The second beta of version 1.60 has been available for about two weeks now and there haven't been any reports of evil bugs yet, so it's probably time to make 1.60 the current release. It's on http://www.aconet.org/tools/ and on http://www.aconet.nl/tools/ (the old site may disappear before the end of October - anyone linking to aconet.org, please change your link). Some notes about the changes since the last revision of 1.59: New features ------------ Modified ID checking Set checking for deferred and headered messages separately. Expanded User Action Header Presenting the rule used through the X-AntiSpam-Action: header. SendWhitelist To add e-mail addresses of recipients of outbound mail to a whitelist. Block list checking Using a seperate module (NSQ) because all current RISC OS resolver modules are broken. More logging Optionally log To and From headers - and Weight - to the logs of deleted and diverted messages. Optionally log messages which turn out smaller than their listed length. Forget flag and button To make the program forget any previously remembered deferred and headered messages. Clean ups --------- - Redesign of the marking mechanism. UIDLs are now the default. The file with marking info is now called IDs instead of TFSM and the expiry time is no longer needed because the program checks the presence of the messages in the IDs file on a subsequent run. See 'An extra word of warning' below. - All marking related options are back on the Scan frame of the Choices window. - Details of known mail transports in Resources.Transports. - Lists scan no longer case sensitive. - Socket_Select calls because repeated calls to Socket_Connect seems to cause problems for (some versions of) VRPC. - Added version check for resource files. Moved the contents of the Docs directory to Resources. Bug fixes (includes bugs fixed between alpha/beta releases of 1.60) --------- - When trialling, very long header lines could cause an error. - Malformed encoded word (see RFC2047) could make the program crash. - Initial NSQ_TestZones could cause a crash if the DNS was unavailable. - Memory allocation routines didn't reserve enough scratch space after the last area. - A single dot as the 251th character of a line would cause problems. - Flushing a skipped message could cause a crash. See the StrongHelp manual for details. As always, let me know if you find anything wrong with it. An extra word of warning... =========================== The IDs file and the old TFSM file have a slightly different internal format. That's the main reason I decided to use a different name. This change will allow you to switch from version 1.59 to 1.60 and back, but there is a snag. Once you've used 1.60, the TFSM file may have lost sync because it will most likely still contain references to messages that have already been dealt with. Also, because 1.60 does not use the TFSM file in any way, its marking window may present messages you've already seen while using the older version of the program. Regards, Frank