Fix GnuTLS privatekey forced fail bug; in both TLS's treat an empty

[users/jgh/exim.git] / doc / doc-misc / WishList

diff --git a/doc/doc-misc/WishList b/doc/doc-misc/WishList
index e101901ad607d42a101e1881f70b40d527ac0bb9..8c058484ca86fca804e1648ae52807fae17db9c2 100644 (file)
--- a/doc/doc-misc/WishList
+++ b/doc/doc-misc/WishList
@@ -1,4 +1,4 @@
-$Cambridge: exim/doc/doc-misc/WishList,v 1.49 2005/09/15 08:27:59 ph10 Exp $
+$Cambridge: exim/doc/doc-misc/WishList,v 1.58 2006/02/13 15:09:30 ph10 Exp $

 
 EXIM 4 WISH LIST
 ----------------
 
 EXIM 4 WISH LIST
 ----------------


@@ -1928,14 +1928,6 @@ This is probably a longish-term thing at the moment. Quotas over 2G are now
 supported, but not individual messages; no doubt one day this will be wanted.
 ------------------------------------------------------------------------------
 
 supported, but not individual messages; no doubt one day this will be wanted.
 ------------------------------------------------------------------------------
 

-(335) 14-Jun-05 T Re-arrange default configuration
-
-A small niggle which might be worth fixing is the ordering of the ACL in the
-default configuration file. The relay_from_hosts and authenticated clauses
-would be better off before the dnslists examples. However, this should be left
-until a x.x0 release, because of the documentation implications.
-------------------------------------------------------------------------------
-

 (336) 16-Jun-05 M Show recipient(s) after header check failure
 
 The mainlog line for "There is no valid sender in any header line" shows the
 (336) 16-Jun-05 M Show recipient(s) after header check failure
 
 The mainlog line for "There is no valid sender in any header line" shows the


@@ -1979,5 +1971,61 @@ second. Maybe something like ${match_any{...}{...}} is needed.
 
 This would re-instate the behaviour prior to change 4.52/TF/02.
 ------------------------------------------------------------------------------
 
 This would re-instate the behaviour prior to change 4.52/TF/02.
 ------------------------------------------------------------------------------

---- HWM 341 ------------------------------------------------------------------
+
+(342) 26-Sep-05 T Log and maybe defer odd values for condition pre-condition
+
+Odd values for "condition" in an ACL cause it to defer. In a router, they are
+treated as "true". At least they should be logged in a router, and perhaps they
+should also defer, for compatibility with ACLs.
+------------------------------------------------------------------------------
+
+(343) 03-Oct-05 M A query-style lookup for scanning flat files
+
+The natural syntax for this would be to use a regex, like this:
+${lookup regex{/some/file regex}{found-string}{not-found-string}}
+------------------------------------------------------------------------------
+
+(344) 10-Oct-05 M Make debug_print work in authenticators
+------------------------------------------------------------------------------
+
+(345) 14-Oct-05 M Standardize rejection messages
+
+"The parsing for rejection lines is a bit of a mess, and fairly
+unmaintainable. Do you think it would be possible to standardise
+rejection/refusal log messages? How about something like:
+
+(<ID>|16 Spaces) *< (Connection|MAIL|RCPT|HELO|EHLO|DATA) rejected (from
+<Address>)?: <Reason> (\(<Detail>\))?"
+------------------------------------------------------------------------------
+
+(346) 20-Oct-05 S Set $domain and $local_part in retry matching
+
+Currently, these variables are unset. Make it like rewrite matching.
+------------------------------------------------------------------------------
+
+(347) 15-Nov-05 M Arrange to expand data from wildlsearch
+
+This would allow keys that are regular expressions to set up numerical
+variables that are included in the data. This has to be done inside the lookup
+code, because of caching. Probably means we have to invent ewildlsearch and
+enwildlsearch.
+------------------------------------------------------------------------------
+
+(348) 17-Nov-05 S Option to allow AUTH when not advertised.
+
+It seems that there are clients that send AUTH when it hasn't been advertised,
+some even after HELO, not even EHLO. Sigh. Possibly this should be an ACL
+control, to enable it to be restricted to certain hosts.
+------------------------------------------------------------------------------
+
+(349) 13-Feb-06 S Add login_cap support to the pipe transport
+
+FreeBSD (and maybe others) have some functions for a per-user resource database
+of which login_cap() is one. A copile-time macro (set for FreeBSD) could enable
+a new pipe transport option to cause setclassresources() to be called to limit
+the resources used. It's a bit tricky, because it has to be done as root, but
+in the forked transport process. There will have to be some higher-level code
+to deal with this.
+------------------------------------------------------------------------------
+--- HWM 349 ------------------------------------------------------------------

 ---------------------------- End of WishList ---------------------------------
 ---------------------------- End of WishList ---------------------------------