[clang] [clang-format] Add an option to format integer and float literal case (PR #151590)

[Andy MacGregor via cfe-commits]

================
@@ -0,0 +1,368 @@
+//===--- NumericLiteralCaseFixer.cpp -----------------------*- C++ -*-===//
+//
+// Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
+// See https://llvm.org/LICENSE.txt for license information.
+// SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
+//
+//===----------------------------------------------------------------------===//
+///
+/// \file
+/// This file implements NumericLiteralCaseFixer that standardizes character
+/// case within numeric literal constants.
+///
+//===----------------------------------------------------------------------===//
+
+#include "NumericLiteralCaseFixer.h"
+
+#include "llvm/ADT/StringExtras.h"
+
+#include <algorithm>
+
+namespace clang {
+namespace format {
+
+using CharTransformFn = char (*)(char C);
+namespace {
+
+/// @brief Collection of std::transform predicates for each part of a numeric
+/// literal
+struct FormatParameters {
+  FormatParameters(FormatStyle::LanguageKind Language,
+                   const FormatStyle::NumericLiteralCaseStyle &CaseStyle);
+
+  CharTransformFn Prefix;
+  CharTransformFn HexDigit;
+  CharTransformFn FloatExponentSeparator;
+  CharTransformFn Suffix;
+
+  char Separator;
+};
+
+/// @brief Parse a single numeric constant from text into ranges that are
+/// appropriate for applying NumericLiteralCaseStyle rules.
+class QuickNumericalConstantParser {
+public:
+  QuickNumericalConstantParser(const StringRef &IntegerLiteral,
+                               const FormatParameters &Transforms);
+
+  /// @brief Reformats the numeric constant if needed.
+  /// Calling this method invalidates the object's state.
+  /// @return std::nullopt if no reformatting is required. std::option<>
+  /// containing the reformatted string otherwise.
+  std::optional<std::string> formatIfNeeded() &&;
+
+private:
+  const StringRef &IntegerLiteral;
+  const FormatParameters &Transforms;
+
+  std::string Formatted;
+
+  std::string::iterator PrefixBegin;
+  std::string::iterator PrefixEnd;
+  std::string::iterator HexDigitBegin;
+  std::string::iterator HexDigitEnd;
+  std::string::iterator FloatExponentSeparatorBegin;
+  std::string::iterator FloatExponentSeparatorEnd;
+  std::string::iterator SuffixBegin;
+  std::string::iterator SuffixEnd;
+
+  void parse();
+  void applyFormatting();
+};
+
+} // namespace
+
+static char noOpTransform(char C) { return C; }
+
+static CharTransformFn getTransform(int8_t config_value) {
+  switch (config_value) {
+  case -1:
+    return llvm::toLower;
+  case 1:
+    return llvm::toUpper;
+  default:
+    return noOpTransform;
+  }
+}
+
+/// @brief Test if Suffix matches a C++ literal reserved by the library.
+/// Matches against all suffixes reserved in the C++23 standard
+static bool matchesReservedSuffix(StringRef Suffix) {
+  static const std::set<StringRef> ReservedSuffixes = {
+      "h", "min", "s", "ms", "us", "ns", "il", "i", "if", "d", "y",
+  };
+
+  return ReservedSuffixes.find(Suffix) != ReservedSuffixes.end();
+}
+
+FormatParameters::FormatParameters(
+    FormatStyle::LanguageKind Language,
+    const FormatStyle::NumericLiteralCaseStyle &CaseStyle)
+    : Prefix(getTransform(CaseStyle.PrefixCase)),
+      HexDigit(getTransform(CaseStyle.HexDigitCase)),
+      FloatExponentSeparator(
+          getTransform(CaseStyle.FloatExponentSeparatorCase)),
+      Suffix(getTransform(CaseStyle.SuffixCase)) {
+  switch (Language) {
+  case FormatStyle::LK_CSharp:
+  case FormatStyle::LK_Java:
+  case FormatStyle::LK_JavaScript:
+    Separator = '_';
+    break;
+  case FormatStyle::LK_C:
+  case FormatStyle::LK_Cpp:
+  case FormatStyle::LK_ObjC:
+  default:
+    Separator = '\'';
+  }
+}
+
+QuickNumericalConstantParser::QuickNumericalConstantParser(
+    const StringRef &IntegerLiteral, const FormatParameters &Transforms)
+    : IntegerLiteral(IntegerLiteral), Transforms(Transforms),
+      Formatted(IntegerLiteral), PrefixBegin(Formatted.begin()),
+      PrefixEnd(Formatted.begin()), HexDigitBegin(Formatted.begin()),
+      HexDigitEnd(Formatted.begin()),
+      FloatExponentSeparatorBegin(Formatted.begin()),
+      FloatExponentSeparatorEnd(Formatted.begin()),
+      SuffixBegin(Formatted.begin()), SuffixEnd(Formatted.begin()) {}
+
+void QuickNumericalConstantParser::parse() {
+  auto Cur = Formatted.begin();
+  auto End = Formatted.cend();
+
+  bool IsHex = false;
+  bool IsFloat = false;
+
+  // Find the range that contains the prefix.
+  PrefixBegin = Cur;
+  if (*Cur != '0') {
+  } else {
+    ++Cur;
+    const char C = *Cur;
+    switch (C) {
+    case 'x':
+    case 'X':
+      IsHex = true;
+      ++Cur;
+      break;
+    case 'b':
+    case 'B':
+      ++Cur;
+      break;
+    case 'o':
+    case 'O':
+      // Javascript uses 0o as octal prefix.
+      ++Cur;
+      break;
+    default:
+      break;
+    }
+  }
+  PrefixEnd = Cur;
+
+  // Find the range that contains hex digits.
+  HexDigitBegin = Cur;
+  if (IsHex) {
+    while (Cur != End) {
+      const char C = *Cur;
+      if (llvm::isHexDigit(C)) {
+      } else if (C == Transforms.Separator) {
+      } else if (C == '.') {
+        IsFloat = true;
+      } else {
+        break;
+      }
+      ++Cur;
+    }
+  }
+  HexDigitEnd = Cur;
----------------
30Wedge wrote:

Uses `NumericLiteralInfo` now. I rebased to be able to use it.

https://github.com/llvm/llvm-project/pull/151590