[clang-format] Add functionality of getting info about numeric literals

clang/lib/Format/CMakeLists.txt

@@ -13,6 +13,7 @@ add_clang_library(clangFormat
   MacroExpander.cpp
   MatchFilePath.cpp
   NamespaceEndCommentsFixer.cpp
+  NumericLiteralInfo.cpp
   ObjCPropertyAttributeOrderFixer.cpp
   QualifierAlignmentFixer.cpp
   SortJavaScriptImports.cpp

clang/lib/Format/NumericLiteralInfo.cpp

@@ -0,0 +1,65 @@
+//===--- NumericLiteralInfo.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 the functionality of getting information about a
+/// numeric literal string, including 0-based positions of the base letter, the
+/// decimal/hexadecimal point, the exponent letter, and the suffix, or npos if
+/// absent.
+///
+//===----------------------------------------------------------------------===//
+
+#include "NumericLiteralInfo.h"
+#include "llvm/ADT/StringExtras.h"
+
+namespace clang {
+namespace format {
+
+using namespace llvm;
+
+NumericLiteralInfo::NumericLiteralInfo(StringRef Text, char Separator) {
+  if (Text.size() < 2)
+    return;
+
+  bool IsHex = false;
+  if (Text[0] == '0') {
+    switch (Text[1]) {
+    case 'x':
+    case 'X':
+      IsHex = true;
+      [[fallthrough]];
+    case 'b':
+    case 'B':
+    case 'o':
+    case 'O':
+      BaseLetterPos = 1; // e.g. 0xF
+      break;
+    }
+  }
+
+  DotPos = Text.find('.', BaseLetterPos + 1); // e.g. 0x.1 or .1
+
+  // e.g. 1.e2 or 0xFp2
+  const auto Pos = DotPos != StringRef::npos ? DotPos + 1 : BaseLetterPos + 2;
+
+  ExponentLetterPos =
+      // Trim C++ user-defined suffix as in `1_Pa`.
+      (Separator == '\'' ? Text.take_front(Text.find('_')) : Text)
+          .find_insensitive(IsHex ? 'p' : 'e', Pos);
+
+  const bool HasExponent = ExponentLetterPos != StringRef::npos;
+  SuffixPos = Text.find_if_not(
+      [&](char C) {
+        return (HasExponent || !IsHex ? isDigit : isHexDigit)(C) ||
+               C == Separator;
+      },
+      HasExponent ? ExponentLetterPos + 2 : Pos); // e.g. 1e-2f
+}
+
+} // namespace format
+} // namespace clang

clang/lib/Format/NumericLiteralInfo.h

@@ -0,0 +1,29 @@
+//===--- NumericLiteralInfo.h -----------------------------------*- 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
+//
+//===----------------------------------------------------------------------===//
+
+#ifndef LLVM_CLANG_LIB_FORMAT_NUMERICLITERALINFO_H
+#define LLVM_CLANG_LIB_FORMAT_NUMERICLITERALINFO_H
+
+#include "llvm/ADT/StringRef.h"
+
+namespace clang {
+namespace format {
+
+struct NumericLiteralInfo {
+  size_t BaseLetterPos = llvm::StringRef::npos;     // as in 0b1, 0xF, etc.
+  size_t DotPos = llvm::StringRef::npos;            // pos of decimal/hex point
+  size_t ExponentLetterPos = llvm::StringRef::npos; // as in 9e9 and 0xFp9
+  size_t SuffixPos = llvm::StringRef::npos;         // starting pos of suffix
+
+  NumericLiteralInfo(llvm::StringRef Text, char Separator = '\'');
+};
+
+} // end namespace format
+} // end namespace clang
+
+#endif

clang/unittests/Format/CMakeLists.txt

@@ -27,6 +27,7 @@ add_distinct_clang_unittest(FormatTests
   MacroExpanderTest.cpp
   MatchFilePathTest.cpp
   NamespaceEndCommentsFixerTest.cpp
+  NumericLiteralInfoTest.cpp
   ObjCPropertyAttributeOrderFixerTest.cpp
   QualifierFixerTest.cpp
   SortImportsTestJS.cpp

clang/unittests/Format/NumericLiteralInfoTest.cpp

@@ -0,0 +1,71 @@
+//===- unittest/Format/NumericLiteralInfoTest.cpp -------------------------===//
+//
+// 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
+//
+//===----------------------------------------------------------------------===//
+
+#include "../../lib/Format/NumericLiteralInfo.h"
+#include "gtest/gtest.h"
+
+namespace clang {
+namespace format {
+namespace {
+
+static constexpr auto npos = llvm::StringRef::npos;
+
+class NumericLiteralInfoTest : public testing::Test {
+protected:
+  bool verifyInfo(const NumericLiteralInfo &Info, size_t BaseLetterPos = npos,
+                  size_t DotPos = npos, size_t ExponentLetterPos = npos,
+                  size_t SuffixPos = npos) {
+    return Info.BaseLetterPos == BaseLetterPos && Info.DotPos == DotPos &&
+           Info.ExponentLetterPos == ExponentLetterPos &&
+           Info.SuffixPos == SuffixPos;
+  }
+};
+
+TEST_F(NumericLiteralInfoTest, IntegerLiteral) {
+  // Decimal.
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("90")));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("9L"), npos, npos, npos, 1));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("9'0U"), npos, npos, npos, 3));
+
+  // Octal.
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("0")));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("07")));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("0z"), npos, npos, npos, 1));
+  // JavaScript.
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("0o7"), 1));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("0O7_0", '_'), 1));
+
+  // Binary.
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("0b1"), 1));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("0B1ul"), 1, npos, npos, 3));
+
+  // Hexadecimal.
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("0xF"), 1));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("0XfZ"), 1, npos, npos, 3));
+}
+
+TEST_F(NumericLiteralInfoTest, FloatingPointLiteral) {
+  // Decimal.
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo(".9"), npos, 0));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("9."), npos, 1));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("9.F"), npos, 1, npos, 2));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("9e9"), npos, npos, 1));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("9E-9f"), npos, npos, 1, 4));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("9.9e+9bf16"), npos, 1, 3, 6));
+
+  // Hexadecimal.
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("0X.Fp9"), 1, 2, 4));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("0xF.P9"), 1, 3, 4));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("0xFp9"), 1, npos, 3));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("0xFp+9F128"), 1, npos, 3, 6));
+  EXPECT_TRUE(verifyInfo(NumericLiteralInfo("0xF.Fp-9_Pa"), 1, 3, 5, 8));
+}
+
+} // namespace
+} // namespace format
+} // namespace clang