[PATCH v2] manual: Update memchr description [BZ #19406]

[Wilco Dijkstra]

Hi Paul,

> This behavior guarantee applies to memchr but not to
> memcmp/wmemchr/etc., and that should probably be documented. Also, "is
> read" clashes with "stops". Something like the following instead, perhaps?
>

Thanks, that looks reasonable. v2 below.

Cheers,
Wilco


v2: Update description after review

Update memchr description to indicate it cannot overread and the
size input may be larger than the object passed.  This is required
since C11.

Suggested-by: Paul Eggert <eggert@cs.ucla.edu>

---

diff --git a/manual/string.texi b/manual/string.texi
index 24004f1f5cc60b70bafde0794351734a4eb1ce2b..b7a9ba0d46a01ea1a945b5c77cdac8825d626efb 100644
--- a/manual/string.texi
+++ b/manual/string.texi
@@ -1803,6 +1803,12 @@ to an @code{unsigned char}) in the initial @var{size} bytes of the
 object beginning at @var{block}.  The return value is a pointer to the
 located byte, or a null pointer if no match was found.
 
+This function behaves as if it reads the bytes sequentially and stops at
+the first match, which means that if the object contains a match the
+behavior is defined even when the object size is less than @var{size}.
+This behavior guarantee is unusual, and is not shared by related
+functions like @code{memcmp} and @code{wmemchr}.
+
 In ISO C23 and later, this function is qualifier-generic:
 that is, it is also implemented as a function-like macro,
 and when the macro is used and @var{block} has a type