Quelle  escaping_test.cc   Sprache: C

// Copyright 2017 The Abseil Authors.
//
// Licensed under the Apache License, Version 2.0 (the "License");
// you may not use this file except in compliance with the License.
// You may obtain a copy of the License at
//
//      https://www.apache.org/licenses/LICENSE-2.0
//
// Unless required by applicable law or agreed to in writing, software
// distributed under the License is distributed on an "AS IS" BASIS,
// WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
// See the License for the specific language governing permissions and
// limitations under the License.

#include "absl/strings/escaping.h"

#include <array>
#include <cstddef>
#include <cstdio>
#include <cstring>
#include <initializer_list>
#include <memory>
#include <string>
#include <vector>

#include "gtest/gtest.h"
#include "absl/log/check.h"
#include "absl/strings/str_cat.h"

#include "absl/strings/internal/escaping_test_common.h"
#include "absl/strings/string_view.h"

namespace {

struct epair {
  std::string escaped;
  std::string unescaped;
};

TEST(CEscape, EscapeAndUnescape) {
  const std::string inputs[] = {
      std::string("foo\nxx\r\b\0023"),
      std::string(""),
      std::string("abc"),
      std::string("\1chad_rules"),
      std::string("\1arnar_drools"),
      std::string("xxxx\r\t'\"\\"),
      std::string("\0xx\0", 4),
      std::string("\x01\x31"),
      std::string("abc\xb\x42\141bc"),
      std::string("123\1\x31\x32\x33"),
      std::string("\xc1\xca\x1b\x62\x19o\xcc\x04"),
      std::string(
          "\\\"\xe8\xb0\xb7\xe6\xad\x8c\\\" is Google\\\'s Chinese name"),
  };
  // Do this twice, once for octal escapes and once for hex escapes.
  for (int kind = 0; kind < 4; kind++) {
    for (const std::string& original : inputs) {
      std::string escaped;
      switch (kind) {
        case 0:
          escaped = absl::CEscape(original);
          break;
        case 1:
          escaped = absl::CHexEscape(original);
          break;
        case 2:
          escaped = absl::Utf8SafeCEscape(original);
          break;
        case 3:
          escaped = absl::Utf8SafeCHexEscape(original);
          break;
      }
      std::string unescaped_str;
      EXPECT_TRUE(absl::CUnescape(escaped, &unescaped_str));
      EXPECT_EQ(unescaped_str, original);

      unescaped_str.erase();
      std::string error;
      EXPECT_TRUE(absl::CUnescape(escaped, &unescaped_str, &error));
      EXPECT_EQ(error, "");

      // Check in-place unescaping
      std::string s = escaped;
      EXPECT_TRUE(absl::CUnescape(s, &s));
      ASSERT_EQ(s, original);
    }
  }
  // Check that all possible two character strings can be escaped then
  // unescaped successfully.
  for (int char0 = 0; char0 < 256; char0++) {
    for (int char1 = 0; char1 < 256; char1++) {
      char chars[2];
      chars[0] = char0;
      chars[1] = char1;
      std::string s(chars, 2);
      std::string escaped = absl::CHexEscape(s);
      std::string unescaped;
      EXPECT_TRUE(absl::CUnescape(escaped, &unescaped));
      EXPECT_EQ(s, unescaped);
    }
  }
}

TEST(CEscape, BasicEscaping) {
  epair oct_values[] = {
      {"foo\\rbar\\nbaz\\t", "foo\rbar\nbaz\t"},
      {"\\'full of \\\"sound\\\" and \\\"fury\\\"\\'",
       "'full of \"sound\" and \"fury\"'"},
      {"signi\\\\fying\\\\ nothing\\\\", "signi\\fying\\ nothing\\"},
      {"\\010\\t\\n\\013\\014\\r", "\010\011\012\013\014\015"}
  };
  epair hex_values[] = {
      {"ubik\\rubik\\nubik\\t", "ubik\rubik\nubik\t"},
      {"I\\\'ve just seen a \\\"face\\\"",
       "I've just seen a \"face\""},
      {"hel\\\\ter\\\\skel\\\\ter\\\\", "hel\\ter\\skel\\ter\\"},
      {"\\x08\\t\\n\\x0b\\x0c\\r", "\010\011\012\013\014\015"}
  };
  epair utf8_oct_values[] = {
      {"\xe8\xb0\xb7\xe6\xad\x8c\\r\xe8\xb0\xb7\xe6\xad\x8c\\nbaz\\t",
       "\xe8\xb0\xb7\xe6\xad\x8c\r\xe8\xb0\xb7\xe6\xad\x8c\nbaz\t"},
      {"\\\"\xe8\xb0\xb7\xe6\xad\x8c\\\" is Google\\\'s Chinese name",
       "\"\xe8\xb0\xb7\xe6\xad\x8c\" is Google\'s Chinese name"},
      {"\xe3\x83\xa1\xe3\x83\xbc\xe3\x83\xab\\\\are\\\\Japanese\\\\chars\\\\",
       "\xe3\x83\xa1\xe3\x83\xbc\xe3\x83\xab\\are\\Japanese\\chars\\"},
      {"\xed\x81\xac\xeb\xa1\xac\\010\\t\\n\\013\\014\\r",
       "\xed\x81\xac\xeb\xa1\xac\010\011\012\013\014\015"}
  };
  epair utf8_hex_values[] = {
      {"\x20\xe4\xbd\xa0\\t\xe5\xa5\xbd,\\r!\\n",
       "\x20\xe4\xbd\xa0\t\xe5\xa5\xbd,\r!\n"},
      {"\xe8\xa9\xa6\xe9\xa8\x93\\\' means \\\"test\\\"",
       "\xe8\xa9\xa6\xe9\xa8\x93\' means \"test\""},
      {"\\\\\xe6\x88\x91\\\\:\\\\\xe6\x9d\xa8\xe6\xac\xa2\\\\",
       "\\\xe6\x88\x91\\:\\\xe6\x9d\xa8\xe6\xac\xa2\\"},
      {"\xed\x81\xac\xeb\xa1\xac\\x08\\t\\n\\x0b\\x0c\\r",
       "\xed\x81\xac\xeb\xa1\xac\010\011\012\013\014\015"}
  };

  for (const epair& val : oct_values) {
    std::string escaped = absl::CEscape(val.unescaped);
    EXPECT_EQ(escaped, val.escaped);
  }
  for (const epair& val : hex_values) {
    std::string escaped = absl::CHexEscape(val.unescaped);
    EXPECT_EQ(escaped, val.escaped);
  }
  for (const epair& val : utf8_oct_values) {
    std::string escaped = absl::Utf8SafeCEscape(val.unescaped);
    EXPECT_EQ(escaped, val.escaped);
  }
  for (const epair& val : utf8_hex_values) {
    std::string escaped = absl::Utf8SafeCHexEscape(val.unescaped);
    EXPECT_EQ(escaped, val.escaped);
  }
}

TEST(Unescape, BasicFunction) {
  epair tests[] =
    {{"", ""},
     {"\\u0030", "0"},
     {"\\u00A3", "\xC2\xA3"},
     {"\\u22FD", "\xE2\x8B\xBD"},
     {"\\U00010000", "\xF0\x90\x80\x80"},
     {"\\U0010FFFD", "\xF4\x8F\xBF\xBD"}};
  for (const epair& val : tests) {
    std::string out;
    EXPECT_TRUE(absl::CUnescape(val.escaped, &out));
    EXPECT_EQ(out, val.unescaped);
  }
  std::string bad[] = {"\\u1",         // too short
                       "\\U1",         // too short
                       "\\Uffffff",    // exceeds 0x10ffff (largest Unicode)
                       "\\U00110000",  // exceeds 0x10ffff (largest Unicode)
                       "\\uD835",      // surrogate character (D800-DFFF)
                       "\\U0000DD04",  // surrogate character (D800-DFFF)
                       "\\777",        // exceeds 0xff
                       "\\xABCD"};     // exceeds 0xff
  for (const std::string& e : bad) {
    std::string error;
    std::string out;
    EXPECT_FALSE(absl::CUnescape(e, &out, &error));
    EXPECT_FALSE(error.empty());

    out.erase();
    EXPECT_FALSE(absl::CUnescape(e, &out));
  }
}

class CUnescapeTest : public testing::Test {
 protected:
  static const char kStringWithMultipleOctalNulls[];
  static const char kStringWithMultipleHexNulls[];
  static const char kStringWithMultipleUnicodeNulls[];

  std::string result_string_;
};

const char CUnescapeTest::kStringWithMultipleOctalNulls[] =
    "\\0\\n"    // null escape \0 plus newline
    "0\\n"      // just a number 0 (not a null escape) plus newline
    "\\00\\12"  // null escape \00 plus octal newline code
    "\\000";    // null escape \000

// This has the same ingredients as kStringWithMultipleOctalNulls
// but with \x hex escapes instead of octal escapes.
const char CUnescapeTest::kStringWithMultipleHexNulls[] =
    "\\x0\\n"
    "0\\n"
    "\\x00\\xa"
    "\\x000";

const char CUnescapeTest::kStringWithMultipleUnicodeNulls[] =
    "\\u0000\\n"    // short-form (4-digit) null escape plus newline
    "0\\n"          // just a number 0 (not a null escape) plus newline
    "\\U00000000";  // long-form (8-digit) null escape

TEST_F(CUnescapeTest, Unescapes1CharOctalNull) {
  std::string original_string = "\\0";
  EXPECT_TRUE(absl::CUnescape(original_string, &result_string_));
  EXPECT_EQ(std::string("\0", 1), result_string_);
}

TEST_F(CUnescapeTest, Unescapes2CharOctalNull) {
  std::string original_string = "\\00";
  EXPECT_TRUE(absl::CUnescape(original_string, &result_string_));
  EXPECT_EQ(std::string("\0", 1), result_string_);
}

TEST_F(CUnescapeTest, Unescapes3CharOctalNull) {
  std::string original_string = "\\000";
  EXPECT_TRUE(absl::CUnescape(original_string, &result_string_));
  EXPECT_EQ(std::string("\0", 1), result_string_);
}

TEST_F(CUnescapeTest, Unescapes1CharHexNull) {
  std::string original_string = "\\x0";
  EXPECT_TRUE(absl::CUnescape(original_string, &result_string_));
  EXPECT_EQ(std::string("\0", 1), result_string_);
}

TEST_F(CUnescapeTest, Unescapes2CharHexNull) {
  std::string original_string = "\\x00";
  EXPECT_TRUE(absl::CUnescape(original_string, &result_string_));
  EXPECT_EQ(std::string("\0", 1), result_string_);
}

TEST_F(CUnescapeTest, Unescapes3CharHexNull) {
  std::string original_string = "\\x000";
  EXPECT_TRUE(absl::CUnescape(original_string, &result_string_));
  EXPECT_EQ(std::string("\0", 1), result_string_);
}

TEST_F(CUnescapeTest, Unescapes4CharUnicodeNull) {
  std::string original_string = "\\u0000";
  EXPECT_TRUE(absl::CUnescape(original_string, &result_string_));
  EXPECT_EQ(std::string("\0", 1), result_string_);
}

TEST_F(CUnescapeTest, Unescapes8CharUnicodeNull) {
  std::string original_string = "\\U00000000";
  EXPECT_TRUE(absl::CUnescape(original_string, &result_string_));
  EXPECT_EQ(std::string("\0", 1), result_string_);
}

TEST_F(CUnescapeTest, UnescapesMultipleOctalNulls) {
  std::string original_string(kStringWithMultipleOctalNulls);
  EXPECT_TRUE(absl::CUnescape(original_string, &result_string_));
  // All escapes, including newlines and null escapes, should have been
  // converted to the equivalent characters.
  EXPECT_EQ(std::string("\0\n"
                        "0\n"
                        "\0\n"
                        "\0",
                        7),
            result_string_);
}


TEST_F(CUnescapeTest, UnescapesMultipleHexNulls) {
  std::string original_string(kStringWithMultipleHexNulls);
  EXPECT_TRUE(absl::CUnescape(original_string, &result_string_));
  EXPECT_EQ(std::string("\0\n"
                        "0\n"
                        "\0\n"
                        "\0",
                        7),
            result_string_);
}

TEST_F(CUnescapeTest, UnescapesMultipleUnicodeNulls) {
  std::string original_string(kStringWithMultipleUnicodeNulls);
  EXPECT_TRUE(absl::CUnescape(original_string, &result_string_));
  EXPECT_EQ(std::string("\0\n"
                        "0\n"
                        "\0",
                        5),
            result_string_);
}

static struct {
  absl::string_view plaintext;
  absl::string_view cyphertext;
} const base64_tests[] = {
    // Empty string.
    {{"", 0}, {"", 0}},
    {{nullptr, 0},
     {"", 0}},  // if length is zero, plaintext ptr must be ignored!

    // Basic bit patterns;
    // values obtained with "echo -n '...' | uuencode -m test"

    {{"\000", 1}, "AA=="},
    {{"\001", 1}, "AQ=="},
    {{"\002", 1}, "Ag=="},
    {{"\004", 1}, "BA=="},
    {{"\010", 1}, "CA=="},
    {{"\020", 1}, "EA=="},
    {{"\040", 1}, "IA=="},
    {{"\100", 1}, "QA=="},
    {{"\200", 1}, "gA=="},

    {{"\377", 1}, "/w=="},
    {{"\376", 1}, "/g=="},
    {{"\375", 1}, "/Q=="},
    {{"\373", 1}, "+w=="},
    {{"\367", 1}, "9w=="},
    {{"\357", 1}, "7w=="},
    {{"\337", 1}, "3w=="},
    {{"\277", 1}, "vw=="},
    {{"\177", 1}, "fw=="},
    {{"\000\000", 2}, "AAA="},
    {{"\000\001", 2}, "AAE="},
    {{"\000\002", 2}, "AAI="},
    {{"\000\004", 2}, "AAQ="},
    {{"\000\010", 2}, "AAg="},
    {{"\000\020", 2}, "ABA="},
    {{"\000\040", 2}, "ACA="},
    {{"\000\100", 2}, "AEA="},
    {{"\000\200", 2}, "AIA="},
    {{"\001\000", 2}, "AQA="},
    {{"\002\000", 2}, "AgA="},
    {{"\004\000", 2}, "BAA="},
    {{"\010\000", 2}, "CAA="},
    {{"\020\000", 2}, "EAA="},
    {{"\040\000", 2}, "IAA="},
    {{"\100\000", 2}, "QAA="},
    {{"\200\000", 2}, "gAA="},

    {{"\377\377", 2}, "//8="},
    {{"\377\376", 2}, "//4="},
    {{"\377\375", 2}, "//0="},
    {{"\377\373", 2}, "//s="},
    {{"\377\367", 2}, "//c="},
    {{"\377\357", 2}, "/+8="},
    {{"\377\337", 2}, "/98="},
    {{"\377\277", 2}, "/78="},
    {{"\377\177", 2}, "/38="},
    {{"\376\377", 2}, "/v8="},
    {{"\375\377", 2}, "/f8="},
    {{"\373\377", 2}, "+/8="},
    {{"\367\377", 2}, "9/8="},
    {{"\357\377", 2}, "7/8="},
    {{"\337\377", 2}, "3/8="},
    {{"\277\377", 2}, "v/8="},
    {{"\177\377", 2}, "f/8="},

    {{"\000\000\000", 3}, "AAAA"},
    {{"\000\000\001", 3}, "AAAB"},
    {{"\000\000\002", 3}, "AAAC"},
    {{"\000\000\004", 3}, "AAAE"},
    {{"\000\000\010", 3}, "AAAI"},
    {{"\000\000\020", 3}, "AAAQ"},
    {{"\000\000\040", 3}, "AAAg"},
    {{"\000\000\100", 3}, "AABA"},
    {{"\000\000\200", 3}, "AACA"},
    {{"\000\001\000", 3}, "AAEA"},
    {{"\000\002\000", 3}, "AAIA"},
    {{"\000\004\000", 3}, "AAQA"},
    {{"\000\010\000", 3}, "AAgA"},
    {{"\000\020\000", 3}, "ABAA"},
    {{"\000\040\000", 3}, "ACAA"},
    {{"\000\100\000", 3}, "AEAA"},
    {{"\000\200\000", 3}, "AIAA"},
    {{"\001\000\000", 3}, "AQAA"},
    {{"\002\000\000", 3}, "AgAA"},
    {{"\004\000\000", 3}, "BAAA"},
    {{"\010\000\000", 3}, "CAAA"},
    {{"\020\000\000", 3}, "EAAA"},
    {{"\040\000\000", 3}, "IAAA"},
    {{"\100\000\000", 3}, "QAAA"},
    {{"\200\000\000", 3}, "gAAA"},

    {{"\377\377\377", 3}, "////"},
    {{"\377\377\376", 3}, "///+"},
    {{"\377\377\375", 3}, "///9"},
    {{"\377\377\373", 3}, "///7"},
    {{"\377\377\367", 3}, "///3"},
    {{"\377\377\357", 3}, "///v"},
    {{"\377\377\337", 3}, "///f"},
    {{"\377\377\277", 3}, "//+/"},
    {{"\377\377\177", 3}, "//9/"},
    {{"\377\376\377", 3}, "//7/"},
    {{"\377\375\377", 3}, "//3/"},
    {{"\377\373\377", 3}, "//v/"},
    {{"\377\367\377", 3}, "//f/"},
    {{"\377\357\377", 3}, "/+//"},
    {{"\377\337\377", 3}, "/9//"},
    {{"\377\277\377", 3}, "/7//"},
    {{"\377\177\377", 3}, "/3//"},
    {{"\376\377\377", 3}, "/v//"},
    {{"\375\377\377", 3}, "/f//"},
    {{"\373\377\377", 3}, "+///"},
    {{"\367\377\377", 3}, "9///"},
    {{"\357\377\377", 3}, "7///"},
    {{"\337\377\377", 3}, "3///"},
    {{"\277\377\377", 3}, "v///"},
    {{"\177\377\377", 3}, "f///"},

    // Random numbers: values obtained with
    //
    //  #! /bin/bash
    //  dd bs=$1 count=1 if=/dev/random of=/tmp/bar.random
    //  od -N $1 -t o1 /tmp/bar.random
    //  uuencode -m test < /tmp/bar.random
    //
    // where $1 is the number of bytes (2, 3)

    {{"\243\361", 2}, "o/E="},
    {{"\024\167", 2}, "FHc="},
    {{"\313\252", 2}, "y6o="},
    {{"\046\041", 2}, "JiE="},
    {{"\145\236", 2}, "ZZ4="},
    {{"\254\325", 2}, "rNU="},
    {{"\061\330", 2}, "Mdg="},
    {{"\245\032", 2}, "pRo="},
    {{"\006\000", 2}, "BgA="},
    {{"\375\131", 2}, "/Vk="},
    {{"\303\210", 2}, "w4g="},
    {{"\040\037", 2}, "IB8="},
    {{"\261\372", 2}, "sfo="},
    {{"\335\014", 2}, "3Qw="},
    {{"\233\217", 2}, "m48="},
    {{"\373\056", 2}, "+y4="},
    {{"\247\232", 2}, "p5o="},
    {{"\107\053", 2}, "Rys="},
    {{"\204\077", 2}, "hD8="},
    {{"\276\211", 2}, "vok="},
    {{"\313\110", 2}, "y0g="},
    {{"\363\376", 2}, "8/4="},
    {{"\251\234", 2}, "qZw="},
    {{"\103\262", 2}, "Q7I="},
    {{"\142\312", 2}, "Yso="},
    {{"\067\211", 2}, "N4k="},
    {{"\220\001", 2}, "kAE="},
    {{"\152\240", 2}, "aqA="},
    {{"\367\061", 2}, "9zE="},
    {{"\133\255", 2}, "W60="},
    {{"\176\035", 2}, "fh0="},
    {{"\032\231", 2}, "Gpk="},

    {{"\013\007\144", 3}, "Cwdk"},
    {{"\030\112\106", 3}, "GEpG"},
    {{"\047\325\046", 3}, "J9Um"},
    {{"\310\160\022", 3}, "yHAS"},
    {{"\131\100\237", 3}, "WUCf"},
    {{"\064\342\134", 3}, "NOJc"},
    {{"\010\177\004", 3}, "CH8E"},
    {{"\345\147\205", 3}, "5WeF"},
    {{"\300\343\360", 3}, "wOPw"},
    {{"\061\240\201", 3}, "MaCB"},
    {{"\225\333\044", 3}, "ldsk"},
    {{"\215\137\352", 3}, "jV/q"},
    {{"\371\147\160", 3}, "+Wdw"},
    {{"\030\320\051", 3}, "GNAp"},
    {{"\044\174\241", 3}, "JHyh"},
    {{"\260\127\037", 3}, "sFcf"},
    {{"\111\045\033", 3}, "SSUb"},
    {{"\202\114\107", 3}, "gkxH"},
    {{"\057\371\042", 3}, "L/ki"},
    {{"\223\247\244", 3}, "k6ek"},
    {{"\047\216\144", 3}, "J45k"},
    {{"\203\070\327", 3}, "gzjX"},
    {{"\247\140\072", 3}, "p2A6"},
    {{"\124\115\116", 3}, "VE1O"},
    {{"\157\162\050", 3}, "b3Io"},
    {{"\357\223\004", 3}, "75ME"},
    {{"\052\117\156", 3}, "Kk9u"},
    {{"\347\154\000", 3}, "52wA"},
    {{"\303\012\142", 3}, "wwpi"},
    {{"\060\035\362", 3}, "MB3y"},
    {{"\130\226\361", 3}, "WJbx"},
    {{"\173\013\071", 3}, "ews5"},
    {{"\336\004\027", 3}, "3gQX"},
    {{"\357\366\234", 3}, "7/ac"},
    {{"\353\304\111", 3}, "68RJ"},
    {{"\024\264\131", 3}, "FLRZ"},
    {{"\075\114\251", 3}, "PUyp"},
    {{"\315\031\225", 3}, "zRmV"},
    {{"\154\201\276", 3}, "bIG+"},
    {{"\200\066\072", 3}, "gDY6"},
    {{"\142\350\267", 3}, "Yui3"},
    {{"\033\000\166", 3}, "GwB2"},
    {{"\210\055\077", 3}, "iC0/"},
    {{"\341\037\124", 3}, "4R9U"},
    {{"\161\103\152", 3}, "cUNq"},
    {{"\270\142\131", 3}, "uGJZ"},
    {{"\337\076\074", 3}, "3z48"},
    {{"\375\106\362", 3}, "/Uby"},
    {{"\227\301\127", 3}, "l8FX"},
    {{"\340\002\234", 3}, "4AKc"},
    {{"\121\064\033", 3}, "UTQb"},
    {{"\157\134\143", 3}, "b1xj"},
    {{"\247\055\327", 3}, "py3X"},
    {{"\340\142\005", 3}, "4GIF"},
    {{"\060\260\143", 3}, "MLBj"},
    {{"\075\203\170", 3}, "PYN4"},
    {{"\143\160\016", 3}, "Y3AO"},
    {{"\313\013\063", 3}, "ywsz"},
    {{"\174\236\135", 3}, "fJ5d"},
    {{"\103\047\026", 3}, "QycW"},
    {{"\365\005\343", 3}, "9QXj"},
    {{"\271\160\223", 3}, "uXCT"},
    {{"\362\255\172", 3}, "8q16"},
    {{"\113\012\015", 3}, "SwoN"},

    // various lengths, generated by this python script:
    //
    // from std::string import lowercase as lc
    // for i in range(27):
    //   print '{ %2d, "%s",%s "%s" },' % (i, lc[:i], ' ' * (26-i),
    //                                     lc[:i].encode('base64').strip())

    {{"", 0}, {"", 0}},
    {"a", "YQ=="},
    {"ab", "YWI="},
    {"abc", "YWJj"},
    {"abcd", "YWJjZA=="},
    {"abcde", "YWJjZGU="},
    {"abcdef", "YWJjZGVm"},
    {"abcdefg", "YWJjZGVmZw=="},
    {"abcdefgh", "YWJjZGVmZ2g="},
    {"abcdefghi", "YWJjZGVmZ2hp"},
    {"abcdefghij", "YWJjZGVmZ2hpag=="},
    {"abcdefghijk", "YWJjZGVmZ2hpams="},
    {"abcdefghijkl", "YWJjZGVmZ2hpamts"},
    {"abcdefghijklm", "YWJjZGVmZ2hpamtsbQ=="},
    {"abcdefghijklmn", "YWJjZGVmZ2hpamtsbW4="},
    {"abcdefghijklmno", "YWJjZGVmZ2hpamtsbW5v"},
    {"abcdefghijklmnop", "YWJjZGVmZ2hpamtsbW5vcA=="},
    {"abcdefghijklmnopq", "YWJjZGVmZ2hpamtsbW5vcHE="},
    {"abcdefghijklmnopqr", "YWJjZGVmZ2hpamtsbW5vcHFy"},
    {"abcdefghijklmnopqrs", "YWJjZGVmZ2hpamtsbW5vcHFycw=="},
    {"abcdefghijklmnopqrst", "YWJjZGVmZ2hpamtsbW5vcHFyc3Q="},
    {"abcdefghijklmnopqrstu", "YWJjZGVmZ2hpamtsbW5vcHFyc3R1"},
    {"abcdefghijklmnopqrstuv", "YWJjZGVmZ2hpamtsbW5vcHFyc3R1dg=="},
    {"abcdefghijklmnopqrstuvw", "YWJjZGVmZ2hpamtsbW5vcHFyc3R1dnc="},
    {"abcdefghijklmnopqrstuvwx", "YWJjZGVmZ2hpamtsbW5vcHFyc3R1dnd4"},
    {"abcdefghijklmnopqrstuvwxy", "YWJjZGVmZ2hpamtsbW5vcHFyc3R1dnd4eQ=="},
    {"abcdefghijklmnopqrstuvwxyz", "YWJjZGVmZ2hpamtsbW5vcHFyc3R1dnd4eXo="},
};

template <typename StringType>
void TestEscapeAndUnescape() {
  // Check the short strings; this tests the math (and boundaries)
  for (const auto& tc : base64_tests) {
    // Test plain base64.
    StringType encoded("this junk should be ignored");
    absl::Base64Escape(tc.plaintext, &encoded);
    EXPECT_EQ(encoded, tc.cyphertext);
    EXPECT_EQ(absl::Base64Escape(tc.plaintext), tc.cyphertext);

    StringType decoded("this junk should be ignored");
    EXPECT_TRUE(absl::Base64Unescape(encoded, &decoded));
    EXPECT_EQ(decoded, tc.plaintext);

    StringType websafe_with_padding(tc.cyphertext);
    for (unsigned int c = 0; c < websafe_with_padding.size(); ++c) {
      if ('+' == websafe_with_padding[c]) websafe_with_padding[c] = '-';
      if ('/' == websafe_with_padding[c]) websafe_with_padding[c] = '_';
      // Intentionally keeping padding aka '='.
    }

    // Test plain websafe (aka without padding).
    StringType websafe(websafe_with_padding);
    for (unsigned int c = 0; c < websafe.size(); ++c) {
      if ('=' == websafe[c]) {
        websafe.resize(c);
        break;
      }
    }
    encoded = "this junk should be ignored";
    absl::WebSafeBase64Escape(tc.plaintext, &encoded);
    EXPECT_EQ(encoded, websafe);
    EXPECT_EQ(absl::WebSafeBase64Escape(tc.plaintext), websafe);

    decoded = "this junk should be ignored";
    EXPECT_TRUE(absl::WebSafeBase64Unescape(websafe, &decoded));
    EXPECT_EQ(decoded, tc.plaintext);
  }

  // Now try the long strings, this tests the streaming
  for (const auto& tc : absl::strings_internal::base64_strings()) {
    StringType buffer;
    absl::WebSafeBase64Escape(tc.plaintext, &buffer);
    EXPECT_EQ(tc.cyphertext, buffer);
    EXPECT_EQ(absl::WebSafeBase64Escape(tc.plaintext), tc.cyphertext);
  }

  // Verify the behavior when decoding bad data
  {
    absl::string_view data_set[] = {"ab-/", absl::string_view("\0bcd", 4),
                                    absl::string_view("abc.\0", 5)};
    for (absl::string_view bad_data : data_set) {
      StringType buf;
      EXPECT_FALSE(absl::Base64Unescape(bad_data, &buf));
      EXPECT_FALSE(absl::WebSafeBase64Unescape(bad_data, &buf));
      EXPECT_TRUE(buf.empty());
    }
  }
}

TEST(Base64, EscapeAndUnescape) {
  TestEscapeAndUnescape<std::string>();
}

TEST(Base64, Padding) {
  // Padding is optional.
  // '.' is an acceptable padding character, just like '='.
  std::initializer_list<absl::string_view> good_padding = {
    "YQ",
    "YQ==",
    "YQ=.",
    "YQ.=",
    "YQ..",
  };
  for (absl::string_view b64 : good_padding) {
    std::string decoded;
    EXPECT_TRUE(absl::Base64Unescape(b64, &decoded));
    EXPECT_EQ(decoded, "a");
    std::string websafe_decoded;
    EXPECT_TRUE(absl::WebSafeBase64Unescape(b64, &websafe_decoded));
    EXPECT_EQ(websafe_decoded, "a");
  }
  std::initializer_list<absl::string_view> bad_padding = {
    "YQ=",
    "YQ.",
    "YQ===",
    "YQ==.",
    "YQ=.=",
    "YQ=..",
    "YQ.==",
    "YQ.=.",
    "YQ..=",
    "YQ...",
    "YQ====",
    "YQ....",
    "YQ=====",
    "YQ.....",
  };
  for (absl::string_view b64 : bad_padding) {
    std::string decoded;
    EXPECT_FALSE(absl::Base64Unescape(b64, &decoded));
    std::string websafe_decoded;
    EXPECT_FALSE(absl::WebSafeBase64Unescape(b64, &websafe_decoded));
  }
}

TEST(Base64, DISABLED_HugeData) {
  const size_t kSize = size_t(3) * 1000 * 1000 * 1000;
  static_assert(kSize % 3 == 0, "kSize must be divisible by 3");
  const std::string huge(kSize, 'x');

  std::string escaped;
  absl::Base64Escape(huge, &escaped);

  // Generates the string that should match a base64 encoded "xxx..." string.
  // "xxx" in base64 is "eHh4".
  std::string expected_encoding;
  expected_encoding.reserve(kSize / 3 * 4);
  for (size_t i = 0; i < kSize / 3; ++i) {
    expected_encoding.append("eHh4");
  }
  EXPECT_EQ(expected_encoding, escaped);

  std::string unescaped;
  EXPECT_TRUE(absl::Base64Unescape(escaped, &unescaped));
  EXPECT_EQ(huge, unescaped);
}

TEST(Escaping, HexStringToBytesBackToHex) {
  std::string bytes, hex;

  constexpr absl::string_view kTestHexLower =  "1c2f0032f40123456789abcdef";
  constexpr absl::string_view kTestHexUpper =  "1C2F0032F40123456789ABCDEF";
  constexpr absl::string_view kTestBytes = absl::string_view(
      "\x1c\x2f\x00\x32\xf4\x01\x23\x45\x67\x89\xab\xcd\xef", 13);

  EXPECT_TRUE(absl::HexStringToBytes(kTestHexLower, &bytes));
  EXPECT_EQ(bytes, kTestBytes);

  EXPECT_TRUE(absl::HexStringToBytes(kTestHexUpper, &bytes));
  EXPECT_EQ(bytes, kTestBytes);

  hex = absl::BytesToHexString(kTestBytes);
  EXPECT_EQ(hex, kTestHexLower);

  // Same buffer.
  // We do not care if this works since we do not promise it in the contract.
  // The purpose of this test is to to see if the program will crash or if
  // sanitizers will catch anything.
  bytes = std::string(kTestHexUpper);
  (void)absl::HexStringToBytes(bytes, &bytes);

  // Length not a multiple of two.
  EXPECT_FALSE(absl::HexStringToBytes("1c2f003", &bytes));

  // Not hex.
  EXPECT_FALSE(absl::HexStringToBytes("1c2f00ft", &bytes));

  // Empty input.
  bytes = "abc";
  EXPECT_TRUE(absl::HexStringToBytes("", &bytes));
  EXPECT_EQ("", bytes);  // Results in empty output.
}

TEST(HexAndBack, HexStringToBytes_and_BytesToHexString) {
  std::string hex_mixed = "0123456789abcdefABCDEF";
  std::string bytes_expected = "\x01\x23\x45\x67\x89\xab\xcd\xef\xAB\xCD\xEF";
  std::string hex_only_lower = "0123456789abcdefabcdef";

  std::string bytes_result = absl::HexStringToBytes(hex_mixed);
  EXPECT_EQ(bytes_expected, bytes_result);

  std::string prefix_valid = hex_mixed + "?";
  std::string prefix_valid_result = absl::HexStringToBytes(
      absl::string_view(prefix_valid.data(), prefix_valid.size() - 1));
  EXPECT_EQ(bytes_expected, prefix_valid_result);

  std::string infix_valid = "?" + hex_mixed + "???";
  std::string infix_valid_result = absl::HexStringToBytes(
      absl::string_view(infix_valid.data() + 1, hex_mixed.size()));
  EXPECT_EQ(bytes_expected, infix_valid_result);

  std::string hex_result = absl::BytesToHexString(bytes_expected);
  EXPECT_EQ(hex_only_lower, hex_result);
}

}  // namespace