Yasca Report

Everything about development and the OpenMW source code.
User avatar
Site Admin
Posts: 1193
Joined: 05 Aug 2011, 22:21
Location: Wroclaw, Poland

Yasca Report

Post by lgromanowski » 13 Aug 2011, 20:00

EmbraceUnity wrote: Dunno if this is helpful to you or not. Hope so.

http://embraceunity.com/wp-content/uplo ... 54536.html

(EDIT: Updated with new RATS and CppCheck results)
Zini wrote: Well, most of it is just pointless gibberish, but some of the points listed may be valid. Among others I see four potential memory leaks in the sound system. Unless it is a false positive, this should be taken care of.
Thanks for bringing it to our attention.

Edit: The stuff about uninitialised variables in loadland.hpp looks rather bad. I think cnt just needs to be set to 0, but I don't know this part of OpenMW well enough to handle it offhand. Nico, you might want to take a look.
nicolay wrote: Wow, never heard of Yasca before. That was a (potentially) very useful list. Most of the points are just telling you places where you need to be careful though, and didn't say anything new (however it doesn't seem to like strlen() - and that is probably a valid point but not a big concern.)

The uninitialized cnt was a pure bug, now fixed. A mistranslation from D, which always initializes vars. The memory "leaks" in the sound files are only in the test programs and are completely unimportant.
best regards,