[PATCH 0/2] Fix Clang warnings

[Hans de Goede]

Hi,

Patches look good, ACK series.

Regards,

Hans


On 03/28/2012 01:06 PM, Pete Batard wrote:
> Just a couple of patches for warnings raised by Clang/MinGW. These only affect Windows/xusb.
>
> With these applied, below is the output from Clang in MinGW:
>
> ------------------------------------------------------------------------
> $ scan-build make -j2
> make all-recursive
> make[1]: Entering directory `/d/libusbx'
> Making all in libusb
> make[2]: Entering directory `/d/libusbx/libusb'
> CC libusb_1_0_la-core.lo
> CC libusb_1_0_la-descriptor.lo
> CC libusb_1_0_la-io.lo
> io.c:1864:8: warning: Call to 'malloc' has an allocation size of 0 bytes
> fds = malloc(sizeof(*fds) * nfds);
> ^ ~~~~~~~~~~~~~~~~~~~
> 1 warning generated.
> CC libusb_1_0_la-sync.lo
> CC libusb_1_0_la-poll_windows.lo
> CC libusb_1_0_la-windows_usb.lo
> GEN libusb-1.0.lo
> CC libusb_1_0_la-threads_windows.lo
> CCLD libusb-1.0.la
> Creating library file: .libs/libusb-1.0.dll.a
> make[2]: Leaving directory `/d/libusbx/libusb'
> Making all in doc
> make[2]: Entering directory `/d/libusbx/doc'
> make[2]: Nothing to be done for `all'.
> make[2]: Leaving directory `/d/libusbx/doc'
> Making all in examples
> make[2]: Entering directory `/d/libusbx/examples'
> CC listdevs.o
> CC xusb.o
> CCLD listdevs.exe
> ./.libs/lt-listdevs.c:692:11: warning: Value stored to 'len' during its initialization is never read
> int len = strlen (new_value);
> ^ ~~~~~~~~~~~~~~~~~~
> CCLD xusb.exe
> 1 warning generated.
> ./.libs/lt-xusb.c:692:11: warning: Value stored to 'len' during its initialization is never read
> int len = strlen (new_value);
> ^ ~~~~~~~~~~~~~~~~~~
> 1 warning generated.
> make[2]: Leaving directory `/d/libusbx/examples'
> make[2]: Entering directory `/d/libusbx'
> make[2]: Leaving directory `/d/libusbx'
> make[1]: Leaving directory `/d/libusbx'
> scan-build: 3 bugs found.
> scan-build: Run 'scan-view /tmp/scan-build-2012-03-28-2' to examine bug reports.
> ------------------------------------------------------------------------
>
> The lt- ones are libtool issues which only libtool can (and hopefully will) address as we're not invoking strlen in our code.
>
> As far as I can tell, the one in io.c is a false positive, as nfds should never be zero. Maybe there's a way to silence Clang on that, but I don't want to touch core for 1.0.9 unless needed. Now, if we can ever get a zero value there, we may have an issue, as the behaviour of malloc(0) is undetermined, and it seems gcc will return a pointer that should not be dereferenced [1].
>
> Regards,
>
> /Pete
>
> PS: I also pushed a fix to address the missing xusb in the VS2005 project files
>
> [1] http://bytes.com/topic/c/answers/578467-what-malloc-0-should-returns
>
> _______________________________________________
> libusbx mailing list
> libusbx at lists.infradead.org
> http://lists.infradead.org/mailman/listinfo/libusbx