[FFmpeg-soc] [soc]: r5719 - mms/mmst.c

[Ronald S. Bultje]

Hi,

On Thu, Mar 25, 2010 at 11:38 AM, zhentan feng <spyfeng at gmail.com> wrote:
> On Thu, Mar 25, 2010 at 11:32 PM, Ronald S. Bultje <rsbultje at gmail.com>wrote:
>> On Thu, Mar 25, 2010 at 11:30 AM, spyfeng <subversion at mplayerhq.hu> wrote:
>> > @@ -207,7 +207,6 @@ static int send_media_file_request(MMSCo
>> >     put_le32(&mms->outgoing_packet_data, 0);
>> >     put_le32(&mms->outgoing_packet_data, 0);
>> >     ff_put_str16_nolen(&mms->outgoing_packet_data, mms->path+1); // +1
>> for skip "/".
>> > -    put_le32(&mms->outgoing_packet_data, 0);
>> >
>> >     return send_command_packet(mms);
>> >  }
>> > @@ -384,7 +383,6 @@ static int send_startup_packet(MMSContex
>> >     insert_command_prefixes(mms, 0, 0x0004000b);
>> >     put_le32(&mms->outgoing_packet_data, 0x0003001c);
>> >     ff_put_str16_nolen(&mms->outgoing_packet_data, data_string);
>> > -    put_le16(&mms->outgoing_packet_data, 0); // double unicode ended
>> string.
>> >
>> >     return send_command_packet(mms);
>> >  }
>>
>> Does this still work?
>
> yes, it works.
> I check it according to the MMS specification 2.2.4.17.
> it doesn't say there are  2 bytes ended string.
> but I check libmms and find that it also have 2 ended bytes .
> Since it also works, I think it could be removed.

OK, that's fine then. Make sure that if you change the data on the
wire, that you add a comment to your commits that you tested it and it
still works with the servers you're testing against.

And yes, the change is good because it decreases the codesize + data
on the wire, that's always good!

Ronald