[FFmpeg-devel] [PATCH v2] lavf/rtsp.c: Fix stimeout option not applied on http tunnel

[Jun Li]

Ping.

On Fri, Apr 5, 2019 at 12:00 PM Jun Li <junli1026 at gmail.com> wrote:

>
>
> On Fri, Apr 5, 2019 at 11:50 AM Jun Li <junli1026 at gmail.com> wrote:
>
>> stimeout option is already used in tcp transport, since
>> http is based on tcp, pass the option to http for tunneling
>> case.
>> ---
>>  libavformat/rtsp.c | 10 ++++++++--
>>  1 file changed, 8 insertions(+), 2 deletions(-)
>>
>> diff --git a/libavformat/rtsp.c b/libavformat/rtsp.c
>> index 033095905d..8349840c96 100644
>> --- a/libavformat/rtsp.c
>> +++ b/libavformat/rtsp.c
>> @@ -1744,6 +1744,9 @@ redirect:
>>          char httpname[1024];
>>          char sessioncookie[17];
>>          char headers[1024];
>> +        AVDictionary *options = NULL;
>> +
>> +        av_dict_set_int(&options, "timeout", rt->stimeout, 0);
>>
>>          ff_url_join(httpname, sizeof(httpname), https_tunnel ? "https" :
>> "http", auth, host, port, "%s", path);
>>          snprintf(sessioncookie, sizeof(sessioncookie), "%08x%08x",
>> @@ -1774,7 +1777,8 @@ redirect:
>>          }
>>
>>          /* complete the connection */
>> -        if (ffurl_connect(rt->rtsp_hd, NULL)) {
>> +        if (ffurl_connect(rt->rtsp_hd, &options)) {
>> +            av_dict_free(&options);
>>              err = AVERROR(EIO);
>>              goto fail;
>>          }
>> @@ -1818,10 +1822,12 @@ redirect:
>>          ff_http_init_auth_state(rt->rtsp_hd_out, rt->rtsp_hd);
>>
>>          /* complete the connection */
>> -        if (ffurl_connect(rt->rtsp_hd_out, NULL)) {
>> +        if (ffurl_connect(rt->rtsp_hd_out, &options)) {
>> +            av_dict_free(&options);
>>              err = AVERROR(EIO);
>>              goto fail;
>>          }
>> +        av_dict_free(&options);
>>      } else {
>>          int ret;
>>          /* open the tcp connection */
>> --
>> 2.17.1
>>
>>
>
> Updated the version, free the memory in error path, patch is here:
> https://patchwork.ffmpeg.org/patch/12620/
> I initially planned to do that then I realized the error path will trigger
> whole process to exit, which may be safe. But I agree with you, it is a
> better practice to clean-up here.
>
> Thanks for review, Michael.
>
> Best Regards,
> Jun
>
>
>