qemu-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [Qemu-devel] [PULL 2/8] ui: convert the SDL2 frontend to keycodemapd


From: Paolo Bonzini
Subject: Re: [Qemu-devel] [PULL 2/8] ui: convert the SDL2 frontend to keycodemapdb
Date: Thu, 1 Feb 2018 13:10:34 -0500
User-agent: Mozilla/5.0 (X11; Linux x86_64; rv:52.0) Gecko/20100101 Thunderbird/52.5.2

On 01/02/2018 12:54, Daniel P. Berrangé wrote:
> On Thu, Feb 01, 2018 at 12:37:26PM -0500, Paolo Bonzini wrote:
>> On 01/02/2018 12:35, Daniel P. Berrangé wrote:
>>> On Thu, Feb 01, 2018 at 12:28:02PM -0500, Paolo Bonzini wrote:
>>>> On 25/01/2018 09:32, Gerd Hoffmann wrote:
>>>>> --- a/include/ui/input.h
>>>>> +++ b/include/ui/input.h
>>>>> @@ -80,4 +80,7 @@ extern const guint16 qemu_input_map_qnum_to_qcode[];
>>>>>  extern const guint qemu_input_map_qcode_to_linux_len;
>>>>>  extern const guint16 qemu_input_map_qcode_to_linux[];
>>>>
>>>> Is there any reason to use the length variables rather than just
>>>> ARRAY_SIZE()?  Coverity isn't able to see through the initializer and
>>>> incorrectly complains that
>>>
>>> ARRAY_SIZE can only be used if you can see the array variable
>>> initialization, but this is the header file so all we have is
>>> the extern declaration and thus the array size is unknown from
>>> the places that need to use the _len variable.
>>>
>>> We do in fact initalize the _len variable based on the ARRAY_SIZE,
>>> but that's in the .c files.
>>
>> Would it make sense for keymap-gen to also generate the header file, so
>> that we can use an "extern const guint x[NNN]" declaration rather
>> instead of "x[]"?
> 
> I could probably add such a feature to it....

I could too if I know it will be accepted!  I'll send you a patch.

Thanks,

Paolo



reply via email to

[Prev in Thread] Current Thread [Next in Thread]