[cfe-dev] Interesting behaviour with UBSAN+libc++

Marshall Clow mclow.lists at gmail.com
Fri Dec 5 13:42:48 PST 2014


> On Dec 5, 2014, at 1:05 AM, İsmail Dönmez <ismail at donmez.ws> wrote:
> 
> 
> 
> On Fri, Dec 5, 2014 at 10:36 AM, Stephan Bergmann <sbergman at redhat.com <mailto:sbergman at redhat.com>> wrote:
> On 12/04/2014 10:39 PM, Richard Smith wrote:
> The UBSan runtime's support for type checks of virtual bases is new and
> possibly incorrect.
> 
> ...which reminds me of <http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20141201/247604.html <http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20141201/247604.html>> "[PATCH] [comiler-rt/ubsan] getVtablePrefix must not sanity-check on Prefix->Offset > 0"
> 
> Ok so just to confirm to use ubsan against c++ programs I do not necessarily recompile libc++ with ubsan support right? 

Correct.

— Marshall


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.llvm.org/pipermail/cfe-dev/attachments/20141205/f96d2fb8/attachment.html>


More information about the cfe-dev mailing list