[FFmpeg-devel] [PATCH] avio: add avio_get_str as a replacement for get_strz

Ronald S. Bultje rsbultje
Thu Mar 3 16:57:57 CET 2011

Hi Reimar,

On Wed, Mar 2, 2011 at 1:21 PM, Reimar D?ffinger
<Reimar.Doeffinger at gmx.de> wrote:
> On Wed, Mar 02, 2011 at 12:12:58PM -0500, Ronald S. Bultje wrote:
>> On Tue, Mar 1, 2011 at 3:38 PM, Reimar D?ffinger
>> <Reimar.Doeffinger at gmx.de> wrote:
>> > On Tue, Mar 01, 2011 at 06:33:56PM +0100, Anton Khirnov wrote:
>> >> ?/**
>> >> + * Read a UTF-8 string from pb. The reading will terminate when either
>> >> + * a NULL character was encountered or maxlen bytes have been read.
>> >> + *
>> >> + * @return number of bytes read (is always <= maxlen).
>> >> + */
>> >> +int avio_get_str(AVIOContext *pb, int maxlen, char *buf, int buflen);
>> >
>> > 1) If you already say UTF-8 string you should definitely use uint8_t as type.
>> Why? libc uses char* for UTF8.
> Because it needlessly risks issues due to undefined signedness.
> And which libc functions are explicitly UTF-8 functions? I am not aware of any.
>> Looks a little nicer indeed. Can anyone (maybe Anton) send a
>> git-formatted patch that does something like this? Also, Reimar, if
>> you don't mind, I can commit this with Anton as author and then your
>> patch (hopefully you'll send one git-formatted) right on top, so that
>> authorship is properly attributed.
> I very much mind wasting my time with that kind of crap, particularly
> since the reason for writing the code here is that I don't want
> to implement and test it.
> I'd _very_ much rather have someone else as author than this.
> Not to mention that I consider it rather questionable to commit something
> suboptimal just for this purpose, that is cluttering the history for politics
> (IMHO, sorry).

Others appear to care more about it, sorry for this. Anton already
corrected order-of-patches so this isn't an issue anymore.


More information about the ffmpeg-devel mailing list