The upcast/downcast tests might want to be codegen tests to ensure that, not only do we not fail to compile it - but we compile it correctly (not emitting any calls to RTTI stuff, etc).
On Mon, Jul 29, 2013 at 9:48 AM, Arnaud A. de Grandmaison <[email protected]> wrote: > With the patch :) > > On 07/29/2013 06:47 PM, Arnaud A. de Grandmaison wrote: >> On 07/29/2013 06:17 PM, Jean-Daniel Dupas wrote: >>> Le 29 juil. 2013 à 17:52, Arnaud A. de Grandmaison <[email protected]> >>> a écrit : >>> >>>> Clang emits an error when typeid is used with -fno-rtti, but does not >>>> perform a similar check for dynamic_cast. >>>> >>>> The attached patch implements the check. >>>> >>>> I have done it late in 'CastOperation::CheckDynamicCast' , as some >>>> dynamic_cast can be resolved statically, and will thus require no >>>> support from the runtime, but we may also want to be more paranoïd. >>>> >>>> Any thoughts ? >>> If we want to support dynamic_cast without rtti when possible (as you do in >>> your patch), you may add such case in the test to make sure clang accepts >>> it. >> Good point. I added a test for the 2 cases we accept (upcast, and cast >> to same type). >> >>> -- Jean-Daniel >>> >>> >>> >>> >> > > > -- > Arnaud A. de Grandmaison > > > _______________________________________________ > cfe-commits mailing list > [email protected] > http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits > _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
