[FFmpeg-devel] [PATCH] Remove printf format string errors in windows builds

[Måns Rullgård]

Larry Robinson <silver-dad at comcast.net> writes:

> On 1/23/2011 4:40 PM, M?ns Rullg?rd wrote:
>> Alex Converse<alex.converse at gmail.com>  writes:
>>
>>> 2011/1/23 M?ns Rullg?rd<mans at mansr.com>:
>>>> Alex Converse<alex.converse at gmail.com>  writes:
>>>>
>>>>> 2011/1/23 M?ns Rullg?rd<mans at mansr.com>:
>>>>>> Diego Elio Petten?<flameeyes at gmail.com>  writes:
>>>>>>
>>>>>>> Il giorno dom, 23/01/2011 alle 14.58 -0800, Larry Robinson ha scritto:
>>>>>>>> This patch uses #ifndef _WIN32 to replace these specifiers with `%Id'
>>>>>>>> (the windows equivalent) for windows builds.
>>>>>>> Aren't there PRI or SCN macros to use to replace those rather than
>>>>>>> doubling the lines with #if _WIN32?
>>>>>> The problem is that windows doesn't implement the C standard.
>>>>>>
>>>>> Doesn't MinGW provide it's own stdint/inttypes headers? If it doesn't
>>>>> I know there are compatibility headers. PRI/SCN constants should be
>>>>> available somehow.
>>>> There are no standard macros for those modifiers for the simple reason
>>>> that the modifiers themselves are already standard.  The PRI macros
>>>> are only for the sized intXX_t types.  These could perhaps also have
>>>> been assigned specific format letters, but maybe the alphabet ran out.
>>> Then let's make FFPRISIZE and FFPRIPTRDIFF macros?
>> First let's ask the mingw devs about supporting the standard.  I don't
>> want to make the code uglier than necessary.
>>
>> Aside from the warnings, does anything actually break?  I thought
>> people had been using FFmpeg on windows for years.
>>
> Just to be clear, I am building on Mingw-w64, which is separate from
> mingw32 and produces 64-bit executables.  It is not a problem on
> mingw32 because you can use -D__USE_MINGW_ANSI_STDIO=1 to force
> conformance (thanks to Jean-Baptiste Kempf for pointing this out), but
> this is not yet implemented for sscanf in Mingw-w64.
>
> What breaks is a number of error messages which will print
> incorrectly, but for a successful run these sections of code are not
> executed.

Given that win64 use is still limited to say the least and the
non-fatality of the errors, I say we wait at least until we know what
plans there are for implementing the proper functions in mingw-w64.
Perhaps you could even do it yourself.

-- 
M?ns Rullg?rd
mans at mansr.com