows it, just like it allows
> > spaces between the name and the colon in alias files.
>
> Thanks to google, I found it in an online Sendmail manual at
The requirement actually dates back over a quarter of a century to the
long obsolete RFC 733. RFC 822 (itself almost 22 years old and also now
obsolete as well) explicitly did not allow any linear whitespace either
in or after the field name, yet for some unknown reason the authors of
RFC 2822 found some reason to _partially_ resurrect it again (though as
a deprecated feature which "MUST NOT" be generated but which "MUST" be
accepted) (i.e. to allow LWSP-chars only _after_ the field name and
prior to the colon):
[[ RFC 733 ]]
c. Field-names
To aid in the creation and reading of field-names, the free
insertion of LWSP-chars is allowed in reasonable places.
Rather than obscuring the syntax specification for field-name
with the explicit syntax for these LWSP-chars, the existence
of a "lexical" analyzer is assumed. The analyzer interprets
the text which comprises the field-name as a sequence of
field-name atoms (fnatoms) separated by LWSP-chars
Note that ONLY LWSP-chars may occur between the fnatoms of a
field-name and that CRLFs may NOT. In addition, comments are
NOT lexically recognized, as such, but parenthesized strings
are legal as part of field-names. These constraints are
different from what is permissible within structured field
bodies. In particular, this means that header field-names
must wholly occur on the FIRST line of a folded header item
and may NOT be split across two or more lines.
[[ RFC 822 ]]
C.1.1. FIELD NAMES
These now must be a sequence of printable characters. They
may not contain any LWSP-chars.
--
Greg A. Woods
+1 416 218-0098 VE3TCP RoboHack <woods@???>
Planix, Inc. <woods@???> Secrets of the Weird <woods@???>
