Re: [patch][c++/PR 31743]

[Dave Brolley <brolley at redhat dot com>]

Hi Mark,

Thanks for reviewing this.

Mark Mitchell wrote:
> Thanks for working on this.  A cleaner fix would be to but a call to
> require_complete_type in build_new (rather than build_new_1).
>
> The TYPE argument to build_new is either the type being allocated, or
> the type of the array elements.  In either case, it must be a complete
> type.  So, after the other checks (for REFERENCE_TYPE and FUNCTION_TYPE)
> in build_new, you could just do:
>
>   type = require_complete_type (type);
>   if (type == error_mark_node)
>     return error_mark_node;
>
> Then, remove the complete_type_or_else call from build_new_1.  (The only
> caller of build_new_1 is build_new.)
>   
I tried this and it yields the inaccurate diagnostic
31743.C:2: error: cannot convert ‘int*’ to ‘int (*)[]’ in initialization

The problem is that require_complete_type expects an *expression* but 
the tree we have in build_new represents a *type*. require_complete_type 
goes on to test the completeness of TREE_TYPE (value) which, in this 
case is simply 'int' and so, it doesn't go on to call 
complete_type_or_else to generate the diagnostic that we want. I suppose 
I could create a dummy expression with the given type to pass to 
require_complete_type, but that seems a bit clunky.

I still believe that my patch is correct since build_new_1 is the place 
where the use of incomplete classes/structs are currently diagnosed and 
a similar (and correct) diagnostic is generated
31743.C:2: error: invalid use of array with unspecified bounds

Please see my more detailed rationale in the PR itself and let me know 
what you think.
> Please also include a testcase with your submission.
>   
Sorry. I figured the test case in the bug report was sufficient. 
Attaching a test case to this message.

Dave

Attachment: 31743.C
Description: application/unknown-content-type-c_auto_file