Hi James!
> Am 18.08.2016 um 06:47 schrieb James Almer <jamrial at gmail.com>:
>
>> On 8/17/2016 1:14 PM, Carl Eugen Hoyos wrote:
>> Hi!
>>
>> 2016-08-17 17:32 GMT+02:00 Michael Niedermayer <michael at niedermayer.cc>:
>>>> + av_opt_get(s->pb, "mime_type", AV_OPT_SEARCH_CHILDREN, &mime_type);
>>>> + if (mime_type && s->iformat->mime_type) {
>>>> + int rate = 0, channels = 0;
>>>> + size_t len = strlen(s->iformat->mime_type);
>>>> + if (!strncmp(s->iformat->mime_type, mime_type, len)) {
>>>> + uint8_t *options = mime_type + len;
>>>> + len = strlen(mime_type);
>>>> + while (options < mime_type + len) {
>>>> + options = strstr(options, ";");
>>>> + if (!options++)
>>>> + break;
>>>> + if (!rate)
>>>> + sscanf(options, " rate=%d", &rate);
>>>> + if (!channels)
>>>> + sscanf(options, " channels=%d", &channels);
>>>> + }
>>>
>>> rate and channels probably should be subject to some sanity checks
>>> like < 0
>>
>> Sanity checks for <0 added, INT_MAX is also possible with the current
>> code.
>>
>> Patch applied, Carl Eugen
>
> This broke msvc
> http://fate.ffmpeg.org/log.cgi?time=20160817230315&log=compile&slot=x86_64-msvc12-windows-native
>
> Moving __VA_ARGS__ to the end of the macro may be enough to fix this,
> but i can't test.
I pushed this change yesterday and fate is happy again.
Thank you!
Carl Eugen