hey Robert, Thanks for fixing this so quickly! I just realized I failed to attach my patch, but looks like you had no trouble dealing with it.
I just wanted to clarify a couple of things, in case you or other readers of this report become a recipient of these bugs in the future. On Mon, 27 Oct 2008 21:53:39 +0100, Robert Millan said: > By chance these were, in fact, meant to be pointers so no harm done. I'll > wipe out that warning for your convenience. Actually, this is the problem. Without a prototype, gcc will assume that the function will return an int - but in fact it is returning a pointer. This causes the pointer to get truncated on the aformentioned architectures, and attempting to use this pointer will cause a segfault. On amd64 you can sometimes get away with it because the upper 32 bits would've been 0s anyway but, on ia64 at least, this is never the case. In other words, the warning is that the returned pointer is getting converted to something else, not the other way around. > severity 503698 minor There are cases where the returned pointer is ignored or that the code in question is guaranteed to never execute. Because of this, I used to always file these bugs as "important" unless I'd verified that the case will actually occur. However, the ia64 buildds have since been configured to reject any builds that fail this test. Whether or not that is a good idea is arguable but, as things stand, these issues are considered FTBFS defects and therefore warrant an RC severity. Anyway, hope that didn't come off too much like a lecture :) Thanks again for fixing this issue, and for your work towards solving the dsp56k fw problem. -- dann frazier -- To UNSUBSCRIBE, email to [EMAIL PROTECTED] with a subject of "unsubscribe". Trouble? Contact [EMAIL PROTECTED]
