Over six months ago, I've reported nasty and easily exploitable overflows in lynx while parsing some URLs - like cso://AAAA... etc. I've given some examples, and it was fixed, but then, month later, I've realized that other protocols, not mentioned in previous post are still buggy in exactly the same way. Another post resulted in patched lynx release. And what now, guess?... Similar problems are present for example when lynx is using proxy server (often sysadm puts proxy server settings in global lynx.cfg) - even in recent 2.8.3dev2x releases - http://AAA... or ftp://AAA... requests with over 2 kb of junk after protocol indentifier (instead of valid hostname) - 0x41414141 SEGV - old, good, exploitable overflow while preparing request for proxy server. AND MORE FOLLOWS - for example some overflows when viewing 'Information about current document' and so on, all related to extremely long URLs. I'm not going to give more examples here, as I'm afraid I might miss one or two that won't be fixed - developers, use your head, take a look at the code and fix every suspected piece of code, not only already published / described bugs. _______________________________________________________ Michal Zalewski * [lcamtufat_private] <=> [AGS WAN SYSADM] [dione.ids.pl SYSADM] <-> [http://lcamtuf.na.export.pl] [+48 22 551 45 93] [+48 603 110 160] bash$ :(){ :|:&};: =-----=> God is real, unless declared integer. <=-----=
This archive was generated by hypermail 2b30 : Fri Apr 13 2001 - 15:37:40 PDT