[FFmpeg-devel] [PATCH 1/2] ffmpeg: don't reconfigure terminal if we're not taking input from stdin

[Rodger Combs]

> On Sep 8, 2016, at 07:49, Nicolas George <george at nsup.org> wrote:
> 
> Le duodi 22 fructidor, an CCXXIV, Rodger Combs a écrit :
>> ---
>> ffmpeg.c | 6 +++++-
>> 1 file changed, 5 insertions(+), 1 deletion(-)
>> 
>> diff --git a/ffmpeg.c b/ffmpeg.c
>> index d858407..1d793fe 100644
>> --- a/ffmpeg.c
>> +++ b/ffmpeg.c
>> @@ -366,7 +366,7 @@ static BOOL WINAPI CtrlHandler(DWORD fdwCtrlType)
>> void term_init(void)
>> {
>> #if HAVE_TERMIOS_H
>> -    if(!run_as_daemon){
>> +    if (!run_as_daemon && stdin_interaction) {
>>         struct termios tty;
>>         if (tcgetattr (0, &tty) == 0) {
>>             oldtty = tty;
>> @@ -4328,6 +4328,10 @@ int main(int argc, char **argv)
>> 
>>     show_banner(argc, argv, options);
>> 
> 
>> +    ret = locate_option(argc, argv, options, "stdin");
>> +    if (ret && !strcmp(argv[ret], "-nostdin"))
>> +        stdin_interaction = 0;
>> +
>>     term_init();
> 
> I think it would be more elegant to move the term_init() call a few lines
> down, after ffmpeg_parse_options(), rather than parse -nostdin out of order.

Agreed in principle; I did it this way because I'm not entirely sure if there would be negative consequences to parsing options (and thus opening I/O) before setting up signal handlers. If you think that's not an issue, I'll make the change you suggest.

> 
> Apart from that, both patches looks right to me.
> 
>> 
>>     /* parse options and open all input/output files */
> 
> Regards,
> 
> -- 
>  Nicolas George
> _______________________________________________
> ffmpeg-devel mailing list
> ffmpeg-devel at ffmpeg.org <mailto:ffmpeg-devel at ffmpeg.org>
> http://ffmpeg.org/mailman/listinfo/ffmpeg-devel <http://ffmpeg.org/mailman/listinfo/ffmpeg-devel>