Peter Maydell wrote:
> On 4 October 2012 12:09, Jim Meyering <j...@meyering.net> wrote:
>> From: Jim Meyering <meyer...@redhat.com>
>>
>> Don't bother with strncpy. There's no need for its zero-fill.
>> Use g_strndup in place of g_malloc+strncpy+NUL-terminate.
>>
>> Signed-off-by: Jim Meyering <meyer...@redhat.com>
>> ---
>> ui/vnc-auth-sasl.c | 4 +---
>> 1 file changed, 1 insertion(+), 3 deletions(-)
>>
>> diff --git a/ui/vnc-auth-sasl.c b/ui/vnc-auth-sasl.c
>> index 8fba770..bfdcb46 100644
>> --- a/ui/vnc-auth-sasl.c
>> +++ b/ui/vnc-auth-sasl.c
>> @@ -432,9 +432,7 @@ static int protocol_client_auth_sasl_start_len(VncState
>> *vs, uint8_t *data, size
>>
>> static int protocol_client_auth_sasl_mechname(VncState *vs, uint8_t *data,
>> size_t len)
>> {
>> - char *mechname = g_malloc(len + 1);
>> - strncpy(mechname, (char*)data, len);
>> - mechname[len] = '\0';
>> + char *mechname = g_strndup((const char *) data, len);
>> VNC_DEBUG("Got client mechname '%s' check against '%s'\n",
>> mechname, vs->sasl.mechlist);
>>
>
> (Does the compiler really insist on that cast?)
Yes. Without it, I get a warning/failure when char is signed.
See below.
> Reviewed-by: Peter Maydell <peter.mayd...@linaro.org>
Thanks for the review.
ui/vnc-auth-sasl.c: In function 'protocol_client_auth_sasl_mechname':
ui/vnc-auth-sasl.c:435:5: error: pointer targets in passing argument 1 of
'g_strndup' differ in signedness [-Werror=pointer-sign]
char *mechname = g_strndup(data, len);
^
In file included from /usr/include/glib-2.0/glib.h:81:0,
from ./qemu-common.h:41,
from ui/vnc.h:30,
from ui/vnc-auth-sasl.c:25:
/usr/include/glib-2.0/glib/gstrfuncs.h:192:8: note: expected 'const gchar
*' but argument is of type 'uint8_t *'
gchar* g_strndup (const gchar *str,
^
cc1: all warnings being treated as errors
make: *** [ui/vnc-auth-sasl.o] Error 1
