[Lldb-commits] [lldb] r198962 - Fixed the assertion test case.

[Ed Maste]

On 10 January 2014 17:22, Greg Clayton <gclayton at apple.com> wrote:
> Author: gclayton
> Date: Fri Jan 10 16:22:44 2014
> New Revision: 198962
>
> URL: http://llvm.org/viewvc/llvm-project?rev=198962&view=rev
> Log:
> Fixed the assertion test case.
...
> +            # Don't expect the function name to be in the disassembly as the assert
> +            # function might be a no-return function where the PC is past the end
> +            # of the function and in the next function. We also can't back the PC up
> +            # because we don't know how much to back it up by on targets with opcodes
> +            # that have differing sizes
>              self.expect("disassemble -a %s" % frame.GetPC(),
> -                substrs = [frame.GetFunctionName()])
> +                substrs = ['->'])

The test fails on FreeBSD after this commit for the reason described
in the new comment: the PC is past the end of __assert, and the
disassembly stops at the end of that function and so does not output
'->'.  I've XFAILed the test on FreeBSD for now, llvm.org/pr18533.

The test comment claims "lldb should be able to disassemble frames
from the inferior after asserting."  I'm not sure if there was a
specific fix this test is trying to verify.  Perhaps we can just
change it to "disassemble -s" instead?