commit 1e54bdd48a5deec7c1c4a0a167fb5abe312c8458 Author: Nick Mathewson nickm@torproject.org Date: Fri Apr 7 09:51:08 2017 -0400
Make the warnings about terminating nuls a bit stronger
It looks like 32_encoded_size/64_encode_size APIs are inconsistent not only in the number of "d"s they have, but also in whether they count the terminating NUL. Taylor noted this in 86477f4e3fedb316, but I think we should note the inconsistently more loudly in order to avoid trouble.
(I ran into trouble with this when writing 30b13fd82e243713c6a0d.) --- src/common/util_format.c | 9 ++++++--- 1 file changed, 6 insertions(+), 3 deletions(-)
diff --git a/src/common/util_format.c b/src/common/util_format.c index 80844af..f6788e9 100644 --- a/src/common/util_format.c +++ b/src/common/util_format.c @@ -22,8 +22,10 @@ #include <stdlib.h>
/* Return the base32 encoded size in bytes using the source length srclen. - * The NUL terminated byte is added as well since every base32 encoding - * requires enough space for it. */ + * + * (WATCH OUT: This API counts the terminating NUL byte, but + * base64_encode_size does not.) + */ size_t base32_encoded_size(size_t srclen) { @@ -133,7 +135,8 @@ base32_decode(char *dest, size_t destlen, const char *src, size_t srclen) /** Return the Base64 encoded size of <b>srclen</b> bytes of data in * bytes. * - * Does <b>NOT</b> count the terminating NUL. + * (WATCH OUT: This API <em>does not</em> count the terminating NUL byte, + * but base64_encoded_size does.) * * If <b>flags</b>&BASE64_ENCODE_MULTILINE is true, return the size * of the encoded output as multiline output (64 character, `\n' terminated