Re: [PATCH] qom/object: Remove circular include dependency

["Philippe Mathieu-Daudé via" <qemu-devel@nongnu.org>]

On 9/5/22 11:41, Peter Maydell wrote:
> On Mon, 9 May 2022 at 09:53, Philippe Mathieu-Daudé
> <philippe.mathieu.daude@gmail.com> wrote:
>>
>> From: Philippe Mathieu-Daudé <f4bug@amsat.org>
>>
>> "qom/object.h" doesn't need to include itself.
>>
>> Signed-off-by: Philippe Mathieu-Daudé <f4bug@amsat.org>
>> ---
>>   include/qom/object.h | 1 -
>>   1 file changed, 1 deletion(-)
>>
>> diff --git a/include/qom/object.h b/include/qom/object.h
>> index 5f3d5b5bf5..ef7258a5e1 100644
>> --- a/include/qom/object.h
>> +++ b/include/qom/object.h
>> @@ -16,7 +16,6 @@
>>
>>   #include "qapi/qapi-builtin-types.h"
>>   #include "qemu/module.h"
>> -#include "qom/object.h"
>>
>>   struct TypeImpl;
>>   typedef struct TypeImpl *Type;
> 
> Accidentally (but harmlessly) 

This makes crash clang tools such clang-format, I'll amend that
to the description.

> added in commit db1015e92e0483 by a
> change generated by a script.
> 
> Reviewed-by: Peter Maydell <peter.maydell@linaro.org>
> 
> Eduardo: is it worth making ./scripts/codeconverter/converter.py
> handle this as a special case, so it doesn't add the include line
> to qom/object.h itself ? Or do we not really expect that script
> to be run on the codebase again in future ?

Oh, I see required_identifiers() adding this include in
scripts/codeconverter/codeconverter/qom_macros.py.

If we do not expect to use these scripts, we should remove it from
the repository. They are in good shape however, and might be still
valuable, so I'll amend a fix to the script.

Thanks for pointing at these scripts,

Phil.