Re: test case against union aliasing

[Julia Lawall]

On Tue, 27 Feb 2018, Andy Shevchenko wrote:

> Hi!
>
> It seems we have some code in kernel which utilizes (rightlessly for
> now) union aliasing.
>
> So, can we create a test case against such? (Three steps: a) replace
> union by struct keyword in data type in question, b) compile, c) run in
> Qemu with proper functional test)
>
> Does it make any sense?

I think so.  If you want to do it more safely than with sed, it should be
possible with Coccinelle, with just

@@
identifier i;
@@

-union i
+struct i

julia