On Tuesday, 25 March 2025, 09:35:47 CET, Yao Zi wrote:
> A character literl, '\0', is assigned to the pointer. The corresponding
> comment doesn't make any sense, since '\0' represents zero and this
> statement does the same thing as assigning value with NULL.
>
> Initializing value with NULL (or zero) is safe and correct here: the
> only case that the initial value of the pointer is passed to
> write_value() is that the if branch isn't executed, where len keeps its
> initial value, zero, as well. With src_len equal to zero, write_value()
> will bail out and src_val won't be dereferenced.
>
> Let's clean up the misleading comment and change right side of the
> assignment to fix compiler warnings about the wrong type,
>
> Fixes: ee750bbaf68c ("client/gatt: proxy_property_changed: check for NULL iterator")
> Signed-off-by: Yao Zi <ziyao@xxxxxxxxxxx>
After reading my own commit message, assigning "value = '\0'" absolutely makes
no sense in this context. Actually my intention was only checking for iter != NULL.
Acked-by: Christian Eggers <ceggers@xxxxxxx>
> ---
> client/gatt.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/client/gatt.c b/client/gatt.c
> index 4dac88590..b18186518 100644
> --- a/client/gatt.c
> +++ b/client/gatt.c
> @@ -3207,7 +3207,7 @@ static void proxy_property_changed(GDBusProxy *proxy, const char *name,
> chrc->path, bt_uuidstr_to_str(chrc->uuid), name);
>
> if (!strcmp(name, "Value")) {
> - uint8_t *value = '\0'; /* don't pass NULL to write_value() */
> + uint8_t *value = NULL;
> int len = 0;
>
> if (iter && dbus_message_iter_get_arg_type(iter) ==
>
