I can't find more details, but I read this at #themanaworld of irc.themanaworld.org: <{4144}> 0.5 just released as new version <{4144}> master branch is heavy development. some thing may not work, and no new features <dark> 0.5 is 0.0.29.1? <dark> i was using the "older" client, but i think the one from git (from master branch) is working better.. <{4144}> no 0.5 is mana 0.5 and not released tmw 0.30 <{4144}> 0.0.29.1 better not use, if you care about security <dark> why? 0.0.29.1 is the version i have at gentoo. so it has a vulnerability? <dark> is there some stable version other than 0.0.29.1? (so that a gentoo ebuild could be made for it) <{4144}> yes, and many many crashes and some freezes <{4144}> also if you have opengl support, any mana versions will work more faster <dark> i came here to report a crash, but it isn't present at the git master <dark> do you have a link for this vulnerability? also, what version should gentoo provide? <{4144}> yes, in 0.0.29.1 many bugs, and no one will fix it. now 0.0.29.1 is outdated version <dark> but what is the latest stable then? <{4144}> 0.5 <{4144}> it will be released after some time as tmw 0.30 <{4144}> now like advertisement. if you want more features, use manaplus. officialy not supported version <dark> where can i find about this 0.0.29.1 vulnerability? (to link at the gentoo bug report) <{4144}> no links. some info may be in commits only <dark> so I will paste this irc log ok :P I'm currently running branch 0.5 from git://gitorious.org/mana/mana.git , that is overall better / less buggy (0.0.29.1 has some segfaults). It is still to be released. If there is really a vulnerability, options would be to mask this and wait for tmw 0.3 be officially released, and maybe to make a ebuild targeting this new version either from git, or from http://manasource.org/files/mana-0.5.0.tar.gz (from http://manasource.org/downloads )
So far I see a lot of hand-waving with no real proof or information.
(In reply to comment #1) > So far I see a lot of hand-waving with no real proof or information. > FWIW, these could be security related, but these are just from scanning the commit log. I have no _real_ data. <-- commit 99c83e7702cc19fe1b1b5ff72499a2f97e98ef7e Author: Thorbjørn Lindeijer <thorbjorn@lindeijer.nl> Date: Sun Jan 2 22:17:14 2011 +0100 Fixed possible crash when receiving player IP Could happen sometimes for GMs when the message is received either before the player instance is created or after it was destroyed. Also changed to a dynamic_cast, putting a little less trust in the server. Reviewed-by: Freeyorp <-- commit cabc69deb2771c1e88fc1f540eb10e89ada8897c Author: Andrei Karas <akaras@inbox.ru> Date: Sat Dec 25 23:24:27 2010 +0200 Fix crash with corrupted links in chat. Reviewed-by: Kage <-- commit c5e341f7b5f5e4efcabd09e00e177bb873db8097 Author: Chuck Miller <shadowmil@gmail.com> Date: Sun Oct 31 14:14:26 2010 -0400 Fix memory leak of localplayer classes in the charserverhandler They are initialized else where in the code Reviewed-by: Jared Adams <-- commit a9bcae9f81d97f67ab21cd26a099449cf0746137 Author: Andrei Karas <akaras@inbox.ru> Date: Tue Oct 19 01:52:47 2010 +0300 Fix crash in drawing incorrect strings. Reviewed-by: cody <-- commit c3370ee4b57d18a62510240f105791ca3b403224 Author: MadCamel <madcamel@gmail.com> Date: Thu Oct 14 10:38:04 2010 +0200 Disable showing of IP addresses while taking a screenshot This could potentially leak confidential information. Reviewed-by: cody <-- commit 000c3799eb85234b9baf5ede86dd966d929112d8 Author: Stefan Dombrowski <stefan@uni-bonn.de> Date: Mon Aug 23 17:28:05 2010 +0200 Fixing segmentation fault in partyhandler Segmentation fault happens if a player leaves your party while he is out of sight. Signed-off-by: Jared Adams <jaxad0127@gmail.com> <-- commit 04e4f0da4b6242f39d09794edad394587ddea620 Author: Andrei Karas <akaras@inbox.ru> Date: Fri Jul 16 23:34:17 2010 +0300 Fix memory corruption in chat on exit. Signed-off-by: Jared Adams <jaxad0127@gmail.com>
(In reply to comment #1) > So far I see a lot of hand-waving with no real proof or information. You're right. I should have done my homework, sorry. I was told that if I wanted to see, he could crash my client. This made me worried :P But I also found it hard to believe they knew about some exploit and haven't advertised how to mitigate it, and/or released fixes. (In reply to comment #2) > FWIW, these could be security related, but these are just from scanning the > commit log. I have no _real_ data. That code were committed after 0.0.29.1 (that is from 10 jun 2009). But I don't know if they was running at 0.0.29.1 client, or if they fix something on unreleased software.
(not sure if I should have done that, but I marked it as NEEDINFO, and changed the title to reflect my lack of information about this)
