Scott Gifford writes:
I was going through an old install of Courier-IMAP, and found a patch
I hade mad way back when (to courier-imap-0.27) at what is now
rfc2045.c:495. I changed:
if ((newp->mime_version=strdup(p->mime_version)) == 0)
rfc2045_enomem();
to
if (p->mime_version) {
if ((newp->mime_version=strdup(p->mime_version)) == 0)
rfc2045_enomem();
} else {
newp->mime_version=NULL;
}
to fix a crash a user was seeing with no mime-version header, or an
empty one, or a malformed one, or something...I can't reproduce the
problem with the latest version (I no longer have the message that
caused the problem), but I couldn't find an obvious guarantee that
this line of code would never be reached with p->mime_version set to
NULL.
Is there some reason why this is guaranteed never to happen, or should
a check for NULL be put in?
A message is not recognized as a MIME message unless the Mime-Version:
header is present, and additional MIME section are not parsed unless the
message is considered as a MIME message.
2 messages in net.sourceforge.lists.courier-users[courier-users] Re: An old patch at r...
