The crash is actually because of an unhandled exception caused by a
modulo by zero operation, and is reproducible if DC++ receives a GET
BLOM command with zero filter size (m) specified (or if parameter #3 of
the GET command is invalid).
As far as I see the (current) protocol extension does not restrict m=0,
however getting a zero sized filter seems illogical to me... so if m=0
is invalid then the fix is easy (send "Unsupported m"). If m=0 is a
valid value then some other action must be taken.
** Changed in: dcplusplus
Status: New => Confirmed
** Changed in: dcplusplus
Importance: Undecided => High
--
Crash when creating bloom filter
https://bugs.launchpad.net/bugs/571914
You received this bug notification because you are a member of
Dcplusplus-team, which is subscribed to DC++.
Status in DC++: Confirmed
Bug description:
[2010-04-20 23:18] <Sulan> http://pastebin.com/tVrwYCZF
[2010-04-20 23:18] <Sulan> DC++ didnt like my buggy bloom filter
[2010-04-21 16:26] <eMTee> Sulan, does this crash depend on a specific GET BLOM
command parameters or on a certain share the filter made from?
[2010-04-21 16:34] <Sulan> dont know, i only got it once
[2010-04-21 16:34] <Sulan> no time to test it more atm
[2010-04-21 16:35] <eMTee> if its reproducible it'd worth to test on a build
made with other than gcc 4.4.1 because the crash is in the 64bit modulo
function in ver. 4.4.1 gcc runtime lib...
_______________________________________________
Mailing list: https://launchpad.net/~linuxdcpp-team
Post to : [email protected]
Unsubscribe : https://launchpad.net/~linuxdcpp-team
More help : https://help.launchpad.net/ListHelp
