[cfe-dev] Clang Analysis of several open source projects.

[Ted Kremenek]

On May 13, 2011, at 10:33 AM, Kaelyn Uhrain wrote:

> Once I added the declaration "void fatal() __attribute__ ((noreturn));", clang's static analyzer no longer complained that x might be uninitialized.  I haven't looked into the code, but I suspect that either the analyzer isn't propagating the noreturn function attribute properly (in cases where the only return path in a function hits a call to another function marked as noreturn) or possibly that the attribute is missing in a key spot within a header.

Right now there is no propagation of the noreturn attribute because the analyzer isn't doing any inter-procedural analysis.

I have two thoughts on this:

1) Eventually do this propagation automatically.  This will not always be correct, however, because we can only infer this when the implementation of the function is available.

2) Warn when a function that never returns does not have the noreturn attribute.  I believe the compiler may already do this.

The benefit of (2) is that it also teaches the compiler that a function doesn't return when called.  This allows it to do a variety of optimizations in the caller.  Beyond whatever it gets from inlining, the compiler will never do full auto-propagation of the noreturn attribute.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20110513/59ff163f/attachment.html>