due to type promotions in C this code was failing on some platforms.
src/device.c:2016:8: error: format specifies type 'unsigned short' but the
argument has type 'uint8_t' (aka 'unsigned char') [-Werror,-Wformat]
properties, uuid_str);
^~~~~~~~~~
/build/falco/usr/include/bits/stdio2.h:39:7: note: expanded from macro 'sprintf'
__VA_ARGS__)
^
---
src/device.c | 2 +-
1 file changed, 1 insertion(+), 1 deletion(-)
diff --git a/src/device.c b/src/device.c
index 0d3a655..b4e6d53 100644
--- a/src/device.c
+++ b/src/device.c
@@ -2015,7 +2015,7 @@ static void store_chrc(struct gatt_db_attribute *attr, void *user_data)
sprintf(handle, "%04hx", handle_num);
bt_uuid_to_string(&uuid, uuid_str, sizeof(uuid_str));
- sprintf(value, GATT_CHARAC_UUID_STR ":%04hx:%02hx:%s", value_handle,
+ sprintf(value, GATT_CHARAC_UUID_STR ":%04hx:%02x:%s", value_handle,
properties, uuid_str);
g_key_file_set_string(key_file, "Attributes", handle, value);
--
2.5.0
Hi Johan,
On Mon, Oct 19, 2015 at 11:51 PM, Johan Hedberg <[email protected]> wrote:
> Hi Jakub,
>
> On Mon, Oct 19, 2015, Jakub Pawlowski wrote:
>> On Mon, Oct 19, 2015 at 10:35 PM, Johan Hedberg <[email protected]> wrote:
>> > Hi Jakub,
>> >
>> > On Mon, Oct 19, 2015, Jakub Pawlowski wrote:
>> >> due to type promotions in C this code was failing on some platforms.
>> >>
>> >> src/device.c:2016:8: error: format specifies type 'unsigned short' but the
>> >> argument has type 'uint8_t' (aka 'unsigned char') [-Werror,-Wformat]
>> >> properties, uuid_str);
>> >> ^~~~~~~~~~
>> >> /build/falco/usr/include/bits/stdio2.h:39:7: note: expanded from macro 'sprintf'
>> >> __VA_ARGS__)
>> >> ^
>> >> ---
>> >> src/device.c | 2 +-
>> >> 1 file changed, 1 insertion(+), 1 deletion(-)
>> >>
>> >> diff --git a/src/device.c b/src/device.c
>> >> index 0d3a655..b4e6d53 100644
>> >> --- a/src/device.c
>> >> +++ b/src/device.c
>> >> @@ -2015,7 +2015,7 @@ static void store_chrc(struct gatt_db_attribute *attr, void *user_data)
>> >>
>> >> sprintf(handle, "%04hx", handle_num);
>> >> bt_uuid_to_string(&uuid, uuid_str, sizeof(uuid_str));
>> >> - sprintf(value, GATT_CHARAC_UUID_STR ":%04hx:%02hx:%s", value_handle,
>> >> + sprintf(value, GATT_CHARAC_UUID_STR ":%04hx:%02x:%s", value_handle,
>> >> properties, uuid_str);
>> >
>> > Wouldn't the right format specifier be %hhx in this case?
>>
>> So it was compiling fine on my machine, but when it was compiling for
>> all possible platforms for ChromeOS those errors came up.
>> The very proper thing would be probably to use "PRIu8" from <inttypes.h>
>> but then I found good explanation:
>>
>> ""printf() is a variadic function. Its optional arguments( and only
>> those ) get promoted according to default argument promotions(
>> 6.5.2.2. p6 ).
>>
>> Since you are asking for integers, integer promotions are applied in
>> this case, and types you mention get promoted to int. ( and not
>> unsigned int because C )""
>> source: http://stackoverflow.com/questions/26362386/why-is-the-format-specifier-for-uint8-t-and-uint16-t-the-same-u
>>
>> so I understand that this uint8_t will get promoted to int, and
>> displayed properly.
>> I manually tested it for 0, 127, and 255 and looks like it's
>> formatting just fine.
>
> It's a bit unclear from your answer: do you still get warnings/errors
> with %hhx or not? Looking at the warning you got:
>
> "argument has type 'uint8_t' (aka 'unsigned char')"
>
> and the printf(3):
>
> "hh A following integer conversion corresponds to a signed char or
> unsigned char argument"
>
> It'd still seem like %hhx shouldn't give a warning. Even though %x might
> work too being able to be more specific in the type specification
> doesn't seem like a bad thing.
Ok, I've build version with %hhx on all platforms we have and it
worked well with no warnings. I'll update patch.
>
> Johan
Hi Jakub,
On Mon, Oct 19, 2015, Jakub Pawlowski wrote:
> On Mon, Oct 19, 2015 at 10:35 PM, Johan Hedberg <[email protected]> wrote:
> > Hi Jakub,
> >
> > On Mon, Oct 19, 2015, Jakub Pawlowski wrote:
> >> due to type promotions in C this code was failing on some platforms.
> >>
> >> src/device.c:2016:8: error: format specifies type 'unsigned short' but the
> >> argument has type 'uint8_t' (aka 'unsigned char') [-Werror,-Wformat]
> >> properties, uuid_str);
> >> ^~~~~~~~~~
> >> /build/falco/usr/include/bits/stdio2.h:39:7: note: expanded from macro 'sprintf'
> >> __VA_ARGS__)
> >> ^
> >> ---
> >> src/device.c | 2 +-
> >> 1 file changed, 1 insertion(+), 1 deletion(-)
> >>
> >> diff --git a/src/device.c b/src/device.c
> >> index 0d3a655..b4e6d53 100644
> >> --- a/src/device.c
> >> +++ b/src/device.c
> >> @@ -2015,7 +2015,7 @@ static void store_chrc(struct gatt_db_attribute *attr, void *user_data)
> >>
> >> sprintf(handle, "%04hx", handle_num);
> >> bt_uuid_to_string(&uuid, uuid_str, sizeof(uuid_str));
> >> - sprintf(value, GATT_CHARAC_UUID_STR ":%04hx:%02hx:%s", value_handle,
> >> + sprintf(value, GATT_CHARAC_UUID_STR ":%04hx:%02x:%s", value_handle,
> >> properties, uuid_str);
> >
> > Wouldn't the right format specifier be %hhx in this case?
>
> So it was compiling fine on my machine, but when it was compiling for
> all possible platforms for ChromeOS those errors came up.
> The very proper thing would be probably to use "PRIu8" from <inttypes.h>
> but then I found good explanation:
>
> ""printf() is a variadic function. Its optional arguments( and only
> those ) get promoted according to default argument promotions(
> 6.5.2.2. p6 ).
>
> Since you are asking for integers, integer promotions are applied in
> this case, and types you mention get promoted to int. ( and not
> unsigned int because C )""
> source: http://stackoverflow.com/questions/26362386/why-is-the-format-specifier-for-uint8-t-and-uint16-t-the-same-u
>
> so I understand that this uint8_t will get promoted to int, and
> displayed properly.
> I manually tested it for 0, 127, and 255 and looks like it's
> formatting just fine.
It's a bit unclear from your answer: do you still get warnings/errors
with %hhx or not? Looking at the warning you got:
"argument has type 'uint8_t' (aka 'unsigned char')"
and the printf(3):
"hh A following integer conversion corresponds to a signed char or
unsigned char argument"
It'd still seem like %hhx shouldn't give a warning. Even though %x might
work too being able to be more specific in the type specification
doesn't seem like a bad thing.
Johan
Hi Johan,
On Mon, Oct 19, 2015 at 10:35 PM, Johan Hedberg <[email protected]> wrote:
> Hi Jakub,
>
> On Mon, Oct 19, 2015, Jakub Pawlowski wrote:
>> due to type promotions in C this code was failing on some platforms.
>>
>> src/device.c:2016:8: error: format specifies type 'unsigned short' but the
>> argument has type 'uint8_t' (aka 'unsigned char') [-Werror,-Wformat]
>> properties, uuid_str);
>> ^~~~~~~~~~
>> /build/falco/usr/include/bits/stdio2.h:39:7: note: expanded from macro 'sprintf'
>> __VA_ARGS__)
>> ^
>> ---
>> src/device.c | 2 +-
>> 1 file changed, 1 insertion(+), 1 deletion(-)
>>
>> diff --git a/src/device.c b/src/device.c
>> index 0d3a655..b4e6d53 100644
>> --- a/src/device.c
>> +++ b/src/device.c
>> @@ -2015,7 +2015,7 @@ static void store_chrc(struct gatt_db_attribute *attr, void *user_data)
>>
>> sprintf(handle, "%04hx", handle_num);
>> bt_uuid_to_string(&uuid, uuid_str, sizeof(uuid_str));
>> - sprintf(value, GATT_CHARAC_UUID_STR ":%04hx:%02hx:%s", value_handle,
>> + sprintf(value, GATT_CHARAC_UUID_STR ":%04hx:%02x:%s", value_handle,
>> properties, uuid_str);
>
> Wouldn't the right format specifier be %hhx in this case?
So it was compiling fine on my machine, but when it was compiling for
all possible platforms for ChromeOS those errors came up.
The very proper thing would be probably to use "PRIu8" from <inttypes.h>
but then I found good explanation:
""printf() is a variadic function. Its optional arguments( and only
those ) get promoted according to default argument promotions(
6.5.2.2. p6 ).
Since you are asking for integers, integer promotions are applied in
this case, and types you mention get promoted to int. ( and not
unsigned int because C )""
source: http://stackoverflow.com/questions/26362386/why-is-the-format-specifier-for-uint8-t-and-uint16-t-the-same-u
so I understand that this uint8_t will get promoted to int, and
displayed properly.
I manually tested it for 0, 127, and 255 and looks like it's
formatting just fine.
Jakub
>
> Johan
Hi Jakub,
On Mon, Oct 19, 2015, Jakub Pawlowski wrote:
> due to type promotions in C this code was failing on some platforms.
>
> src/device.c:2016:8: error: format specifies type 'unsigned short' but the
> argument has type 'uint8_t' (aka 'unsigned char') [-Werror,-Wformat]
> properties, uuid_str);
> ^~~~~~~~~~
> /build/falco/usr/include/bits/stdio2.h:39:7: note: expanded from macro 'sprintf'
> __VA_ARGS__)
> ^
> ---
> src/device.c | 2 +-
> 1 file changed, 1 insertion(+), 1 deletion(-)
>
> diff --git a/src/device.c b/src/device.c
> index 0d3a655..b4e6d53 100644
> --- a/src/device.c
> +++ b/src/device.c
> @@ -2015,7 +2015,7 @@ static void store_chrc(struct gatt_db_attribute *attr, void *user_data)
>
> sprintf(handle, "%04hx", handle_num);
> bt_uuid_to_string(&uuid, uuid_str, sizeof(uuid_str));
> - sprintf(value, GATT_CHARAC_UUID_STR ":%04hx:%02hx:%s", value_handle,
> + sprintf(value, GATT_CHARAC_UUID_STR ":%04hx:%02x:%s", value_handle,
> properties, uuid_str);
Wouldn't the right format specifier be %hhx in this case?
Johan