re2 1.20.0 → 1.20.2

package/vendor/abseil-cpp/absl/strings/internal/str_format/convert_test.cc

@@ -0,0 +1,1257 @@
+// Copyright 2020 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 <errno.h>
+#include <stdarg.h>
+#include <stdio.h>
+
+#include <cctype>
+#include <cmath>
+#include <limits>
+#include <string>
+#include <thread>  // NOLINT
+
+#include "gmock/gmock.h"
+#include "gtest/gtest.h"
+#include "absl/base/attributes.h"
+#include "absl/base/internal/raw_logging.h"
+#include "absl/strings/internal/str_format/bind.h"
+#include "absl/strings/match.h"
+#include "absl/types/optional.h"
+
+namespace absl {
+ABSL_NAMESPACE_BEGIN
+namespace str_format_internal {
+namespace {
+
+struct NativePrintfTraits {
+  bool hex_float_has_glibc_rounding;
+  bool hex_float_prefers_denormal_repr;
+  bool hex_float_uses_minimal_precision_when_not_specified;
+  bool hex_float_optimizes_leading_digit_bit_count;
+};
+
+template <typename T, size_t N>
+size_t ArraySize(T (&)[N]) {
+  return N;
+}
+
+std::string LengthModFor(float) { return ""; }
+std::string LengthModFor(double) { return ""; }
+std::string LengthModFor(long double) { return "L"; }
+std::string LengthModFor(char) { return "hh"; }
+std::string LengthModFor(signed char) { return "hh"; }
+std::string LengthModFor(unsigned char) { return "hh"; }
+std::string LengthModFor(short) { return "h"; }           // NOLINT
+std::string LengthModFor(unsigned short) { return "h"; }  // NOLINT
+std::string LengthModFor(int) { return ""; }
+std::string LengthModFor(unsigned) { return ""; }
+std::string LengthModFor(long) { return "l"; }                 // NOLINT
+std::string LengthModFor(unsigned long) { return "l"; }        // NOLINT
+std::string LengthModFor(long long) { return "ll"; }           // NOLINT
+std::string LengthModFor(unsigned long long) { return "ll"; }  // NOLINT
+
+std::string EscCharImpl(int v) {
+  if (std::isprint(static_cast<unsigned char>(v))) {
+    return std::string(1, static_cast<char>(v));
+  }
+  char buf[64];
+  int n = snprintf(buf, sizeof(buf), "\\%#.2x",
+                   static_cast<unsigned>(v & 0xff));
+  assert(n > 0 && n < sizeof(buf));
+  return std::string(buf, n);
+}
+
+std::string Esc(char v) { return EscCharImpl(v); }
+std::string Esc(signed char v) { return EscCharImpl(v); }
+std::string Esc(unsigned char v) { return EscCharImpl(v); }
+
+template <typename T>
+std::string Esc(const T &v) {
+  std::ostringstream oss;
+  oss << v;
+  return oss.str();
+}
+
+void StrAppendV(std::string *dst, const char *format, va_list ap) {
+  // First try with a small fixed size buffer
+  static const int kSpaceLength = 1024;
+  char space[kSpaceLength];
+
+  // It's possible for methods that use a va_list to invalidate
+  // the data in it upon use.  The fix is to make a copy
+  // of the structure before using it and use that copy instead.
+  va_list backup_ap;
+  va_copy(backup_ap, ap);
+  int result = vsnprintf(space, kSpaceLength, format, backup_ap);
+  va_end(backup_ap);
+  if (result < kSpaceLength) {
+    if (result >= 0) {
+      // Normal case -- everything fit.
+      dst->append(space, result);
+      return;
+    }
+    if (result < 0) {
+      // Just an error.
+      return;
+    }
+  }
+
+  // Increase the buffer size to the size requested by vsnprintf,
+  // plus one for the closing \0.
+  int length = result + 1;
+  char *buf = new char[length];
+
+  // Restore the va_list before we use it again
+  va_copy(backup_ap, ap);
+  result = vsnprintf(buf, length, format, backup_ap);
+  va_end(backup_ap);
+
+  if (result >= 0 && result < length) {
+    // It fit
+    dst->append(buf, result);
+  }
+  delete[] buf;
+}
+
+void StrAppend(std::string *, const char *, ...) ABSL_PRINTF_ATTRIBUTE(2, 3);
+void StrAppend(std::string *out, const char *format, ...) {
+  va_list ap;
+  va_start(ap, format);
+  StrAppendV(out, format, ap);
+  va_end(ap);
+}
+
+std::string StrPrint(const char *, ...) ABSL_PRINTF_ATTRIBUTE(1, 2);
+std::string StrPrint(const char *format, ...) {
+  va_list ap;
+  va_start(ap, format);
+  std::string result;
+  StrAppendV(&result, format, ap);
+  va_end(ap);
+  return result;
+}
+
+NativePrintfTraits VerifyNativeImplementationImpl() {
+  NativePrintfTraits result;
+
+  // >>> hex_float_has_glibc_rounding. To have glibc's rounding behavior we need
+  // to meet three requirements:
+  //
+  //   - The threshold for rounding up is 8 (for e.g. MSVC uses 9).
+  //   - If the digits lower than than the 8 are non-zero then we round up.
+  //   - If the digits lower than the 8 are all zero then we round toward even.
+  //
+  // The numbers below represent all the cases covering {below,at,above} the
+  // threshold (8) with both {zero,non-zero} lower bits and both {even,odd}
+  // preceding digits.
+  const double d0079 = 65657.0;  // 0x1.0079p+16
+  const double d0179 = 65913.0;  // 0x1.0179p+16
+  const double d0080 = 65664.0;  // 0x1.0080p+16
+  const double d0180 = 65920.0;  // 0x1.0180p+16
+  const double d0081 = 65665.0;  // 0x1.0081p+16
+  const double d0181 = 65921.0;  // 0x1.0181p+16
+  result.hex_float_has_glibc_rounding =
+      StartsWith(StrPrint("%.2a", d0079), "0x1.00") &&
+      StartsWith(StrPrint("%.2a", d0179), "0x1.01") &&
+      StartsWith(StrPrint("%.2a", d0080), "0x1.00") &&
+      StartsWith(StrPrint("%.2a", d0180), "0x1.02") &&
+      StartsWith(StrPrint("%.2a", d0081), "0x1.01") &&
+      StartsWith(StrPrint("%.2a", d0181), "0x1.02");
+
+  // >>> hex_float_prefers_denormal_repr. Formatting `denormal` on glibc yields
+  // "0x0.0000000000001p-1022", whereas on std libs that don't use denormal
+  // representation it would either be 0x1p-1074 or 0x1.0000000000000-1074.
+  const double denormal = std::numeric_limits<double>::denorm_min();
+  result.hex_float_prefers_denormal_repr =
+      StartsWith(StrPrint("%a", denormal), "0x0.0000000000001");
+
+  // >>> hex_float_uses_minimal_precision_when_not_specified. Some (non-glibc)
+  // libs will format the following as "0x1.0079000000000p+16".
+  result.hex_float_uses_minimal_precision_when_not_specified =
+      (StrPrint("%a", d0079) == "0x1.0079p+16");
+
+  // >>> hex_float_optimizes_leading_digit_bit_count. The number 1.5, when
+  // formatted by glibc should yield "0x1.8p+0" for `double` and "0xcp-3" for
+  // `long double`, i.e., number of bits in the leading digit is adapted to the
+  // number of bits in the mantissa.
+  const double d_15 = 1.5;
+  const long double ld_15 = 1.5;
+  result.hex_float_optimizes_leading_digit_bit_count =
+      StartsWith(StrPrint("%a", d_15), "0x1.8") &&
+      StartsWith(StrPrint("%La", ld_15), "0xc");
+
+  return result;
+}
+
+const NativePrintfTraits &VerifyNativeImplementation() {
+  static NativePrintfTraits native_traits = VerifyNativeImplementationImpl();
+  return native_traits;
+}
+
+class FormatConvertTest : public ::testing::Test { };
+
+template <typename T>
+void TestStringConvert(const T& str) {
+  const FormatArgImpl args[] = {FormatArgImpl(str)};
+  struct Expectation {
+    const char *out;
+    const char *fmt;
+  };
+  const Expectation kExpect[] = {
+    {"hello",  "%1$s"      },
+    {"",       "%1$.s"     },
+    {"",       "%1$.0s"    },
+    {"h",      "%1$.1s"    },
+    {"he",     "%1$.2s"    },
+    {"hello",  "%1$.10s"   },
+    {" hello", "%1$6s"     },
+    {"   he",  "%1$5.2s"   },
+    {"he   ",  "%1$-5.2s"  },
+    {"hello ", "%1$-6.10s" },
+  };
+  for (const Expectation &e : kExpect) {
+    UntypedFormatSpecImpl format(e.fmt);
+    EXPECT_EQ(e.out, FormatPack(format, absl::MakeSpan(args)));
+  }
+}
+
+TEST_F(FormatConvertTest, BasicString) {
+  TestStringConvert("hello");  // As char array.
+  TestStringConvert(static_cast<const char*>("hello"));
+  TestStringConvert(std::string("hello"));
+  TestStringConvert(string_view("hello"));
+#if defined(ABSL_HAVE_STD_STRING_VIEW)
+  TestStringConvert(std::string_view("hello"));
+#endif  // ABSL_HAVE_STD_STRING_VIEW
+}
+
+TEST_F(FormatConvertTest, NullString) {
+  const char* p = nullptr;
+  UntypedFormatSpecImpl format("%s");
+  EXPECT_EQ("", FormatPack(format, {FormatArgImpl(p)}));
+}
+
+TEST_F(FormatConvertTest, StringPrecision) {
+  // We cap at the precision.
+  char c = 'a';
+  const char* p = &c;
+  UntypedFormatSpecImpl format("%.1s");
+  EXPECT_EQ("a", FormatPack(format, {FormatArgImpl(p)}));
+
+  // We cap at the NUL-terminator.
+  p = "ABC";
+  UntypedFormatSpecImpl format2("%.10s");
+  EXPECT_EQ("ABC", FormatPack(format2, {FormatArgImpl(p)}));
+}
+
+// Pointer formatting is implementation defined. This checks that the argument
+// can be matched to `ptr`.
+MATCHER_P(MatchesPointerString, ptr, "") {
+  if (ptr == nullptr && arg == "(nil)") {
+    return true;
+  }
+  void* parsed = nullptr;
+  if (sscanf(arg.c_str(), "%p", &parsed) != 1) {
+    ABSL_RAW_LOG(FATAL, "Could not parse %s", arg.c_str());
+  }
+  return ptr == parsed;
+}
+
+TEST_F(FormatConvertTest, Pointer) {
+  static int x = 0;
+  const int *xp = &x;
+  char c = 'h';
+  char *mcp = &c;
+  const char *cp = "hi";
+  const char *cnil = nullptr;
+  const int *inil = nullptr;
+  using VoidF = void (*)();
+  VoidF fp = [] {}, fnil = nullptr;
+  volatile char vc;
+  volatile char *vcp = &vc;
+  volatile char *vcnil = nullptr;
+  const FormatArgImpl args_array[] = {
+      FormatArgImpl(xp),   FormatArgImpl(cp),  FormatArgImpl(inil),
+      FormatArgImpl(cnil), FormatArgImpl(mcp), FormatArgImpl(fp),
+      FormatArgImpl(fnil), FormatArgImpl(vcp), FormatArgImpl(vcnil),
+  };
+  auto args = absl::MakeConstSpan(args_array);
+
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%p"), args),
+              MatchesPointerString(&x));
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%20p"), args),
+              MatchesPointerString(&x));
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%.1p"), args),
+              MatchesPointerString(&x));
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%.20p"), args),
+              MatchesPointerString(&x));
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%30.20p"), args),
+              MatchesPointerString(&x));
+
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%-p"), args),
+              MatchesPointerString(&x));
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%-20p"), args),
+              MatchesPointerString(&x));
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%-.1p"), args),
+              MatchesPointerString(&x));
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%.20p"), args),
+              MatchesPointerString(&x));
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%-30.20p"), args),
+              MatchesPointerString(&x));
+
+  // const char*
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%2$p"), args),
+              MatchesPointerString(cp));
+  // null const int*
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%3$p"), args),
+              MatchesPointerString(nullptr));
+  // null const char*
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%4$p"), args),
+              MatchesPointerString(nullptr));
+  // nonconst char*
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%5$p"), args),
+              MatchesPointerString(mcp));
+
+  // function pointers
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%6$p"), args),
+              MatchesPointerString(reinterpret_cast<const void*>(fp)));
+  EXPECT_THAT(
+      FormatPack(UntypedFormatSpecImpl("%8$p"), args),
+      MatchesPointerString(reinterpret_cast<volatile const void *>(vcp)));
+
+  // null function pointers
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%7$p"), args),
+              MatchesPointerString(nullptr));
+  EXPECT_THAT(FormatPack(UntypedFormatSpecImpl("%9$p"), args),
+              MatchesPointerString(nullptr));
+}
+
+struct Cardinal {
+  enum Pos { k1 = 1, k2 = 2, k3 = 3 };
+  enum Neg { kM1 = -1, kM2 = -2, kM3 = -3 };
+};
+
+TEST_F(FormatConvertTest, Enum) {
+  const Cardinal::Pos k3 = Cardinal::k3;
+  const Cardinal::Neg km3 = Cardinal::kM3;
+  const FormatArgImpl args[] = {FormatArgImpl(k3), FormatArgImpl(km3)};
+  UntypedFormatSpecImpl format("%1$d");
+  UntypedFormatSpecImpl format2("%2$d");
+  EXPECT_EQ("3", FormatPack(format, absl::MakeSpan(args)));
+  EXPECT_EQ("-3", FormatPack(format2, absl::MakeSpan(args)));
+}
+
+template <typename T>
+class TypedFormatConvertTest : public FormatConvertTest { };
+
+TYPED_TEST_SUITE_P(TypedFormatConvertTest);
+
+std::vector<std::string> AllFlagCombinations() {
+  const char kFlags[] = {'-', '#', '0', '+', ' '};
+  std::vector<std::string> result;
+  for (size_t fsi = 0; fsi < (1ull << ArraySize(kFlags)); ++fsi) {
+    std::string flag_set;
+    for (size_t fi = 0; fi < ArraySize(kFlags); ++fi)
+      if (fsi & (1ull << fi))
+        flag_set += kFlags[fi];
+    result.push_back(flag_set);
+  }
+  return result;
+}
+
+TYPED_TEST_P(TypedFormatConvertTest, AllIntsWithFlags) {
+  typedef TypeParam T;
+  typedef typename std::make_unsigned<T>::type UnsignedT;
+  using remove_volatile_t = typename std::remove_volatile<T>::type;
+  const T kMin = std::numeric_limits<remove_volatile_t>::min();
+  const T kMax = std::numeric_limits<remove_volatile_t>::max();
+  const T kVals[] = {
+      remove_volatile_t(1),
+      remove_volatile_t(2),
+      remove_volatile_t(3),
+      remove_volatile_t(123),
+      remove_volatile_t(-1),
+      remove_volatile_t(-2),
+      remove_volatile_t(-3),
+      remove_volatile_t(-123),
+      remove_volatile_t(0),
+      kMax - remove_volatile_t(1),
+      kMax,
+      kMin + remove_volatile_t(1),
+      kMin,
+  };
+  const char kConvChars[] = {'d', 'i', 'u', 'o', 'x', 'X'};
+  const std::string kWid[] = {"", "4", "10"};
+  const std::string kPrec[] = {"", ".", ".0", ".4", ".10"};
+
+  const std::vector<std::string> flag_sets = AllFlagCombinations();
+
+  for (size_t vi = 0; vi < ArraySize(kVals); ++vi) {
+    const T val = kVals[vi];
+    SCOPED_TRACE(Esc(val));
+    const FormatArgImpl args[] = {FormatArgImpl(val)};
+    for (size_t ci = 0; ci < ArraySize(kConvChars); ++ci) {
+      const char conv_char = kConvChars[ci];
+      for (size_t fsi = 0; fsi < flag_sets.size(); ++fsi) {
+        const std::string &flag_set = flag_sets[fsi];
+        for (size_t wi = 0; wi < ArraySize(kWid); ++wi) {
+          const std::string &wid = kWid[wi];
+          for (size_t pi = 0; pi < ArraySize(kPrec); ++pi) {
+            const std::string &prec = kPrec[pi];
+
+            const bool is_signed_conv = (conv_char == 'd' || conv_char == 'i');
+            const bool is_unsigned_to_signed =
+                !std::is_signed<T>::value && is_signed_conv;
+            // Don't consider sign-related flags '+' and ' ' when doing
+            // unsigned to signed conversions.
+            if (is_unsigned_to_signed &&
+                flag_set.find_first_of("+ ") != std::string::npos) {
+              continue;
+            }
+
+            std::string new_fmt("%");
+            new_fmt += flag_set;
+            new_fmt += wid;
+            new_fmt += prec;
+            // old and new always agree up to here.
+            std::string old_fmt = new_fmt;
+            new_fmt += conv_char;
+            std::string old_result;
+            if (is_unsigned_to_signed) {
+              // don't expect agreement on unsigned formatted as signed,
+              // as printf can't do that conversion properly. For those
+              // cases, we do expect agreement with printf with a "%u"
+              // and the unsigned equivalent of 'val'.
+              UnsignedT uval = val;
+              old_fmt += LengthModFor(uval);
+              old_fmt += "u";
+              old_result = StrPrint(old_fmt.c_str(), uval);
+            } else {
+              old_fmt += LengthModFor(val);
+              old_fmt += conv_char;
+              old_result = StrPrint(old_fmt.c_str(), val);
+            }
+
+            SCOPED_TRACE(std::string() + " old_fmt: \"" + old_fmt +
+                         "\"'"
+                         " new_fmt: \"" +
+                         new_fmt + "\"");
+            UntypedFormatSpecImpl format(new_fmt);
+            EXPECT_EQ(old_result, FormatPack(format, absl::MakeSpan(args)));
+          }
+        }
+      }
+    }
+  }
+}
+
+TYPED_TEST_P(TypedFormatConvertTest, Char) {
+  // Pass a bunch of values of type TypeParam to both FormatPack and libc's
+  // vsnprintf("%c", ...) (wrapped in StrPrint) to make sure we get the same
+  // value.
+  typedef TypeParam T;
+  using remove_volatile_t = typename std::remove_volatile<T>::type;
+  std::vector<remove_volatile_t> vals = {
+      remove_volatile_t(1),  remove_volatile_t(2),  remove_volatile_t(10),   //
+      remove_volatile_t(-1), remove_volatile_t(-2), remove_volatile_t(-10),  //
+      remove_volatile_t(0),
+  };
+
+  // We'd like to test values near std::numeric_limits::min() and
+  // std::numeric_limits::max(), too, but vsnprintf("%c", ...) can't handle
+  // anything larger than an int. Add in the most extreme values we can without
+  // exceeding that range.
+  static const T kMin =
+      static_cast<remove_volatile_t>(std::numeric_limits<int>::min());
+  static const T kMax =
+      static_cast<remove_volatile_t>(std::numeric_limits<int>::max());
+  vals.insert(vals.end(), {kMin + 1, kMin, kMax - 1, kMax});
+
+  for (const T c : vals) {
+    const FormatArgImpl args[] = {FormatArgImpl(c)};
+    UntypedFormatSpecImpl format("%c");
+    EXPECT_EQ(StrPrint("%c", static_cast<int>(c)),
+              FormatPack(format, absl::MakeSpan(args)));
+  }
+}
+
+REGISTER_TYPED_TEST_SUITE_P(TypedFormatConvertTest, AllIntsWithFlags, Char);
+
+typedef ::testing::Types<
+    int, unsigned, volatile int,
+    short, unsigned short,
+    long, unsigned long,
+    long long, unsigned long long,
+    signed char, unsigned char, char>
+    AllIntTypes;
+INSTANTIATE_TYPED_TEST_SUITE_P(TypedFormatConvertTestWithAllIntTypes,
+                               TypedFormatConvertTest, AllIntTypes);
+TEST_F(FormatConvertTest, VectorBool) {
+  // Make sure vector<bool>'s values behave as bools.
+  std::vector<bool> v = {true, false};
+  const std::vector<bool> cv = {true, false};
+  EXPECT_EQ("1,0,1,0",
+            FormatPack(UntypedFormatSpecImpl("%d,%d,%d,%d"),
+                       absl::Span<const FormatArgImpl>(
+                           {FormatArgImpl(v[0]), FormatArgImpl(v[1]),
+                            FormatArgImpl(cv[0]), FormatArgImpl(cv[1])})));
+}
+
+
+TEST_F(FormatConvertTest, Int128) {
+  absl::int128 positive = static_cast<absl::int128>(0x1234567890abcdef) * 1979;
+  absl::int128 negative = -positive;
+  absl::int128 max = absl::Int128Max(), min = absl::Int128Min();
+  const FormatArgImpl args[] = {FormatArgImpl(positive),
+                                FormatArgImpl(negative), FormatArgImpl(max),
+                                FormatArgImpl(min)};
+
+  struct Case {
+    const char* format;
+    const char* expected;
+  } cases[] = {
+      {"%1$d", "2595989796776606496405"},
+      {"%1$30d", "        2595989796776606496405"},
+      {"%1$-30d", "2595989796776606496405        "},
+      {"%1$u", "2595989796776606496405"},
+      {"%1$x", "8cba9876066020f695"},
+      {"%2$d", "-2595989796776606496405"},
+      {"%2$30d", "       -2595989796776606496405"},
+      {"%2$-30d", "-2595989796776606496405       "},
+      {"%2$u", "340282366920938460867384810655161715051"},
+      {"%2$x", "ffffffffffffff73456789f99fdf096b"},
+      {"%3$d", "170141183460469231731687303715884105727"},
+      {"%3$u", "170141183460469231731687303715884105727"},
+      {"%3$x", "7fffffffffffffffffffffffffffffff"},
+      {"%4$d", "-170141183460469231731687303715884105728"},
+      {"%4$x", "80000000000000000000000000000000"},
+  };
+
+  for (auto c : cases) {
+    UntypedFormatSpecImpl format(c.format);
+    EXPECT_EQ(c.expected, FormatPack(format, absl::MakeSpan(args)));
+  }
+}
+
+TEST_F(FormatConvertTest, Uint128) {
+  absl::uint128 v = static_cast<absl::uint128>(0x1234567890abcdef) * 1979;
+  absl::uint128 max = absl::Uint128Max();
+  const FormatArgImpl args[] = {FormatArgImpl(v), FormatArgImpl(max)};
+
+  struct Case {
+    const char* format;
+    const char* expected;
+  } cases[] = {
+      {"%1$d", "2595989796776606496405"},
+      {"%1$30d", "        2595989796776606496405"},
+      {"%1$-30d", "2595989796776606496405        "},
+      {"%1$u", "2595989796776606496405"},
+      {"%1$x", "8cba9876066020f695"},
+      {"%2$d", "340282366920938463463374607431768211455"},
+      {"%2$u", "340282366920938463463374607431768211455"},
+      {"%2$x", "ffffffffffffffffffffffffffffffff"},
+  };
+
+  for (auto c : cases) {
+    UntypedFormatSpecImpl format(c.format);
+    EXPECT_EQ(c.expected, FormatPack(format, absl::MakeSpan(args)));
+  }
+}
+
+template <typename Floating>
+void TestWithMultipleFormatsHelper(const std::vector<Floating> &floats,
+                                   const std::set<Floating> &skip_verify) {
+  const NativePrintfTraits &native_traits = VerifyNativeImplementation();
+  // Reserve the space to ensure we don't allocate memory in the output itself.
+  std::string str_format_result;
+  str_format_result.reserve(1 << 20);
+  std::string string_printf_result;
+  string_printf_result.reserve(1 << 20);
+
+  const char *const kFormats[] = {
+      "%",  "%.3", "%8.5", "%500",   "%.5000", "%.60", "%.30",   "%03",
+      "%+", "% ",  "%-10", "%#15.3", "%#.0",   "%.0",  "%1$*2$", "%1$.*2$"};
+
+  for (const char *fmt : kFormats) {
+    for (char f : {'f', 'F',  //
+                   'g', 'G',  //
+                   'a', 'A',  //
+                   'e', 'E'}) {
+      std::string fmt_str = std::string(fmt) + f;
+
+      if (fmt == absl::string_view("%.5000") && f != 'f' && f != 'F' &&
+          f != 'a' && f != 'A') {
+        // This particular test takes way too long with snprintf.
+        // Disable for the case we are not implementing natively.
+        continue;
+      }
+
+      if ((f == 'a' || f == 'A') &&
+          !native_traits.hex_float_has_glibc_rounding) {
+        continue;
+      }
+
+      for (Floating d : floats) {
+        if (!native_traits.hex_float_prefers_denormal_repr &&
+            (f == 'a' || f == 'A') && std::fpclassify(d) == FP_SUBNORMAL) {
+          continue;
+        }
+        int i = -10;
+        FormatArgImpl args[2] = {FormatArgImpl(d), FormatArgImpl(i)};
+        UntypedFormatSpecImpl format(fmt_str);
+
+        string_printf_result.clear();
+        StrAppend(&string_printf_result, fmt_str.c_str(), d, i);
+        str_format_result.clear();
+
+        {
+          AppendPack(&str_format_result, format, absl::MakeSpan(args));
+        }
+
+#ifdef _MSC_VER
+        // MSVC has a different rounding policy than us so we can't test our
+        // implementation against the native one there.
+        continue;
+#elif defined(__APPLE__)
+        // Apple formats NaN differently (+nan) vs. (nan)
+        if (std::isnan(d)) continue;
+#endif
+        if (string_printf_result != str_format_result &&
+            skip_verify.find(d) == skip_verify.end()) {
+          // We use ASSERT_EQ here because failures are usually correlated and a
+          // bug would print way too many failed expectations causing the test
+          // to time out.
+          ASSERT_EQ(string_printf_result, str_format_result)
+              << fmt_str << " " << StrPrint("%.18g", d) << " "
+              << StrPrint("%a", d) << " " << StrPrint("%.50f", d);
+        }
+      }
+    }
+  }
+}
+
+TEST_F(FormatConvertTest, Float) {
+  std::vector<float> floats = {0.0f,
+                               -0.0f,
+                               .9999999f,
+                               9999999.f,
+                               std::numeric_limits<float>::max(),
+                               -std::numeric_limits<float>::max(),
+                               std::numeric_limits<float>::min(),
+                               -std::numeric_limits<float>::min(),
+                               std::numeric_limits<float>::lowest(),
+                               -std::numeric_limits<float>::lowest(),
+                               std::numeric_limits<float>::epsilon(),
+                               std::numeric_limits<float>::epsilon() + 1.0f,
+                               std::numeric_limits<float>::infinity(),
+                               -std::numeric_limits<float>::infinity(),
+                               std::nanf("")};
+
+  // Some regression tests.
+  floats.push_back(0.999999989f);
+
+  if (std::numeric_limits<float>::has_denorm != std::denorm_absent) {
+    floats.push_back(std::numeric_limits<float>::denorm_min());
+    floats.push_back(-std::numeric_limits<float>::denorm_min());
+  }
+
+  for (float base :
+       {1.f, 12.f, 123.f, 1234.f, 12345.f, 123456.f, 1234567.f, 12345678.f,
+        123456789.f, 1234567890.f, 12345678901.f, 12345678.f, 12345678.f}) {
+    for (int exp = -123; exp <= 123; ++exp) {
+      for (int sign : {1, -1}) {
+        floats.push_back(sign * std::ldexp(base, exp));
+      }
+    }
+  }
+
+  for (int exp = -300; exp <= 300; ++exp) {
+    const float all_ones_mantissa = 0xffffff;
+    floats.push_back(std::ldexp(all_ones_mantissa, exp));
+  }
+
+  // Remove duplicates to speed up the logic below.
+  std::sort(floats.begin(), floats.end());
+  floats.erase(std::unique(floats.begin(), floats.end()), floats.end());
+
+  TestWithMultipleFormatsHelper(floats, {});
+}
+
+TEST_F(FormatConvertTest, Double) {
+  // For values that we know won't match the standard library implementation we
+  // skip verification, but still run the algorithm to catch asserts/sanitizer
+  // bugs.
+  std::set<double> skip_verify;
+  std::vector<double> doubles = {0.0,
+                                 -0.0,
+                                 .99999999999999,
+                                 99999999999999.,
+                                 std::numeric_limits<double>::max(),
+                                 -std::numeric_limits<double>::max(),
+                                 std::numeric_limits<double>::min(),
+                                 -std::numeric_limits<double>::min(),
+                                 std::numeric_limits<double>::lowest(),
+                                 -std::numeric_limits<double>::lowest(),
+                                 std::numeric_limits<double>::epsilon(),
+                                 std::numeric_limits<double>::epsilon() + 1,
+                                 std::numeric_limits<double>::infinity(),
+                                 -std::numeric_limits<double>::infinity(),
+                                 std::nan("")};
+
+  // Some regression tests.
+  doubles.push_back(0.99999999999999989);
+
+  if (std::numeric_limits<double>::has_denorm != std::denorm_absent) {
+    doubles.push_back(std::numeric_limits<double>::denorm_min());
+    doubles.push_back(-std::numeric_limits<double>::denorm_min());
+  }
+
+  for (double base :
+       {1., 12., 123., 1234., 12345., 123456., 1234567., 12345678., 123456789.,
+        1234567890., 12345678901., 123456789012., 1234567890123.}) {
+    for (int exp = -123; exp <= 123; ++exp) {
+      for (int sign : {1, -1}) {
+        doubles.push_back(sign * std::ldexp(base, exp));
+      }
+    }
+  }
+
+  // Workaround libc bug.
+  // https://sourceware.org/bugzilla/show_bug.cgi?id=22142
+  const bool gcc_bug_22142 =
+      StrPrint("%f", std::numeric_limits<double>::max()) !=
+      "1797693134862315708145274237317043567980705675258449965989174768031"
+      "5726078002853876058955863276687817154045895351438246423432132688946"
+      "4182768467546703537516986049910576551282076245490090389328944075868"
+      "5084551339423045832369032229481658085593321233482747978262041447231"
+      "68738177180919299881250404026184124858368.000000";
+
+  for (int exp = -300; exp <= 300; ++exp) {
+    const double all_ones_mantissa = 0x1fffffffffffff;
+    doubles.push_back(std::ldexp(all_ones_mantissa, exp));
+    if (gcc_bug_22142) {
+      skip_verify.insert(doubles.back());
+    }
+  }
+
+  if (gcc_bug_22142) {
+    using L = std::numeric_limits<double>;
+    skip_verify.insert(L::max());
+    skip_verify.insert(L::min());  // NOLINT
+    skip_verify.insert(L::denorm_min());
+    skip_verify.insert(-L::max());
+    skip_verify.insert(-L::min());  // NOLINT
+    skip_verify.insert(-L::denorm_min());
+  }
+
+  // Remove duplicates to speed up the logic below.
+  std::sort(doubles.begin(), doubles.end());
+  doubles.erase(std::unique(doubles.begin(), doubles.end()), doubles.end());
+
+  TestWithMultipleFormatsHelper(doubles, skip_verify);
+}
+
+TEST_F(FormatConvertTest, DoubleRound) {
+  std::string s;
+  const auto format = [&](const char *fmt, double d) -> std::string & {
+    s.clear();
+    FormatArgImpl args[1] = {FormatArgImpl(d)};
+    AppendPack(&s, UntypedFormatSpecImpl(fmt), absl::MakeSpan(args));
+#if !defined(_MSC_VER)
+    // MSVC has a different rounding policy than us so we can't test our
+    // implementation against the native one there.
+    EXPECT_EQ(StrPrint(fmt, d), s);
+#endif  // _MSC_VER
+
+    return s;
+  };
+  // All of these values have to be exactly represented.
+  // Otherwise we might not be testing what we think we are testing.
+
+  // These values can fit in a 64bit "fast" representation.
+  const double exact_value = 0.00000000000005684341886080801486968994140625;
+  assert(exact_value == std::pow(2, -44));
+  // Round up at a 5xx.
+  EXPECT_EQ(format("%.13f", exact_value), "0.0000000000001");
+  // Round up at a >5
+  EXPECT_EQ(format("%.14f", exact_value), "0.00000000000006");
+  // Round down at a <5
+  EXPECT_EQ(format("%.16f", exact_value), "0.0000000000000568");
+  // Nine handling
+  EXPECT_EQ(format("%.35f", exact_value),
+            "0.00000000000005684341886080801486969");
+  EXPECT_EQ(format("%.36f", exact_value),
+            "0.000000000000056843418860808014869690");
+  // Round down the last nine.
+  EXPECT_EQ(format("%.37f", exact_value),
+            "0.0000000000000568434188608080148696899");
+  EXPECT_EQ(format("%.10f", 0.000003814697265625), "0.0000038147");
+  // Round up the last nine
+  EXPECT_EQ(format("%.11f", 0.000003814697265625), "0.00000381470");
+  EXPECT_EQ(format("%.12f", 0.000003814697265625), "0.000003814697");
+
+  // Round to even (down)
+  EXPECT_EQ(format("%.43f", exact_value),
+            "0.0000000000000568434188608080148696899414062");
+  // Exact
+  EXPECT_EQ(format("%.44f", exact_value),
+            "0.00000000000005684341886080801486968994140625");
+  // Round to even (up), let make the last digits 75 instead of 25
+  EXPECT_EQ(format("%.43f", exact_value + std::pow(2, -43)),
+            "0.0000000000001705302565824240446090698242188");
+  // Exact, just to check.
+  EXPECT_EQ(format("%.44f", exact_value + std::pow(2, -43)),
+            "0.00000000000017053025658242404460906982421875");
+
+  // This value has to be small enough that it won't fit in the uint128
+  // representation for printing.
+  const double small_exact_value =
+      0.000000000000000000000000000000000000752316384526264005099991383822237233803945956334136013765601092018187046051025390625;  // NOLINT
+  assert(small_exact_value == std::pow(2, -120));
+  // Round up at a 5xx.
+  EXPECT_EQ(format("%.37f", small_exact_value),
+            "0.0000000000000000000000000000000000008");
+  // Round down at a <5
+  EXPECT_EQ(format("%.38f", small_exact_value),
+            "0.00000000000000000000000000000000000075");
+  // Round up at a >5
+  EXPECT_EQ(format("%.41f", small_exact_value),
+            "0.00000000000000000000000000000000000075232");
+  // Nine handling
+  EXPECT_EQ(format("%.55f", small_exact_value),
+            "0.0000000000000000000000000000000000007523163845262640051");
+  EXPECT_EQ(format("%.56f", small_exact_value),
+            "0.00000000000000000000000000000000000075231638452626400510");
+  EXPECT_EQ(format("%.57f", small_exact_value),
+            "0.000000000000000000000000000000000000752316384526264005100");
+  EXPECT_EQ(format("%.58f", small_exact_value),
+            "0.0000000000000000000000000000000000007523163845262640051000");
+  // Round down the last nine
+  EXPECT_EQ(format("%.59f", small_exact_value),
+            "0.00000000000000000000000000000000000075231638452626400509999");
+  // Round up the last nine
+  EXPECT_EQ(format("%.79f", small_exact_value),
+            "0.000000000000000000000000000000000000"
+            "7523163845262640050999913838222372338039460");
+
+  // Round to even (down)
+  EXPECT_EQ(format("%.119f", small_exact_value),
+            "0.000000000000000000000000000000000000"
+            "75231638452626400509999138382223723380"
+            "394595633413601376560109201818704605102539062");
+  // Exact
+  EXPECT_EQ(format("%.120f", small_exact_value),
+            "0.000000000000000000000000000000000000"
+            "75231638452626400509999138382223723380"
+            "3945956334136013765601092018187046051025390625");
+  // Round to even (up), let make the last digits 75 instead of 25
+  EXPECT_EQ(format("%.119f", small_exact_value + std::pow(2, -119)),
+            "0.000000000000000000000000000000000002"
+            "25694915357879201529997415146671170141"
+            "183786900240804129680327605456113815307617188");
+  // Exact, just to check.
+  EXPECT_EQ(format("%.120f", small_exact_value + std::pow(2, -119)),
+            "0.000000000000000000000000000000000002"
+            "25694915357879201529997415146671170141"
+            "1837869002408041296803276054561138153076171875");
+}
+
+TEST_F(FormatConvertTest, DoubleRoundA) {
+  const NativePrintfTraits &native_traits = VerifyNativeImplementation();
+  std::string s;
+  const auto format = [&](const char *fmt, double d) -> std::string & {
+    s.clear();
+    FormatArgImpl args[1] = {FormatArgImpl(d)};
+    AppendPack(&s, UntypedFormatSpecImpl(fmt), absl::MakeSpan(args));
+    if (native_traits.hex_float_has_glibc_rounding) {
+      EXPECT_EQ(StrPrint(fmt, d), s);
+    }
+    return s;
+  };
+
+  // 0x1.00018000p+100
+  const double on_boundary_odd = 1267679614447900152596896153600.0;
+  EXPECT_EQ(format("%.0a", on_boundary_odd), "0x1p+100");
+  EXPECT_EQ(format("%.1a", on_boundary_odd), "0x1.0p+100");
+  EXPECT_EQ(format("%.2a", on_boundary_odd), "0x1.00p+100");
+  EXPECT_EQ(format("%.3a", on_boundary_odd), "0x1.000p+100");
+  EXPECT_EQ(format("%.4a", on_boundary_odd), "0x1.0002p+100");  // round
+  EXPECT_EQ(format("%.5a", on_boundary_odd), "0x1.00018p+100");
+  EXPECT_EQ(format("%.6a", on_boundary_odd), "0x1.000180p+100");
+
+  // 0x1.00028000p-2
+  const double on_boundary_even = 0.250009536743164062500;
+  EXPECT_EQ(format("%.0a", on_boundary_even), "0x1p-2");
+  EXPECT_EQ(format("%.1a", on_boundary_even), "0x1.0p-2");
+  EXPECT_EQ(format("%.2a", on_boundary_even), "0x1.00p-2");
+  EXPECT_EQ(format("%.3a", on_boundary_even), "0x1.000p-2");
+  EXPECT_EQ(format("%.4a", on_boundary_even), "0x1.0002p-2");  // no round
+  EXPECT_EQ(format("%.5a", on_boundary_even), "0x1.00028p-2");
+  EXPECT_EQ(format("%.6a", on_boundary_even), "0x1.000280p-2");
+
+  // 0x1.00018001p+1
+  const double slightly_over = 2.00004577683284878730773925781250;
+  EXPECT_EQ(format("%.0a", slightly_over), "0x1p+1");
+  EXPECT_EQ(format("%.1a", slightly_over), "0x1.0p+1");
+  EXPECT_EQ(format("%.2a", slightly_over), "0x1.00p+1");
+  EXPECT_EQ(format("%.3a", slightly_over), "0x1.000p+1");
+  EXPECT_EQ(format("%.4a", slightly_over), "0x1.0002p+1");
+  EXPECT_EQ(format("%.5a", slightly_over), "0x1.00018p+1");
+  EXPECT_EQ(format("%.6a", slightly_over), "0x1.000180p+1");
+
+  // 0x1.00017fffp+0
+  const double slightly_under = 1.000022887950763106346130371093750;
+  EXPECT_EQ(format("%.0a", slightly_under), "0x1p+0");
+  EXPECT_EQ(format("%.1a", slightly_under), "0x1.0p+0");
+  EXPECT_EQ(format("%.2a", slightly_under), "0x1.00p+0");
+  EXPECT_EQ(format("%.3a", slightly_under), "0x1.000p+0");
+  EXPECT_EQ(format("%.4a", slightly_under), "0x1.0001p+0");
+  EXPECT_EQ(format("%.5a", slightly_under), "0x1.00018p+0");
+  EXPECT_EQ(format("%.6a", slightly_under), "0x1.000180p+0");
+  EXPECT_EQ(format("%.7a", slightly_under), "0x1.0001800p+0");
+
+  // 0x1.1b3829ac28058p+3
+  const double hex_value = 8.85060580848964661981881363317370414733886718750;
+  EXPECT_EQ(format("%.0a", hex_value), "0x1p+3");
+  EXPECT_EQ(format("%.1a", hex_value), "0x1.2p+3");
+  EXPECT_EQ(format("%.2a", hex_value), "0x1.1bp+3");
+  EXPECT_EQ(format("%.3a", hex_value), "0x1.1b4p+3");
+  EXPECT_EQ(format("%.4a", hex_value), "0x1.1b38p+3");
+  EXPECT_EQ(format("%.5a", hex_value), "0x1.1b383p+3");
+  EXPECT_EQ(format("%.6a", hex_value), "0x1.1b382ap+3");
+  EXPECT_EQ(format("%.7a", hex_value), "0x1.1b3829bp+3");
+  EXPECT_EQ(format("%.8a", hex_value), "0x1.1b3829acp+3");
+  EXPECT_EQ(format("%.9a", hex_value), "0x1.1b3829ac3p+3");
+  EXPECT_EQ(format("%.10a", hex_value), "0x1.1b3829ac28p+3");
+  EXPECT_EQ(format("%.11a", hex_value), "0x1.1b3829ac280p+3");
+  EXPECT_EQ(format("%.12a", hex_value), "0x1.1b3829ac2806p+3");
+  EXPECT_EQ(format("%.13a", hex_value), "0x1.1b3829ac28058p+3");
+  EXPECT_EQ(format("%.14a", hex_value), "0x1.1b3829ac280580p+3");
+  EXPECT_EQ(format("%.15a", hex_value), "0x1.1b3829ac2805800p+3");
+  EXPECT_EQ(format("%.16a", hex_value), "0x1.1b3829ac28058000p+3");
+  EXPECT_EQ(format("%.17a", hex_value), "0x1.1b3829ac280580000p+3");
+  EXPECT_EQ(format("%.18a", hex_value), "0x1.1b3829ac2805800000p+3");
+  EXPECT_EQ(format("%.19a", hex_value), "0x1.1b3829ac28058000000p+3");
+  EXPECT_EQ(format("%.20a", hex_value), "0x1.1b3829ac280580000000p+3");
+  EXPECT_EQ(format("%.21a", hex_value), "0x1.1b3829ac2805800000000p+3");
+
+  // 0x1.0818283848586p+3
+  const double hex_value2 = 8.2529488658208371987257123691961169242858886718750;
+  EXPECT_EQ(format("%.0a", hex_value2), "0x1p+3");
+  EXPECT_EQ(format("%.1a", hex_value2), "0x1.1p+3");
+  EXPECT_EQ(format("%.2a", hex_value2), "0x1.08p+3");
+  EXPECT_EQ(format("%.3a", hex_value2), "0x1.082p+3");
+  EXPECT_EQ(format("%.4a", hex_value2), "0x1.0818p+3");
+  EXPECT_EQ(format("%.5a", hex_value2), "0x1.08183p+3");
+  EXPECT_EQ(format("%.6a", hex_value2), "0x1.081828p+3");
+  EXPECT_EQ(format("%.7a", hex_value2), "0x1.0818284p+3");
+  EXPECT_EQ(format("%.8a", hex_value2), "0x1.08182838p+3");
+  EXPECT_EQ(format("%.9a", hex_value2), "0x1.081828385p+3");
+  EXPECT_EQ(format("%.10a", hex_value2), "0x1.0818283848p+3");
+  EXPECT_EQ(format("%.11a", hex_value2), "0x1.08182838486p+3");
+  EXPECT_EQ(format("%.12a", hex_value2), "0x1.081828384858p+3");
+  EXPECT_EQ(format("%.13a", hex_value2), "0x1.0818283848586p+3");
+  EXPECT_EQ(format("%.14a", hex_value2), "0x1.08182838485860p+3");
+  EXPECT_EQ(format("%.15a", hex_value2), "0x1.081828384858600p+3");
+  EXPECT_EQ(format("%.16a", hex_value2), "0x1.0818283848586000p+3");
+  EXPECT_EQ(format("%.17a", hex_value2), "0x1.08182838485860000p+3");
+  EXPECT_EQ(format("%.18a", hex_value2), "0x1.081828384858600000p+3");
+  EXPECT_EQ(format("%.19a", hex_value2), "0x1.0818283848586000000p+3");
+  EXPECT_EQ(format("%.20a", hex_value2), "0x1.08182838485860000000p+3");
+  EXPECT_EQ(format("%.21a", hex_value2), "0x1.081828384858600000000p+3");
+}
+
+TEST_F(FormatConvertTest, LongDoubleRoundA) {
+  if (std::numeric_limits<long double>::digits % 4 != 0) {
+    // This test doesn't really make sense to run on platforms where a long
+    // double has a different mantissa size (mod 4) than Prod, since then the
+    // leading digit will be formatted differently.
+    return;
+  }
+  const NativePrintfTraits &native_traits = VerifyNativeImplementation();
+  std::string s;
+  const auto format = [&](const char *fmt, long double d) -> std::string & {
+    s.clear();
+    FormatArgImpl args[1] = {FormatArgImpl(d)};
+    AppendPack(&s, UntypedFormatSpecImpl(fmt), absl::MakeSpan(args));
+    if (native_traits.hex_float_has_glibc_rounding &&
+        native_traits.hex_float_optimizes_leading_digit_bit_count) {
+      EXPECT_EQ(StrPrint(fmt, d), s);
+    }
+    return s;
+  };
+
+  // 0x8.8p+4
+  const long double on_boundary_even = 136.0;
+  EXPECT_EQ(format("%.0La", on_boundary_even), "0x8p+4");
+  EXPECT_EQ(format("%.1La", on_boundary_even), "0x8.8p+4");
+  EXPECT_EQ(format("%.2La", on_boundary_even), "0x8.80p+4");
+  EXPECT_EQ(format("%.3La", on_boundary_even), "0x8.800p+4");
+  EXPECT_EQ(format("%.4La", on_boundary_even), "0x8.8000p+4");
+  EXPECT_EQ(format("%.5La", on_boundary_even), "0x8.80000p+4");
+  EXPECT_EQ(format("%.6La", on_boundary_even), "0x8.800000p+4");
+
+  // 0x9.8p+4
+  const long double on_boundary_odd = 152.0;
+  EXPECT_EQ(format("%.0La", on_boundary_odd), "0xap+4");
+  EXPECT_EQ(format("%.1La", on_boundary_odd), "0x9.8p+4");
+  EXPECT_EQ(format("%.2La", on_boundary_odd), "0x9.80p+4");
+  EXPECT_EQ(format("%.3La", on_boundary_odd), "0x9.800p+4");
+  EXPECT_EQ(format("%.4La", on_boundary_odd), "0x9.8000p+4");
+  EXPECT_EQ(format("%.5La", on_boundary_odd), "0x9.80000p+4");
+  EXPECT_EQ(format("%.6La", on_boundary_odd), "0x9.800000p+4");
+
+  // 0x8.80001p+24
+  const long double slightly_over = 142606352.0;
+  EXPECT_EQ(format("%.0La", slightly_over), "0x9p+24");
+  EXPECT_EQ(format("%.1La", slightly_over), "0x8.8p+24");
+  EXPECT_EQ(format("%.2La", slightly_over), "0x8.80p+24");
+  EXPECT_EQ(format("%.3La", slightly_over), "0x8.800p+24");
+  EXPECT_EQ(format("%.4La", slightly_over), "0x8.8000p+24");
+  EXPECT_EQ(format("%.5La", slightly_over), "0x8.80001p+24");
+  EXPECT_EQ(format("%.6La", slightly_over), "0x8.800010p+24");
+
+  // 0x8.7ffffp+24
+  const long double slightly_under = 142606320.0;
+  EXPECT_EQ(format("%.0La", slightly_under), "0x8p+24");
+  EXPECT_EQ(format("%.1La", slightly_under), "0x8.8p+24");
+  EXPECT_EQ(format("%.2La", slightly_under), "0x8.80p+24");
+  EXPECT_EQ(format("%.3La", slightly_under), "0x8.800p+24");
+  EXPECT_EQ(format("%.4La", slightly_under), "0x8.8000p+24");
+  EXPECT_EQ(format("%.5La", slightly_under), "0x8.7ffffp+24");
+  EXPECT_EQ(format("%.6La", slightly_under), "0x8.7ffff0p+24");
+  EXPECT_EQ(format("%.7La", slightly_under), "0x8.7ffff00p+24");
+
+  // 0xc.0828384858688000p+128
+  const long double eights = 4094231060438608800781871108094404067328.0;
+  EXPECT_EQ(format("%.0La", eights), "0xcp+128");
+  EXPECT_EQ(format("%.1La", eights), "0xc.1p+128");
+  EXPECT_EQ(format("%.2La", eights), "0xc.08p+128");
+  EXPECT_EQ(format("%.3La", eights), "0xc.083p+128");
+  EXPECT_EQ(format("%.4La", eights), "0xc.0828p+128");
+  EXPECT_EQ(format("%.5La", eights), "0xc.08284p+128");
+  EXPECT_EQ(format("%.6La", eights), "0xc.082838p+128");
+  EXPECT_EQ(format("%.7La", eights), "0xc.0828385p+128");
+  EXPECT_EQ(format("%.8La", eights), "0xc.08283848p+128");
+  EXPECT_EQ(format("%.9La", eights), "0xc.082838486p+128");
+  EXPECT_EQ(format("%.10La", eights), "0xc.0828384858p+128");
+  EXPECT_EQ(format("%.11La", eights), "0xc.08283848587p+128");
+  EXPECT_EQ(format("%.12La", eights), "0xc.082838485868p+128");
+  EXPECT_EQ(format("%.13La", eights), "0xc.0828384858688p+128");
+  EXPECT_EQ(format("%.14La", eights), "0xc.08283848586880p+128");
+  EXPECT_EQ(format("%.15La", eights), "0xc.082838485868800p+128");
+  EXPECT_EQ(format("%.16La", eights), "0xc.0828384858688000p+128");
+}
+
+// We don't actually store the results. This is just to exercise the rest of the
+// machinery.
+struct NullSink {
+  friend void AbslFormatFlush(NullSink *sink, string_view str) {}
+};
+
+template <typename... T>
+bool FormatWithNullSink(absl::string_view fmt, const T &... a) {
+  NullSink sink;
+  FormatArgImpl args[] = {FormatArgImpl(a)...};
+  return FormatUntyped(&sink, UntypedFormatSpecImpl(fmt), absl::MakeSpan(args));
+}
+
+TEST_F(FormatConvertTest, ExtremeWidthPrecision) {
+  for (const char *fmt : {"f"}) {
+    for (double d : {1e-100, 1.0, 1e100}) {
+      constexpr int max = std::numeric_limits<int>::max();
+      EXPECT_TRUE(FormatWithNullSink(std::string("%.*") + fmt, max, d));
+      EXPECT_TRUE(FormatWithNullSink(std::string("%1.*") + fmt, max, d));
+      EXPECT_TRUE(FormatWithNullSink(std::string("%*") + fmt, max, d));
+      EXPECT_TRUE(FormatWithNullSink(std::string("%*.*") + fmt, max, max, d));
+    }
+  }
+}
+
+TEST_F(FormatConvertTest, LongDouble) {
+  const NativePrintfTraits &native_traits = VerifyNativeImplementation();
+  const char *const kFormats[] = {"%",    "%.3", "%8.5", "%9",  "%.5000",
+                                  "%.60", "%+",  "% ",   "%-10"};
+
+  std::vector<long double> doubles = {
+      0.0,
+      -0.0,
+      std::numeric_limits<long double>::max(),
+      -std::numeric_limits<long double>::max(),
+      std::numeric_limits<long double>::min(),
+      -std::numeric_limits<long double>::min(),
+      std::numeric_limits<long double>::infinity(),
+      -std::numeric_limits<long double>::infinity()};
+
+  for (long double base : {1.L, 12.L, 123.L, 1234.L, 12345.L, 123456.L,
+                           1234567.L, 12345678.L, 123456789.L, 1234567890.L,
+                           12345678901.L, 123456789012.L, 1234567890123.L,
+                           // This value is not representable in double, but it
+                           // is in long double that uses the extended format.
+                           // This is to verify that we are not truncating the
+                           // value mistakenly through a double.
+                           10000000000000000.25L}) {
+    for (int exp : {-1000, -500, 0, 500, 1000}) {
+      for (int sign : {1, -1}) {
+        doubles.push_back(sign * std::ldexp(base, exp));
+        doubles.push_back(sign / std::ldexp(base, exp));
+      }
+    }
+  }
+
+  // Regression tests
+  //
+  // Using a string literal because not all platforms support hex literals or it
+  // might be out of range.
+  doubles.push_back(std::strtold("-0xf.ffffffb5feafffbp-16324L", nullptr));
+
+  for (const char *fmt : kFormats) {
+    for (char f : {'f', 'F',  //
+                   'g', 'G',  //
+                   'a', 'A',  //
+                   'e', 'E'}) {
+      std::string fmt_str = std::string(fmt) + 'L' + f;
+
+      if (fmt == absl::string_view("%.5000") && f != 'f' && f != 'F' &&
+          f != 'a' && f != 'A') {
+        // This particular test takes way too long with snprintf.
+        // Disable for the case we are not implementing natively.
+        continue;
+      }
+
+      if (f == 'a' || f == 'A') {
+        if (!native_traits.hex_float_has_glibc_rounding ||
+            !native_traits.hex_float_optimizes_leading_digit_bit_count) {
+          continue;
+        }
+      }
+
+      for (auto d : doubles) {
+        FormatArgImpl arg(d);
+        UntypedFormatSpecImpl format(fmt_str);
+        std::string result = FormatPack(format, {&arg, 1});
+
+#ifdef _MSC_VER
+        // MSVC has a different rounding policy than us so we can't test our
+        // implementation against the native one there.
+        continue;
+#endif  // _MSC_VER
+
+        // We use ASSERT_EQ here because failures are usually correlated and a
+        // bug would print way too many failed expectations causing the test to
+        // time out.
+        ASSERT_EQ(StrPrint(fmt_str.c_str(), d), result)
+            << fmt_str << " " << StrPrint("%.18Lg", d) << " "
+            << StrPrint("%La", d) << " " << StrPrint("%.1080Lf", d);
+      }
+    }
+  }
+}
+
+TEST_F(FormatConvertTest, IntAsDouble) {
+  const NativePrintfTraits &native_traits = VerifyNativeImplementation();
+  const int kMin = std::numeric_limits<int>::min();
+  const int kMax = std::numeric_limits<int>::max();
+  const int ia[] = {
+    1, 2, 3, 123,
+    -1, -2, -3, -123,
+    0, kMax - 1, kMax, kMin + 1, kMin };
+  for (const int fx : ia) {
+    SCOPED_TRACE(fx);
+    const FormatArgImpl args[] = {FormatArgImpl(fx)};
+    struct Expectation {
+      int line;
+      std::string out;
+      const char *fmt;
+    };
+    const double dx = static_cast<double>(fx);
+    std::vector<Expectation> expect = {
+        {__LINE__, StrPrint("%f", dx), "%f"},
+        {__LINE__, StrPrint("%12f", dx), "%12f"},
+        {__LINE__, StrPrint("%.12f", dx), "%.12f"},
+        {__LINE__, StrPrint("%.12a", dx), "%.12a"},
+    };
+    if (native_traits.hex_float_uses_minimal_precision_when_not_specified) {
+      Expectation ex = {__LINE__, StrPrint("%12a", dx), "%12a"};
+      expect.push_back(ex);
+    }
+    for (const Expectation &e : expect) {
+      SCOPED_TRACE(e.line);
+      SCOPED_TRACE(e.fmt);
+      UntypedFormatSpecImpl format(e.fmt);
+      EXPECT_EQ(e.out, FormatPack(format, absl::MakeSpan(args)));
+    }
+  }
+}
+
+template <typename T>
+bool FormatFails(const char* test_format, T value) {
+  std::string format_string = std::string("<<") + test_format + ">>";
+  UntypedFormatSpecImpl format(format_string);
+
+  int one = 1;
+  const FormatArgImpl args[] = {FormatArgImpl(value), FormatArgImpl(one)};
+  EXPECT_EQ(FormatPack(format, absl::MakeSpan(args)), "")
+      << "format=" << test_format << " value=" << value;
+  return FormatPack(format, absl::MakeSpan(args)).empty();
+}
+
+TEST_F(FormatConvertTest, ExpectedFailures) {
+  // Int input
+  EXPECT_TRUE(FormatFails("%p", 1));
+  EXPECT_TRUE(FormatFails("%s", 1));
+  EXPECT_TRUE(FormatFails("%n", 1));
+
+  // Double input
+  EXPECT_TRUE(FormatFails("%p", 1.));
+  EXPECT_TRUE(FormatFails("%s", 1.));
+  EXPECT_TRUE(FormatFails("%n", 1.));
+  EXPECT_TRUE(FormatFails("%c", 1.));
+  EXPECT_TRUE(FormatFails("%d", 1.));
+  EXPECT_TRUE(FormatFails("%x", 1.));
+  EXPECT_TRUE(FormatFails("%*d", 1.));
+
+  // String input
+  EXPECT_TRUE(FormatFails("%n", ""));
+  EXPECT_TRUE(FormatFails("%c", ""));
+  EXPECT_TRUE(FormatFails("%d", ""));
+  EXPECT_TRUE(FormatFails("%x", ""));
+  EXPECT_TRUE(FormatFails("%f", ""));
+  EXPECT_TRUE(FormatFails("%*d", ""));
+}
+
+// Sanity check to make sure that we are testing what we think we're testing on
+// e.g. the x86_64+glibc platform.
+TEST_F(FormatConvertTest, GlibcHasCorrectTraits) {
+#if !defined(__GLIBC__) || !defined(__x86_64__)
+  return;
+#endif
+  const NativePrintfTraits &native_traits = VerifyNativeImplementation();
+  // If one of the following tests break then it is either because the above PP
+  // macro guards failed to exclude a new platform (likely) or because something
+  // has changed in the implemention of glibc sprintf float formatting behavior.
+  // If the latter, then the code that computes these flags needs to be
+  // revisited and/or possibly the StrFormat implementation.
+  EXPECT_TRUE(native_traits.hex_float_has_glibc_rounding);
+  EXPECT_TRUE(native_traits.hex_float_prefers_denormal_repr);
+  EXPECT_TRUE(
+      native_traits.hex_float_uses_minimal_precision_when_not_specified);
+  EXPECT_TRUE(native_traits.hex_float_optimizes_leading_digit_bit_count);
+}
+
+}  // namespace
+}  // namespace str_format_internal
+ABSL_NAMESPACE_END
+}  // namespace absl