Add tests and build infra for the new ip address parser and formatter

1 files changed, 217 insertions, 0 deletions

diff --git a/tests/t_ip.c b/tests/t_ip.c
new file mode 100644
index 0000000..2fe3a72
--- /dev/null
+++ b/tests/t_ip.c
@@ -0,0 +1,217 @@
+/* This software was written by Dirk Engling <erdgeist@erdgeist.org>
+   It is considered beerware. Prost. Skol. Cheers or whatever.
+
+   $id$ */
+
+/* System */
+#include <stdint.h>
+#include <string.h>
+
+/* Opentracker */
+#include "ot_ip.h"
+#include "test.h"
+
+static char text[64];
+
+static const char *formatted(const char *ip) {
+  text[ip_format(text, ip)] = 0;
+  return text;
+}
+
+/* ------------------------------------------------------------------ IPv4 */
+
+/* ------------------------------------------------------- RFC 5952 output */
+
+static void case_format_rfc5952(void) {
+  test_case("4.1 leading zeros in a group are suppressed");
+  CHECK_STR(formatted("\x20\x01\x0d\xb8\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01"), "2001:db8::1");
+
+  test_case("4.2.1 the longest run of zero groups becomes ::");
+  CHECK_STR(formatted("\x20\x01\x0d\xb8\x00\x00\x00\x00\x00\x00\x00\x00\x00\x02\x00\x01"), "2001:db8::2:1");
+
+  test_case("4.2.2 a single zero group is never shortened");
+  CHECK_STR(formatted("\x20\x01\x0d\xb8\x00\x00\x00\x01\x00\x01\x00\x01\x00\x01\x00\x01"), "2001:db8:0:1:1:1:1:1");
+
+  test_case("4.2.1 a longer run wins over an earlier shorter one");
+  CHECK_STR(formatted("\x20\x01\x00\x00\x00\x00\x00\x01\x00\x00\x00\x00\x00\x00\x00\x01"), "2001:0:0:1::1");
+
+  test_case("4.2.3 of two runs of equal length the first is shortened");
+  CHECK_STR(formatted("\x20\x01\x00\x00\x00\x00\x00\x01\x00\x00\x00\x00\x00\x01\x00\x01"), "2001::1:0:0:1:1");
+
+  test_case("4.3 hexadecimal digits are lower case");
+  CHECK_STR(formatted("\xab\xcd\xef\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01"), "abcd:ef01::1");
+
+  test_case("the unspecified and loopback addresses");
+  CHECK_STR(formatted("\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00"), "::");
+  CHECK_STR(formatted("\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01"), "::1");
+
+  test_case("a full address with no zero group");
+  CHECK_STR(formatted("\x11\x11\x22\x22\x33\x33\x44\x44\x55\x55\x66\x66\x77\x77\x88\x88"), "1111:2222:3333:4444:5555:6666:7777:8888");
+}
+
+static void case_format_trailing_zero_group(void) {
+  /* libowfat's formatter looked two bytes past the address to decide this
+     one, so a peer address ending in a zero group came out as "::" whenever
+     the bytes that happened to follow were not themselves zero. In
+     ot_stats.c those bytes are the peer's port. */
+  const char ip[] = "\x20\x01\x0d\xb8\x00\x01\x00\x02\x00\x03\x00\x04\x00\x05\x00\x00";
+
+  test_case("a single trailing zero group is written out, not shortened");
+  CHECK_STR(formatted(ip), "2001:db8:1:2:3:4:5:0");
+
+  test_case("and nothing beyond the sixteen bytes is consulted");
+  {
+    char padded[20];
+    memcpy(padded, ip, 16);
+    memset(padded + 16, 0xff, 4); /* a port and flags, as a peer record has */
+    CHECK_STR(formatted(padded), "2001:db8:1:2:3:4:5:0");
+    memset(padded + 16, 0x00, 4);
+    CHECK_STR(formatted(padded), "2001:db8:1:2:3:4:5:0");
+  }
+}
+
+static void case_v4mapped(void) {
+  const char mapped[]   = "\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\xff\xff\x0a\x00\x00\x01";
+  const char unmapped[] = "\x20\x01\x0d\xb8\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01";
+
+  test_case("a v4 mapped address is recognised");
+  CHECK(ip_is_v4mapped(mapped));
+  CHECK_INT(ip_is_v4mapped(unmapped), 0);
+
+  /* The tracker has only ever shown the plain quad, so that is the one
+     form the formatter produces; "::ffff:10.0.0.1" is not written anywhere */
+  test_case("a mapped address is written as the plain dotted quad");
+  CHECK_STR(formatted(mapped), "10.0.0.1");
+  CHECK_STR(formatted("\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\xff\xff\x00\x00\x00\x00"), "0.0.0.0");
+  CHECK_STR(formatted("\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\xff\xff\xff\xff\xff\xff"), "255.255.255.255");
+
+  test_case("anything else follows RFC 5952");
+  CHECK_STR(formatted(unmapped), "2001:db8::1");
+}
+
+/* -------------------------------------------------------- RFC 4291 input */
+
+static void case_scan_forms(void) {
+  char ip[16];
+
+  test_case("2.2 form 1, every group spelled out");
+  CHECK_INT(ip_scan("1111:2222:3333:4444:5555:6666:7777:8888", ip), 39);
+  CHECK_MEM(ip, "\x11\x11\x22\x22\x33\x33\x44\x44\x55\x55\x66\x66\x77\x77\x88\x88", 16);
+
+  test_case("2.2 form 2, zero groups abbreviated");
+  CHECK_INT(ip_scan("2001:db8::1", ip), 11);
+  CHECK_MEM(ip, "\x20\x01\x0d\xb8\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01", 16);
+  CHECK_INT(ip_scan("::1", ip), 3);
+  CHECK_MEM(ip, "\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01", 16);
+  CHECK_INT(ip_scan("::", ip), 2);
+  CHECK_MEM(ip, "\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00", 16);
+  CHECK_INT(ip_scan("1::", ip), 3);
+  CHECK_MEM(ip, "\x00\x01\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00", 16);
+
+  test_case("2.2 form 3, a dotted quad in the last two groups");
+  CHECK_INT(ip_scan("::ffff:1.2.3.4", ip), 14);
+  CHECK_MEM(ip, "\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\xff\xff\x01\x02\x03\x04", 16);
+  CHECK_INT(ip_scan("1:2:3:4:5:6:1.2.3.4", ip), 19);
+  CHECK_MEM(ip, "\x00\x01\x00\x02\x00\x03\x00\x04\x00\x05\x00\x06\x01\x02\x03\x04", 16);
+
+  test_case("upper case input is accepted");
+  CHECK_INT(ip_scan("2001:DB8::1", ip), 11);
+  CHECK_MEM(ip, "\x20\x01\x0d\xb8\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\x01", 16);
+}
+
+static void case_scan_accepts_bare_ipv4(void) {
+  char ip[16];
+
+  /* Every address opentracker holds is sixteen bytes, so an operator writing
+     a plain IPv4 address in the config has to land as the v4 mapped form */
+  test_case("a bare IPv4 address is stored v4 mapped");
+  CHECK_INT(ip_scan("10.0.0.1", ip), 8);
+  CHECK_MEM(ip, "\x00\x00\x00\x00\x00\x00\x00\x00\x00\x00\xff\xff\x0a\x00\x00\x01", 16);
+  CHECK(ip_is_v4mapped(ip));
+}
+
+static void case_scan_reports_length(void) {
+  char ip[16];
+
+  /* The config parser continues from where the address ended, which is how
+     "address:port" and "address/bits" are read */
+  test_case("scanning stops at the character after the address");
+  CHECK_INT(ip_scan("10.0.0.1:6969", ip), 8);
+  CHECK_INT(ip_scan("2001:db8::1/64", ip), 11);
+  CHECK_INT(ip_scan("::1]", ip), 3);
+  CHECK_INT(ip_scan("1:2:3:4:5:6:7:8 trailing", ip), 15);
+}
+
+static void case_scan_rejects(void) {
+  char ip[16];
+
+  test_case("an empty or junk string is refused");
+  CHECK_INT(ip_scan("", ip), 0);
+  CHECK_INT(ip_scan("x", ip), 0);
+  CHECK_INT(ip_scan(":", ip), 0);
+  CHECK_INT(ip_scan(":1:2", ip), 0);
+
+  test_case("too few groups without :: are refused");
+  CHECK_INT(ip_scan("1:2:3:4:5:6:7", ip), 0);
+  CHECK_INT(ip_scan("2001:db8:1", ip), 0);
+
+  /* libowfat accepted a prefix of each of these and carried on. Taking only
+     part of a malformed address is worse than refusing it, especially where
+     the text comes from a config file or an X-Forwarded-For header. */
+  test_case("a second :: is refused rather than truncated");
+  CHECK_INT(ip_scan("1::2::3", ip), 0);
+  CHECK_INT(ip_scan("::1::", ip), 0);
+  CHECK_INT(ip_scan("::::", ip), 0);
+
+  test_case("more than eight groups is refused");
+  CHECK_INT(ip_scan("1:2:3:4:5:6:7:8:9", ip), 0);
+
+  test_case("a group of more than four digits is refused, not truncated");
+  CHECK_INT(ip_scan("12345::1", ip), 0);
+  CHECK_INT(ip_scan("fffff::1", ip), 0);
+}
+
+static void case_round_trip(void) {
+  /* Whatever the formatter writes, the scanner has to read back */
+  static const char *const samples[] = {"::",
+                                        "::1",
+                                        "1::",
+                                        "2001:db8::1",
+                                        "2001:db8:1:2:3:4:5:0",
+                                        "1111:2222:3333:4444:5555:6666:7777:8888",
+                                        "2001::1:0:0:1",
+                                        "abcd:ef01::1",
+                                        NULL};
+  char                     ip[16], again[16];
+  size_t                   i;
+
+  test_case("every formatted address scans back to the same bytes");
+  for (i = 0; samples[i]; ++i) {
+    size_t n = ip_scan(samples[i], ip);
+    if (n != strlen(samples[i])) {
+      test_failf(__FILE__, __LINE__, "\"%s\" consumed %zu of %zu", samples[i], n, strlen(samples[i]));
+      return;
+    }
+    text[ip_format(text, ip)] = 0;
+    if (strcmp(text, samples[i])) {
+      test_failf(__FILE__, __LINE__, "\"%s\" formatted back as \"%s\"", samples[i], text);
+      return;
+    }
+    if (!ip_scan(text, again) || memcmp(ip, again, 16)) {
+      test_failf(__FILE__, __LINE__, "\"%s\" did not survive a second pass", samples[i]);
+      return;
+    }
+  }
+  ++test_checks;
+}
+
+void test_ip(void) {
+  case_format_rfc5952();
+  case_format_trailing_zero_group();
+  case_v4mapped();
+  case_scan_forms();
+  case_scan_accepts_bare_ipv4();
+  case_scan_reports_length();
+  case_scan_rejects();
+  case_round_trip();
+}