I've noticed from commit f93983a7 &
the NEWS file that libdvbpsi is on version 1.2.0 now, but in the header
file dvbpsi.h, DVBPSI_VERSION still says 1.1.2 and DVBPSI_VERSION_INT
still says 1.1.1 ...
Signed-off-by: Michael Ira Krufky <mkrufky@linuxtv.org>
Signed-off-by: Jean-Paul Saman <jpsaman@videolan.org>

Fix error: unknown type name 'ssize_t'; did you mean 'size_t' when building on
OSX Maverick.
Signed-off-by: Jean-Paul Saman <jpsaman@videolan.org>

Signed-off-by: Michael Ira Krufky <mkrufky@linuxtv.org>

check_cc_pid: fix error: format specifies type 'long' but the argument has type 'int64_t' (aka 'long long') [-Werror,-Wformat]

gcc -DHAVE_CONFIG_H -I. -I..     -g -O2 -Wall -Werror --std=gnu99 -D_GNU_SOURCE -Wpointer-arith -Wcast-align -Wcast-qual -Wstrict-prototypes -Wshadow -Waggregate-return -Wmissing-prototypes -Wnested-externs -Wsign-compare -DDVBPSI_DIST -MT check_cc_pid-check_cc_pid.o -MD -MP -MF .deps/check_cc_pid-check_cc_pid.Tpo -c -o check_cc_pid-check_cc_pid.o `test -f 'check_cc_pid.c' || echo './'`check_cc_pid.c
check_cc_pid.c:69:20: error: format specifies type 'long' but the argument has type 'int64_t' (aka 'long long') [-Werror,-Wformat]
                   n, pid, pid, cc,
                   ^
1 error generated.
make[3]: *** [check_cc_pid-check_cc_pid.o] Error 1
make[2]: *** [all-recursive] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2
Signed-off-by: Michael Ira Krufky <mkrufky@linuxtv.org>

dump_pids: fix error: format specifies type 'long' but the argument has type 'int64_t' (aka 'long long') [-Werror,-Wformat]

gcc -DHAVE_CONFIG_H -I. -I..     -g -O2 -Wall -Werror --std=gnu99 -D_GNU_SOURCE -Wpointer-arith -Wcast-align -Wcast-qual -Wstrict-prototypes -Wshadow -Waggregate-return -Wmissing-prototypes -Wnested-externs -Wsign-compare -DDVBPSI_DIST -MT dump_pids-dump_pids.o -MD -MP -MF .deps/dump_pids-dump_pids.Tpo -c -o dump_pids-dump_pids.o `test -f 'dump_pids.c' || echo './'`dump_pids.c
dump_pids.c:56:54: error: format specifies type 'long' but the argument has type 'int64_t' (aka 'long long') [-Werror,-Wformat]
        printf("packet %ld, pid %u (0x%x), cc %d\n", n, pid, pid, cc );
                       ~~~                           ^
                       %lld
1 error generated.
make[3]: *** [dump_pids-dump_pids.o] Error 1
make[2]: *** [all-recursive] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2
Signed-off-by: Michael Ira Krufky <mkrufky@linuxtv.org>

Beware using the AdressSanitizer from GCC 4.8 uses a lot of memory. You need
more then 8GB of memory in your machine.

Use strtoll(optarg, NULL, 10).

Check for p_stream to be a valid pointer.

Forgot to terminate the case for option 'a'.

Check filedescriptors for having a valid value.

Check i_fd to have a valid value.

pf_write() returns a ssize_t which was stored inside a size_t (an unsigned).
Later on this value was used in a check against 0.

The pointer to p_subdec should have been evaluated on line 119.

The wrong pointer was checked for having a null value leading to the
null pointer dereference bug.

The bug occurs due to a missing check for allocation failure of p_decoded->i_text.
Instead p_decoded is checked again.

The assert() dereferenced the p_dvbpsi pointer before it was checked for null-value.

The code does not leak a socket descriptor while looping over addrinfo pointer structs.However initializing s_ctl to -1 again after calling close(s_ctl) is good practise.

The code does not leak a socket descriptor while looping over addrinfo pointer structs.However initializing s_ctl to -1 again after calling close(s_ctl) is good practise.

The error path for dvbpsi_AttachDemux() in handle_atsc_MGT failed to release
allocated resources.

The error path for dvbpsi_pmt_attach() in handle_PAT failed to release
allocated resources.

The value of i_nb_programs could exceed the MAX_PROGRAMS limit and thus
produce a buffer overrun.

- syslog.h is not available on Windows platforms.
- O_NONBLOCK is not defined for Windows platforms.