[llvm-commits] [llvm] r62528 - in /llvm/trunk: lib/Support/APFloat.cpp test/CodeGen/X86/2009-01-19-fmod-constant-float.ll

Chris Lattner clattner at apple.com
Mon Jan 19 14:04:42 PST 2009


On Jan 19, 2009, at 1:17 PM, Dale Johannesen wrote:

> Author: johannes
> Date: Mon Jan 19 15:17:05 2009
> New Revision: 62528
>
> URL: http://llvm.org/viewvc/llvm-project?rev=62528&view=rev
> Log:
> compile-time fmod was done incorrectly.  PR 3316.

Thanks Dale!

> +++ llvm/trunk/test/CodeGen/X86/2009-01-19-fmod-constant-float.ll  
> Mon Jan 19 15:17:05 2009
> @@ -0,0 +1,77 @@
> +; RUN: llvm-as < %s | opt -std-compile-opts | llc | grep 1036831949  
> | count 2
> +; RUN: llvm-as < %s | opt -std-compile-opts | llc | grep 3184315597  
> | count 2
> +; check constant folding for 'frem'.  PR 3316.

There is no reason to run all of -std-compile-opts here, and doing so  
will make the test depend on *what* std-compile-opts includes.  Can  
you please reduce this to just running the constant folding pass or  
something?  Also, this is not an x86-specific test so it would be nice  
to run just constant folding + llvm-dis instead of running llc.

-Chris




More information about the llvm-commits mailing list