Use number instead of "min_len", which is easier to review (with help of
the documented parameter sizes).
valuelen is redundant as *vlen can be used directly.
Both userspace and kernel enforce a minimum ATT MTU of 23 octets, which
is also used as minimum size for buffers passed to ATT encoding
functions. Therefore, it is unnecessary to perform these checks on ATT
requests that are guaranteed to fit into 23 octets.
Also document ATT parameter lengths where a constant is being used for
calculating the PDU length.
Use standard C types instead of GLib ones (which are unnecessary here)
and move the "status" parameter to the first position, so it is
consistent with other callbacks.
The __func__ macro is part of the C99 standard whereas __FUNCTION__ is a
legacy gcc specific alias for it:
http://gcc.gnu.org/onlinedocs/gcc/Function-Names.html
Additionally, checkpatch.pl that's commonly used to verify coding style
also recommends to use __func__ instead of __FUNCTION__.
Custom GATT services typically use 128-bits UUIDs; this patch allows for
using gatt_service_add to define such services by adding a new option
GATT_OPT_CHR_UUID, which allows for passing bt_uuid_t* values. Note, a
previous patch renamed the old GATT_OPT_CHR_UUID into
GATT_OPT_CHR_UUID16.
To paraphrase Anderson Lizardo: "h" was overflowing for 128-bit services
because the handle is incremented after each attribute is added so at
the end we had: 0xffff + 1 == 0 (h is uint16_t).
As it seems more complex to try to remove the overflow without affecting
the logic too much, we update the assertion.
Previously, if no callback was given to gatt_write_char(), it was
assumed that a "Write Without Response" (which uses Write Command)
should be used instead of Write Request.
This "shortcut" is unnecessary (there is gatt_write_cmd() for the
situations where Write Without Response is required) and just duplicates
code.
This commit also fixes the few places where gatt_write_cmd() should be
used.
The old name referred to the underlying ATT PDU (Find Information
Request and Find Information Response) and not to the actual GATT
procedure (Discover All Characteristic Descriptors).
Passing the special character single quote:
$gatttool -I
[ ][ ][LE]> '
Segmentation fault (core dumped)
Checking g_shell_parse_argv fail status (return FALSE on failure), works
fine.
Now to exit from gatttool at interactive mode, user must use SIGTERM(15)
or press Ctrl+D. SIGINT(2) or Ctrl+C just flushes typed characters and
breaks line.
When trying to connect, the "connecting" message is not set as prompt
anymore. Instead, it is printed just like other status messages and the
prompt is kept, allowing the user to send commands while connection is
being established.
Use rl_printf() from bluetoothctl to consistently display messages. This
is specially important for asynchronous callbacks, where using printf()
may mess up the prompt created by readline.
g_atomic_* end up using G_STATIC_ASSERT, causing gcc 4.8 to yell due to
-Wunused-local-typedefs.
/usr/include/glib-2.0/glib/gmacros.h:162:53: error: typedef ‘_GStaticAssertCompileTimeAssertion_2’ locally defined but not used [-Werror=unused-local-typedefs]
#define G_STATIC_ASSERT(expr) typedef char G_PASTE (_GStaticAssertCompileTimeAssertion_, __COUNTER__)[(expr) ? 1 : -1]
Most of the uses of atomic operations were wrong. They were fixed as
well. If we are using atomic operations, reading the variable again
later for logging is not an option, we should use the return of the
atomic function used to fetch the variable.
This check is more appropriate to be done by callers of gatt_connect().
This is already done on interactive mode in attrib/interactive.c and is
now also done in non-interactive mode.
