On Tue, Jun 15, 2010 at 02:34:15PM -0500, Patrick Goetz wrote: > On 06/14/2010 09:46 PM, Wesley Craig wrote: > >Seems like a reasonable suggestion. Where was it reported, exactly? > > > > These are patches from an experimental Debian build which can be found here: > > https://mail.incase.de/svn/cyrus22/branches/cyrus23/cyrus-imapd-2.3-hmh/ > > These package builds have been languishing for years as no one in > the Debian cyrus-devel group currently has time to work on them. > Previously, they were stalled because of concerns about how to > migrate people automatically from 2.1.x and 2.2.x. > > Because of all this cyrus is being abandoned by people using > debian/Ubuntu. After a month of testing (I use both cyrus and > dovecot on different servers), I've determined that cyrus is still a > better solution than dovecot in many situations and have taken on > the task of trying to get these packages up to production quality > for distribution. One of the tasks is going through dozens of > patches (some involving security issues) which have been around for > years. > > Right now I'm scratching my head over this one, as both these values > seem kind of arbitrary: > > ~/imap/imapparse.c > line 53: > - MAXLITERAL = INT_MAX / 20 > + MAXLITERAL = INT_MAX / 10
Yeah - that's odd... Hmm - and still untouched in my latest codebase too. We've got a configurable maximum line length now. That's basically what - 100M, changed up to 200M? Or am I out by an order of magnitude or so? Maximum size for a single email. I'd be happy to apply it (and also to replication...) - I don't see the danger with either option, but it will block super-large emails. Bron.
