VooDoo cIRCle

File Release Notes and Changelog

Release Name: 1.1.beta27

Notes:
Release version 1.1.beta27
Date: 2007-03-31
*** WARNING *** Do not forget to update your "lang01.txt" file in directory "lang" as well, since it has been changed. (If you are upgrading from previous version.)
Bug fix: Various off-by-one errors (without buffer overflow) in sent BOTNET packets could lead to BOTNET connection close after received packet sanity check, if there is overly long string, such as overly long channel name in user definition in logic.txt, or overly long partyline channel name.
Bug fix: Various off-by-one errors in DNS resolution, if malicious DNS server returns DNS name that has exact length as buffer, could lead into buffer reading overflow.
Bug fix: Mostly *NIX: Crash after unsuccessful thread creation (e.g. resource limits), now it goes to queue and tries to create it later.
Bug fix: *NIX: After bad password on SSL connection of client, next client was rejected due to OS's reuse of FD for the next socket.
Bug fix: After an error in SSL connection, socket could not have been closed (resource leak).
Bug fix: When BOTNET connection did not succeed to link, there could be rarely forgotten to close socket (resource leak).
Bug fix: Documentation: Adjusted documentation, there was forgotten obsolete "access_to_procedure" access right, now it is handled in another (simpler) way.
Bug fix: File "lang/lang01.txt": Fixed bug, editing pass.txt, ".5" for editing remote password, not local.
Semi-bug fix: On telnet-SSL client connection (telnet), there was doubled message about accepting client's certificate.
Improvement: Adjusted behavior of character case mapping (see documentation).
Improvement: On telnet, and telnet-SSL connection, all client connections now linger (bot waits few seconds before destroying socket to allow SSL layer to transfer send queue).
Improvement: When on telnet-SSL connection maximum send queue is exceeded, the buffer is discarded and that fact is reported (not disconnected as in past versions).
Improvement: When there is timer with "0:00:00:00" duration, we treat it now like it were "0:00:00:01" (because this mistake could consume a lot of CPU, and, possibly, an infinite loop).
Improvement: Improved speed of evaluation of functions in internal scripting by detection if result of pre- and post-increment and -decrement operators is really needed.

Changes: