Fixed #7: Documented state parameter as non-empty

Note that I would not really consider this as a backwards
compatibility problem, as there is a documented way of
returning error codes, and an error code exists for this
situation. One would expect any reasonable implementation
to check for non-null as a sanity check, as it so cheap.

The exception would be if it was documented as a performance
critical method (which it is, but we don't care about
performance), where you would probably either put a blanket
warning "No pre-conditions will be checked" for the module,
or ideally on just the performance critical functions.

1 files changed, 3 insertions, 1 deletions

diff --git a/include/mips_cpu.h b/include/mips_cpu.h
index 8af9827..5b66ab9 100644
--- a/include/mips_cpu.h
+++ b/include/mips_cpu.h
@@ -104,7 +104,9 @@ mips_error mips_cpu_get_pc(
 	difficult, so _try_ to maintain it, but don't worry too
 	much if under some exceptions it doesn't quite work.
 */
-mips_error mips_cpu_step(mips_cpu_h state);
+mips_error mips_cpu_step(
+	mips_cpu_h state	//! Valid (non-empty) handle to a CPU
+);
 
 /*! Controls printing of diagnostic and debug messages.