Re: Uninitialized out parameters.

[ahlan@marriott.org]

On Wednesday, April 6, 2016 at 1:37:41 PM UTC+2, AdaMagica wrote:
> Hm, from the point of view of the RM, the RM has nothing to say about warnings - only whether the code is legal or illegal. So giving warnings is up to the compiler.
> 
> Now for scalar variables, the RM says, in mode variables are copied in, out mode variables are copied out, in out variables are both. So as I understand, even a null procedure with an out parameter does something to the parameter: it copies out an uninitialized variable (which can have any value whatsoever). Language lawyers - please?

In the annotated RM, in the section concerning Null Procedures, it says "There are no null functions because the return value has to be constructed somehow; a function that always raises Program_Error doesn't seem very useful or worth the complication"
If there are no null functions because a return value has to be constructed then surely the same applies to null procedures that have out parameters. These too should be prohibited because they too require that a value has to be constructed. If not prohibited then at least Program_Error should be raised and ideally a warning is generated by the compiler warning that calling the null procedure will always raise Program_Error.
Currently GNAT simply returns "something" - zero or null that sometimes escapes detection. This is NOT desirable.