Open ghost opened 5 years ago
ringbuf_findchr
returns ringbuf_bytes_used
when there is no match, so it only returns 0
when the ring buffer is empty. (If the ring buffer is not empty and there is no match, then the value returned will be > 0
.)
It is expected that you will compare the value returned by ringbuf_findchr
to the value returned by ringbuf_bytes_used
to determine whether a match has been found, e.g.:
https://github.com/dhess/echoev/blob/c74ff579483e3aab5e10b088ef0ba63445dac0bf/echoevc.c#L337
If you compare the two values with <
, then there is no ambiguity.
The reason that ringbuf_findchr
returns ringbuf_bytes_used
when there is no match, rather than returning -1
or some other sentinel value, is so that the type of the returned value can be size_t
rather than ssize_t
; and the reason for that is for type safety, since ringbuf_findchr
itself takes size_t
offset values. (Negative offsets make no sense.)
So while this convention is slightly clumsier and a bit more expensive than checking for -1
(because this convention requires an additional call to ringbuf_bytes_used
), I think it's worth the type safety.
However, in retrospect, this is all a bit confusing, and I should probably make a note of this convention in the documentation.
hello everyone, don't you think the fuction ringbuf_findchr should return -1 if there is no match ? Ror now it return 0 if there is no match and if the match is at index 0.