Pages:<< prev 1 next >>


Black Hand

GroupAdministrators
Posts3,707
JoinedJan 1, 2002
Bug: Array boundary exceeded during socket buffer flush
Danger: High, with critical potential. Possible crashes or memory corruption due to array boundary overflow.
Discovered in: AFKMud 2.1
Found by: GCC 4.3
Fixed by: Samson
---
descriptor.cpp, descriptor_data::flush_buffer
Locate:
Change to:
GCC 4.3 issues the following warning:
Standard array buffer overflow. Guess it's never caused a problem before since it's never been reported. However NULLing memory outside of the array's boundary? Probably not such a hot idea.
Apparently some of the new warnings that have been enabled are actually useful.
Danger: High, with critical potential. Possible crashes or memory corruption due to array boundary overflow.
Discovered in: AFKMud 2.1
Found by: GCC 4.3
Fixed by: Samson
---
descriptor.cpp, descriptor_data::flush_buffer
Locate:
buf[4096] = '\0';
Change to:
buf[4095] = '\0';
GCC 4.3 issues the following warning:
descriptor.cpp: In member function 'bool descriptor_data::flush_buffer(bool)':
descriptor.cpp:528: warning: array subscript is above array bounds
Standard array buffer overflow. Guess it's never caused a problem before since it's never been reported. However NULLing memory outside of the array's boundary? Probably not such a hot idea.
Apparently some of the new warnings that have been enabled are actually useful.
Pages:<< prev 1 next >>