Martin Jackson [Sat, 9 Sep 2017 22:37:48 +0000 (00:37 +0200)]
windows: wc_to_utf8(): Fix a WCHAR related issue causing crashes.
In wc_to_utf8() in windows.c, the zero terminator is written to an invalid
array index, which results in 2 bytes being zeroed in a random place in the
stack. This sometimes causes a crash when running sp_list_ports() (depending
on string length and compiler optimisation settings).
sizeof(wc_str) returns the size in bytes, so cannot be used directly as an
index into that array, it should be divided by sizeof(WCHAR). Otherwise the
zero terminator index is approximately twice what it should be.
Aurelien Jacobs [Fri, 14 Oct 2016 21:50:08 +0000 (23:50 +0200)]
use readdir() instead of the deprecated readir_r()
readir() is threadsafe on both linux and freebsd anyway.
The rationale behind the readdir_r() deprecation is in the glibc manual:
https://www.gnu.org/software/libc/manual/html_node/Reading_002fClosing-Directory.html
This fixes the following warning with recent glibc:
linux.c: In function ‘list_ports’:
linux.c:197:2: warning: ‘readdir_r’ is deprecated [-Wdeprecated-declarations]
while (!readdir_r(dir, &entry, &result) && result) {
^~~~~
Uwe Hermann [Sat, 9 Jan 2016 00:13:45 +0000 (01:13 +0100)]
Bump libtool version (not package version) to 1:0:1.
The last release (0.1.0) had the libtool version (current:revision:age)
set to 0:0:0. Since this release doesn't change/remove any
interfaces (it does *add* new interfaces, though), 'current' and 'age'
are increased and 'revision' is set to 0, resulting in 1:0:1.
Martino Facchin [Wed, 16 Dec 2015 18:07:49 +0000 (19:07 +0100)]
windows: Fix iSerial for composite devices.
USB composite devices can contain an ACM serial interface.
On Windows, the correct iSerial descriptor field is assigned to the parent
(composite) device instead of to the actual serial interface. A bogus
value is returned if the serial interface is asked to provide the S/N.
This patch provides a fallback for this kind of device (tested on Android
with adb + cdc gadgets and on Arduino Zero Programming Port)
Martino Facchin [Fri, 18 Dec 2015 11:40:58 +0000 (12:40 +0100)]
macosx: Fix port listing on ElCapitan.
On OSX 10.11 (ElCapitan) the query for IOProviderClass fails to list ACM
devices as USB.
Add a fallback query using IOClass to correctly recognize these devices.
The fix has no effect on previous OSX versions (tested on Mavericks)
Daniel Elstner [Wed, 26 Aug 2015 17:46:43 +0000 (19:46 +0200)]
Build: Reduce autogen.sh to trivial stub
Use autoreconf instead of invoking the various Auto tools
separately. Get rid of the Darwin-specific guesswork -- it does
not make sense to handle this at the level of libserialport.
People should set up their ACLOCAL_PATH themselves as appropriate
for their own system; just as they already need to set up various
other paths.
Uwe Hermann [Sat, 11 Apr 2015 19:03:53 +0000 (21:03 +0200)]
Fix a compiler warning when -Wshadow is used.
CC serialport.lo
In file included from ../serialport.c:25:0:
../serialport.c: In function 'get_config':
../libserialport_internal.h:227:25: warning: declaration of 'ret' shadows a previous local [-Wshadow]
#define TRY(x) do { int ret = x; if (ret != SP_OK) RETURN_CODEVAL(ret); } while (0)
^
../serialport.c:1566:3: note: in expansion of macro 'TRY'
TRY(get_baudrate(port->fd, &config->baudrate));
^
../serialport.c:1543:6: warning: shadowed declaration is here [-Wshadow]
int ret = get_flow(port->fd, data);
^
Also, add -Wshadow to the list of default compiler options.
Uwe Hermann [Mon, 6 Apr 2015 18:25:42 +0000 (20:25 +0200)]
README: Refer to MSYS2 only.
The "old" MSYS related to the "old" and unsupported MinGW (from
mingw.org) may or may not work properly, so only refer to MSYS2 which
we currently recommend and which has been tested.
Uwe Hermann [Sat, 4 Apr 2015 16:10:18 +0000 (18:10 +0200)]
configure.ac: Use $host_os instead of $target_os.
$build and $build_os refer to the machine on which the software is
built, $host and $host_os refer to the machine for which the software
is (cross-)built.
The variables $target and $target_os (despite the confusing names)
are only relevant in the context of building cross-compilers.
Martin Ling [Tue, 31 Mar 2015 23:24:27 +0000 (00:24 +0100)]
Clarify sp_list_ports() code.
This also fixes the following scan-build warning:
serialport.c:335:15: warning: Result of 'malloc' is converted to a
pointer of type 'struct sp_port *', which is incompatible with sizeof
operand type 'struct sp_port **'
if (!(list = malloc(sizeof(struct sp_port **))))
^~~~~~ ~~~~~~~~~~~~~~~~~~~~~~~~~
Martin Ling [Fri, 27 Mar 2015 22:43:14 +0000 (22:43 +0000)]
windows: Set fAbortOnError=FALSE.
Since we're not handling and clearing receive errors (framing, parity and
overrun), we should turn them off to avoid crashing out with
ERROR_OPERATION_ABORTED in various calls if they occur. Invalid data
will then simply not end up in the buffer. This is consistent with our
current behaviour on posix/termios systems.
It might be nice to be able to inform calling code about receive errors, but I
can't see a good way to do this in a cross-platform way at the moment.