Tests fail (length of long) / compiler warning

[N1IOX]

Unit tests were failing:

tests.cpp:68:19: error: ‘void* memcpy(void*, const void*, size_t)’ reading 10 bytes from a region of size 8 [-Werror=stringop-overread]
   68 |             memcpy(&buffer[i], &randomLong, 10);
      |             ~~~~~~^~~~~~~~~~~~~~~~~~~~~~~~~~~~~
tests.cpp:67:18: note: source object ‘randomLong’ of size 8
   67 |             long randomLong = random();
      |                  ^~~~~~~~~~

Use sizeof(randomLong)

Note, though, that after fixing this, the unit test can still fail - warned as much in the code: "This may SOMETIMES fail if the random gets lucky and makes something valid"
At least on my machine, this seems to happen more than sometimes...this particular test might need revisiting.

---

Also address a compiler warning:

../src/BERDecode.cpp:92:25: warning: operation on ‘tempVal’ may be undefined [-Wsequence-point]
   92 |                 tempVal = tempVal |= 0xFF000000;
      |                 ~~~~~~~~^~~~~~~~~~~~~~~~~~~~~~~

[N1IOX]

@0neblock

Per the comment in the code, this one specific unit test does fail, and not just sometimes:

    SECTION( "Should fail to parse a corrupt buffer "){
        SNMPPacket* readPacket = new SNMPPacket();
        for(int i = 25; i < 133; i+= 10){
            char old[10] = {0};
            memcpy(old, &buffer[i], 10);
            long randomLong = random();
            memcpy(&buffer[i], &randomLong, sizeof(randomLong));
            // This may SOMETIMES fail if the random gets lucky and makes something valid
            REQUIRE( readPacket->parseFrom(buffer, 200) != SNMP_ERROR_OK );

            memcpy(&buffer[i], old, 10);
            REQUIRE( readPacket->parseFrom(buffer, 200) == SNMP_ERROR_OK );
        }
    }

Both on the automated test machine used on GH, and on a machine I have here, it ends up failing on the 7th pass through that loop...succeeding 6 times before that 7th pass. Consistently.

Results output:

tests.cpp:70: PASSED:
  REQUIRE( readPacket->parseFrom(buffer, 200) != 1 )
with expansion:
  -13 != 1
...
tests.cpp:70: PASSED:
...
tests.cpp:70: PASSED:
...
tests.cpp:70: PASSED:
...
tests.cpp:70: PASSED:
...
tests.cpp:70: PASSED:
...
tests.cpp:70: FAILED:
  REQUIRE( readPacket->parseFrom(buffer, 200) != 1 )
with expansion:
  1 != 1

I was going to merely assume that the 8 bytes of "corruption" landed in the middle of one of the OIDs, whilst still leaving the packet parsable...but you know what they say about those who assume ;)
So instead of that, I added a quick hex dump routine to tests.cpp, and called it before and corruption.
Sure enough, that is the case. Hex dumps attached:

Original.txt
Corrupted.txt

The latter contains the hex dump from 7th pass. Both are "manually" parsed.

The diff of these ends up landing in the final 8 bytes used by test OID ".1.3.6.1.4.1.52420.9999999":

06 0C 2B 06 01 04 01 83 99 44 84 E2 AC 7F
06 0C 2B 06 01 04 F2 41 B1 2E 00 00 00 00

I suggest disabling this one test for now, and revisiting it later.
Agreed?