Attention is currently required from: Nico Huber, Edward O'Callaghan, Angel Pons, Anastasia Klimchuk.
10 comments:
File tests/init_shutdown.c:
Patch Set #5, Line 217: if (strcmp(pathname, "/dev/i2c-254") != 0)
Actually, given that this is test, let's make it an assert. […]
Done
Patch Set #5, Line 219: if ((flags & O_RDWR) != O_RDWR)
assert_int_equal
Done
Patch Set #5, Line 229: if (fd != REALTEK_MST_MOCK_FD)
assert_int_equal
Done
Patch Set #5, Line 231: if (request != I2C_SLAVE)
assert_int_equal
Done
Patch Set #5, Line 235: if (addr != 0x4a)
assert_int_equal
Done
Patch Set #5, Line 243: if (fd != REALTEK_MST_MOCK_FD || sz != 1)
assert_int_equal twice (for FD and for sz)
Done
Patch Set #5, Line 250: if (fd != REALTEK_MST_MOCK_FD)
assert_int_equal
Done
Patch Set #5, Line 252: if (sz == 1 || sz == 2)
assert_in_set
Done
File tests/io_mock.h:
Patch Set #5, Line 78: int (*open)(void *state, const char *pathname, int flags);
In addition to my comment in tests.c: one more line here: […]
Ack
File tests/tests.c:
Patch Set #5, Line 93: return __wrap_open(pathname, flags);
When I said "wrap doesn't need to redirect to another wrap" I meant this: […]
I think it's so unlikely a test will ever care about the difference that it's better to pretend they're the same. The Linux manual for open(2) doesn't describe open64() at all, because whether open() or open64() is used is generally an implementation detail of the C library determined by whether _FILE_OFFSET_BITS is 32 or 64 at build-time, and that further is only relevant to 32-bit systems (per feature_test_macros(7)).
If a test cared about the difference, that would probably indicate a bug in the implementation that would cause it to fail on some systems when built with some settings!
To view, visit change 56911. To unsubscribe, or for help writing mail filters, visit settings.