Make 32-bit version of the unit tests pass.
With corrupt_data being
uint8_t corrupt_data[4] = {0xff, 0xff, 0xff, 0xff};
then "-sizeof corrupt_data" when passed as a int64_t will be passed as
4294967292 instead of -4. I'm not entirely sure why this is but
passing the literal value -4 fixes it. This was discovered while
reviewing https://android-review.googlesource.com/#/c/315014/
Test: All unit tests pass on 64- and 32-bit.
Bug: None.
Change-Id: Ifdd440a45926284d9e9cdb8d4013887143aebc4a
diff --git a/test/avb_slot_verify_unittest.cc b/test/avb_slot_verify_unittest.cc
index c74d0d8..3224024 100644
--- a/test/avb_slot_verify_unittest.cc
+++ b/test/avb_slot_verify_unittest.cc
@@ -185,7 +185,7 @@
EXPECT_EQ(AVB_IO_RESULT_OK,
ops_.avb_ops()->write_to_partition(
ops_.avb_ops(), "vbmeta_a",
- -sizeof corrupt_data, // offset from end
+ -4, // offset from end
sizeof corrupt_data, corrupt_data));
AvbSlotVerifyData* slot_data = NULL;