deterministic stress test

Makefile

@@ -6,6 +6,7 @@ SRCS := tests/cobs_encode_max_c.c \
 		tests/test_cobs_decode.cc \
 		tests/test_cobs_decode_inc.cc \
 		tests/test_cobs_decode_inplace.cc \
+		tests/test_many_random_payloads.cc \
 		tests/test_paper_figures.cc \
 		tests/test_wikipedia.cc \
 		tests/unittest_main.cc

make-win.bat

@@ -8,6 +8,7 @@ cl.exe /W4 /WX /MP /EHsc /std:c++20 ^
     tests/test_cobs_encode.cc ^
     tests/test_cobs_encode_inc.cc ^
     tests/test_cobs_encode_inplace.cc ^
+    tests/test_many_random_payloads.cc ^
     tests/test_paper_figures.cc ^
     tests/test_wikipedia.cc ^
     tests/unittest_main.cc ^

tests/test_many_random_payloads.cc

@@ -0,0 +1,61 @@
+#include "../cobs.h"
+#include "byte_vec.h"
+
+#include "doctest.h"
+
+#include <algorithm>
+#include <atomic>
+#include <cstring>
+#include <random>
+#include <thread>
+
+using std::atomic;
+using std::generate;
+using std::max;
+using std::mt19937;
+using std::none_of;
+using std::thread;
+using std::vector;
+
+auto constexpr LEN{ 4u * 1024u * 1024u };
+
+namespace {
+atomic<int> s_iterations{ 250u };
+}
+
+TEST_CASE("many random payloads") {
+  auto const thread_proc{ [](unsigned seed) {
+    mt19937 mt{ seed };  // deterministic
+    byte_vec_t src(LEN), dec(LEN), enc(COBS_ENCODE_MAX(LEN));
+
+    while (--s_iterations > 0) {
+      generate(src.begin(), src.end(), [&]() { return byte_t(mt()); });
+      memset(src.data() + 1000, 0xAA, 256 * 10);  // nonzero run
+
+      size_t enc_len{ 0u };
+      REQUIRE(cobs_encode(src.data(), src.size(), enc.data(), enc.size(), &enc_len) ==
+              COBS_RET_SUCCESS);
+
+      REQUIRE(enc_len >= LEN);
+      REQUIRE(enc_len <= COBS_ENCODE_MAX(LEN));
+      REQUIRE(none_of(enc.data(), enc.data() + enc_len - 1, [](byte_t b) { return !b; }));
+      REQUIRE(enc[enc_len - 1] == 0);
+
+      size_t dec_len{ 0u };
+      REQUIRE(cobs_decode(enc.data(), enc_len, dec.data(), dec.size(), &dec_len) ==
+              COBS_RET_SUCCESS);
+
+      REQUIRE(dec_len == LEN);
+      REQUIRE(src == dec);
+    }
+  } };
+
+  vector<thread> threads;
+  for (auto i{ 0u }, n{ max(1u, thread::hardware_concurrency() - 1) }; i < n; ++i) {
+    threads.emplace_back(thread_proc, i);
+  }
+
+  for (auto& t : threads) {
+    t.join();
+  }
+}