Stephen J Turnbull <turnbull(a)sk.tsukuba.ac.jp> writes in xemacs-beta(a)xemacs.org:
When attempting to preview the attached message (and anything else
with that X-Mailer header), VM barfs with
Signaling: (error "Invalid search bound (wrong side of
point)")
...
I am guessing that since the Mule representation of the Japanese for
"Nikko" is much shorter than the MIME encoded word, the search fails.
Possibly. It might also be due to how the line gets broken up. My
Author Copies look like:
X-Mailer: Gnus v5.6.45/XEmacs 21.0(beta65) - "20 minutes to
=?ISO-2022-JP?B?GyRCRnw4dxsoQiI=?=
Gnus + TM handles this O.K.
There is a possibly related bug in which VM does not find successive
encoded words in the same header. Eg,
Subject: =?ISO-2022-JP?B?GyRCRnw4dxsoQiI=?=
=?ISO-2022-JP?B?GyRCRnw4dxsoQiI=?=
should be translated to
Subject: <NIKKO>"<NIKKO>"
(the <NIKKO> is actually Japanese characters, the double-quote
mark is an
artifact of my cut-and-paste) but instead becomes
Subject: <NIKKO>"
=?ISO-2022-JP?B?GyRCRnw4dxsoQiI=?=
I suspect this is an XEmacs bug, since the backtrace says that the
bound is a marker, which ought to adjust for MIME munging. Since I'm
not sure, I'm reporting to both lists.
Do you have an example of an email message that bombs that I can test
against Gnus?