Hi Nicolas! On 2017-03-30 22:12 +0200, Nicolas George wrote: > Le decadi 10 germinal, an CCXXV, Alexander Strasser a écrit : > > If expiry is zero terminated and you are going through it one byte at a > > time, > > you could omit the strlen call and just check if expiry[i] is non zero. > > > > It's maybe the more common idiom too. > > On the other hand, code that does not need a 0-terminated string is more > generic.
This code depends already on a 0-terminated string. Otherwise the call of strlen before would be wrong. Also the later loop is also written against a 0-terminated string. IMHO for this patch it's not really important to avoid relying on 0-terminated string. > I am thinking of the ideas I posted a few months ago about reworking the > options system to avoid escaping hell. Parsers for various types must be > able to work in the middle of strings with foreign delimiters. Code that > is already capable of working in the middle of a string would be easier > to integrate. > > Of course, all this is in the far future. I will not insist on all new > code to be able to work like that, but I would like it nonetheless. And > if it is already written that way, let us keep it. I am not in general against getting rid of 0-terminated string in some places. I often thought of this too. > All in all, 0-terminated strings were a terrible terrible idea. Yes, they probably still are a bad idea. Alexander _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel