On 2013-09-09 11:29:28, Steve Beattie wrote: > On Fri, Sep 06, 2013 at 11:57:51AM -0700, Tyler Hicks wrote: > > Nice catch! (and shame on me for not using valgrind) > > Well, we don't really make it easy to run valgrind regularly. In the > interest of fixing that, attached is a simple script for doing so, > along with a separate patch for fixing a few leaks that it found; > most occur in error paths, though there's a couple of legitimate ones > it found in parser_yacc.y. There are also more reported leaks that > I have not fully investigated yet.
Great idea! > > It's also finding what may be some false positives around > invalid reads around profile serialization (e.g. on > simple_tests/conditional/else_if_5.sd) and conditional > jump or move depends on uninitialized value(s) (e.g. > simple_tests/network/network_ok_1.sd) that need more investigation, > to determine whether it's a legitimate issue or a valgrind check that > hopefully can be disabled. The memory leak fixes get my Acked-by: Tyler Hicks <[email protected]> I rebased the C++ patches on that patch, so please go ahead and commit it when you get a chance. I need some more time to look at the valgrind test patch. It doesn't like the C++ conversion and ever test was failing before I killed it. I'm not immediately sure what is happening. Running the parser through valgrind on the same testcases shows no issue. I'll have to look into more. Tyler
signature.asc
Description: Digital signature
-- AppArmor mailing list [email protected] Modify settings or unsubscribe at: https://lists.ubuntu.com/mailman/listinfo/apparmor
