On Jul 7, 2009, at 6:40 PM, Eli Friedman wrote: > On Tue, Jul 7, 2009 at 6:08 PM, Chris Lattner<[email protected]> wrote: >> - // FIXME: Vector compare support in the LLVM backend is not >> fully reliable, >> - // just reject all vector comparisons for now. >> - if (1) { >> - Diag(Loc, diag::err_typecheck_vector_comparison) >> - << lType << rType << lex->getSourceRange() << rex- >> >getSourceRange(); >> - return QualType(); >> - } > > Do we want to allow this for non-extvectors?
I don't have a strong opinion. The only bad part about allowing them for non-ext-vectors is if the gcc folks decide to do something differently in the future. This actually came up recently, and they are now leaning towards following opencl's lead since it is a standard (as opposed to the target-specific Cell vector bindings). What do you think? -Chris _______________________________________________ cfe-commits mailing list [email protected] http://lists.cs.uiuc.edu/mailman/listinfo/cfe-commits
