From mboxrd@z Thu Jan 1 00:00:00 1970 From: Heinrich Schuchardt Date: Wed, 24 Apr 2019 19:40:16 +0200 Subject: [U-Boot] [PATCH v2 02/11] lib: charset: add u16_strncmp() In-Reply-To: <20190424063045.14443-3-takahiro.akashi@linaro.org> References: <20190424063045.14443-1-takahiro.akashi@linaro.org> <20190424063045.14443-3-takahiro.akashi@linaro.org> Message-ID: <681cca2e-d2a9-d3c4-c949-203d4e55c464@gmx.de> List-Id: MIME-Version: 1.0 Content-Type: text/plain; charset="utf-8" Content-Transfer-Encoding: 8bit To: u-boot@lists.denx.de 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 > --- > 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; >