The revisions are small changes that don't apply to every one. If they did, there would be a version 1.04. The changes were made and the .gz replaced so first time downloaders would have them. So the .gz you see at this site, is the last revision listed here and includes all the changes of all the revisions. revision 1 act_string for smaug102 uses 5 parameters, smaug14 uses 6 a conditional compile compile statement was put in to accommodate the difference. Added a revision number display on the ISTAT command. No need to download revision unless you are SMAUG14, Its probably easy to apply the fix by hand, the fix was posted on the IMC BBS. at http://moodyg.erols.com/imc/ revision 2 By request, the process_rfinger was set up like the process_rwho so you could keep it in your act_info.c and not have to redo it when you upgrade. Also there was a problem with free() that showed up on some OS's compile. Statements like if (pointer) free(pointer) causes problems if free(pointer) doesnt NULL out the pointer. pointers that are freed are now set to NULL in the code. You only need to download a new version if you are experiencing instability / crashes. revision 3 On input/output buffer overflow, the buffer is now flushed instead of disconnecting the mud. The assumption that it is an intentional spam is more false than true. Hub owners should get a new version. revision 4 I over compensated on the buffer flushing, my bad, this fixes it. I you have downloaded revsion 3 , then get revision 4. revision 5 Problem with the original spam protection routine and the number of version .10 muds/hubs on the network. You need this version if you a hub owner, or you are plagued by input buffer overflow log messages. revision 6 Nothing new except support for Rot 1.4 Muds No need for anyone to upgrade. revision 7 Added conditional compiles for FreeBSD Some optimization, garbage collecting and memory No need for anyone to upgrade unless you have memory (RAM) concerns. revision 8 Changed the IMC mud code to never use shared memory string functions like SSM. I ran into a problem with SSM plus IMC should have never used this type of string handling as no two packetes are likely to be alike, no sense in eating up the mud's hash string space. Muds that have a ssm.c module should download this revision.