Fix the compile error once enabled Werror

[xkqian]

On 32 bit operating system, execute the following command "linux32 ./config -Werror -g3 --prefix=xxx; make", it reported error "type mismatch in the file test/cmp_ctx_test.c: TEST_note("total=%d len=%ld msg='%s'\n", msg_total_size, strlen(msg), msg);".
On 32 bit system, size_t is defined as unsigned int, this is the return type of strlen(), but the code use %ld to printf it, when compiling, type mismatch warning will be reported. For printf, %zu should be the right specifier of size_t, change to avoid the warning.

[kroeckx]

We target C89/C90. As far as I know, C90 doesn't know about the z modifier, and so we can't use it. This would mean that all other code making use of TEST_note() with %z needs to be modified.

Note that our BIO_*printf() functions can use it, it has it's own implementation, but test_note() ends up calling vfprintf().

[xkqian]

We target C89/C90. As far as I know, C90 doesn't know about the z modifier, and so we can't use it. This would mean that all other code making use of TEST_note() with %z needs to be modified.

Note that our BIO_*printf() functions can use it, it has it's own implementation, but test_note() ends up calling vfprintf().

Actually, we can find many places use %z in OpenSSL for TEST_note(). Example as below:

./test/poly1305_internal_test.c:1561: TEST_info("Poly1305 test #%d/%zu+%zu failed.",
./test/sparse_array_test.c:52: TEST_note("iteration %zu", i + 1);
./test/sparse_array_test.c:57: TEST_note("iteration %zu / %zu", i + 1, j + 1);
./test/sparse_array_test.c:164: TEST_note("failed at iteration %zu", i + 1);
./test/drbgtest.c:1377: TEST_note("DRBG %zd case %zd block %zd", n / crngt_num_cases,
./test/testutil/tests.c:238:DEFINE_COMPARISONS(size_t, size_t, "%zu")
./test/ssl_cert_table_internal_test.c:32: TEST_error("Invalid table entry for certificate type %s, index %zu",
./test/bioprinttest.c:117: { SIZE_MAX, "%zu", (sizeof(size_t) == 4 ? "4294967295"
./test/bioprinttest.c:124: { SIZE_MAX / 2 + 1, "%zi", (sizeof(size_t) == 4 ? "-2147483648"
./test/bioprinttest.c:127: { 0, "%zu", "0" },
./test/bioprinttest.c:128: { 0, "%zi", "0" },
./test/params_api_test.c:471: TEST_note("iteration %zu var %s", j + 1, int_names[j]);
./test/params_api_test.c:482: TEST_note("iteration %zu var %s", j + 1, uint_names[j]);
./test/siphash_internal_test.c:184: TEST_info("size %zu vs %d and %d", expectedlen,
./test/siphash_internal_test.c:248: TEST_info("SipHash test #%d/%zu+%zu failed.",
./test/dtls_mtu_test.c:107: TEST_info("record %zu for payload %zu", reclen, s);

Fristly we try to use type cast, but to align with the OpenSSL exist style, we change it like this. As we know, %zu is one c99 addition, can anyone help answer whether we are targeting c89/c90, and whether the exist %zu is proper?

[xkqian]

We target C89/C90. As far as I know, C90 doesn't know about the z modifier, and so we can't use it. This would mean that all other code making use of TEST_note() with %z needs to be modified.

Note that our BIO_*printf() functions can use it, it has it's own implementation, but test_note()
ends up calling vfprintf().

---

Sorry, do you mean we should discard all of the z% in the OpenSSL code except in the functions BIO_*printf()?
As we know, %z is one good specifier for size_t, but only be available after c99, it can dynamically detect the bits of the operating system, once we change it like %lu, or ld%, maybe force type cast needed, this will make the code look ugly, do you have any good suggestions about how to change them? I can see many places use the %z.

[mattcaswell]

Note that our BIO_*printf() functions can use it, it has it's own implementation, but test_note() ends up calling vfprintf().

As far as I can see TEST_note is a macro for test_note, which calls test_vprintf_stderr which calls BIO_vprintf. The only place where this is not the case is in bioprinttest.c which has its own custom implementation of test_vprintf_stderr that does call vfprintf directly.

So, I think actually we are ok.

[kroeckx]

The only place where this is not the case is in bioprinttest.c

I of course only found that one, not the one in basic_output.c

[openssl-machine]

24 hours has passed since 'approval: done' was set, but as this PR has been updated in that time the label 'approval: ready to merge' is not being automatically set. Please review the updates and set the label manually.

[xkqian]

24 hours has passed since 'approval: done' was set, but as this PR has been updated in that time the label 'approval: ready to merge' is not being automatically set. Please review the updates and set the label manually.

---

Hi mattcaswell&t8m, What further actions should I do? @mattcaswell @t8m

[t8m]

Merged to master. Thank you.