[U-Boot] [PATCH v2 02/11] lib: charset: add u16_strncmp()
Heinrich Schuchardt
xypron.glpk at gmx.de
Wed Apr 24 17:40:16 UTC 2019
On 4/24/19 8:30 AM, AKASHI Takahiro wrote:
> u16_strncmp() works like u16_strcmp() but only at most n characters
> (in u16) are compared.
> This function will be used in a later patch.
>
> Signed-off-by: AKASHI Takahiro <takahiro.akashi at linaro.org>
> ---
> include/charset.h | 5 +++++
> lib/charset.c | 13 +++++++++++++
> 2 files changed, 18 insertions(+)
>
> diff --git a/include/charset.h b/include/charset.h
> index 747a9b376c03..49842a88bc8b 100644
> --- a/include/charset.h
> +++ b/include/charset.h
> @@ -171,6 +171,11 @@ s32 utf_to_upper(const s32 code);
> */
> int u16_strcmp(const u16 *s1, const u16 *s2);
>
> +/*
> + * u16_strncmp() - strncmp() for u16 strings
* u16_strncmp() - compare two u16 string
*
* @s1: first string to compare
* @s2: second string to compare
* @n1: maximum number of u16 to compare
* Return: 0 if the first n u16 are the same in s1 and s2
* < 0 if the first different u16 in s1 is less than the
* corresponding u16 in s2
* > 0 if the first different u16 in s1 is greater than the
* corresponding u16 in s2
> + */
> +int u16_strncmp(const u16 *s1, const u16 *s2, size_t n);
> +
> /**
> * u16_strlen - count non-zero words
> *
> diff --git a/lib/charset.c b/lib/charset.c
> index 4a25ac0bdb9c..85f08db68fe2 100644
> --- a/lib/charset.c
> +++ b/lib/charset.c
> @@ -345,6 +345,19 @@ int u16_strcmp(const u16 *s1, const u16 *s2)
> return (*(uint16_t *)s1 - *(uint16_t *)s2);
> }
>
> +int u16_strncmp(const u16 *s1, const u16 *s2, size_t n)
> +{
> + while ((n-- > 0) && (*s1 == *s2++)) {
> + if (*s1++ == 0)
> + return 0;
> + if (!n)
> + return 0;
> + }
> + --s2;
For u16_strncmp() called with n = 0 I would prefer to see 0 as return
value instead of the result of an illegal memory access.
> +
> + return (*(uint16_t *)s1 - *(uint16_t *)s2);
No need for a conversion here.
Let's avoid the superfluous increment/decrement, provide 0 for n = 0,
and make sure that the compiler calculates the difference only once per
loop:
int u16_strncmp(const u16 *s1, const u16 *s2, size_t n)
{
int ret = 0;
for (; n; --n, ++s1, ++s2) {
ret = *s1 - *s2;
if (ret || !*s1)
break;
}
return ret;
}
I would like to see a unit test in test/unicode_ut.c.
The test should include the n = 0 case.
Best regards
Heinrich
> +}
> +
> size_t u16_strlen(const u16 *in)
> {
> size_t i;
>
More information about the U-Boot
mailing list