aboutsummaryrefslogtreecommitdiffstats
path: root/imap.c
diff options
context:
space:
mode:
authorEric S. Raymond <esr@thyrsus.com>2000-03-21 11:40:55 +0000
committerEric S. Raymond <esr@thyrsus.com>2000-03-21 11:40:55 +0000
commit1ae9b3525fcdbf175988b927f921b442412550fe (patch)
tree1f0a2c2f4952f50ac607e0f8356fc5ed5d2e84bf /imap.c
parentd87386367c61296ed1c648cfe11a96915d9257e4 (diff)
downloadfetchmail-1ae9b3525fcdbf175988b927f921b442412550fe.tar.gz
fetchmail-1ae9b3525fcdbf175988b927f921b442412550fe.tar.bz2
fetchmail-1ae9b3525fcdbf175988b927f921b442412550fe.zip
Work around InterChange.
svn path=/trunk/; revision=2830
Diffstat (limited to 'imap.c')
-rw-r--r--imap.c14
1 files changed, 14 insertions, 0 deletions
diff --git a/imap.c b/imap.c
index c9df435e..fe8bffe7 100644
--- a/imap.c
+++ b/imap.c
@@ -1246,15 +1246,29 @@ static int imap_fetch_body(int sock, struct query *ctl, int number, int *lenp)
* In that case, marking the seen flag is the only way to prevent the
* message from being re-fetched on subsequent runs (and according
* to RFC2060 p.43 this fetch should set Seen as a side effect).
+ *
+ * According to RFC2060, and Mark Crispin the IMAP maintainer,
+ * FETCH %d BODY[TEXT] and RFC822.TEXT are "functionally
+ * equivalent". However, we know of at least one server that
+ * treats them differently in the presence of MIME attachments;
+ * the latter form downloads the attachment, the former does not.
+ * The server is InterChange, and the fool who implemented this
+ * misfeature ought to be strung up by his thumbs.
+ *
+ * To work around this, we disable use of the 4rev1 form. It's
+ * all too easy to imagine other clever idiots deciding that the
+ * message body doesn't include attachments.
*/
switch (imap_version)
{
case IMAP4rev1: /* RFC 2060 */
+#ifdef __UNUSED__
if (!ctl->keep)
gen_send(sock, "FETCH %d BODY.PEEK[TEXT]", number);
else
gen_send(sock, "FETCH %d BODY[TEXT]", number);
break;
+#endif /* UNUSED -- FALL THROGH */
case IMAP4: /* RFC 1730 */
if (!ctl->keep)