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

[owenca]

No description provided.

[llvmbot]

@llvm/pr-subscribers-clang-format

Author: owenca (owenca)

Changes

---

Full diff: https://github.com/llvm/llvm-project/pull/152878.diff

5 Files Affected:

• (modified) clang/lib/Format/CMakeLists.txt (+1)
• (added) clang/lib/Format/NumericLiteralInfo.cpp (+67)
• (added) clang/lib/Format/NumericLiteralInfo.h (+41)
• (modified) clang/unittests/Format/CMakeLists.txt (+1)
• (added) clang/unittests/Format/NumericLiteralInfoTest.cpp (+64)

diff --git a/clang/lib/Format/CMakeLists.txt b/clang/lib/Format/CMakeLists.txt
index 9f4939824fdb8..24f435d2caee1 100644
--- a/clang/lib/Format/CMakeLists.txt
+++ b/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
diff --git a/clang/lib/Format/NumericLiteralInfo.cpp b/clang/lib/Format/NumericLiteralInfo.cpp
new file mode 100644
index 0000000000000..2c7a7601ed633
--- /dev/null
+++ b/clang/lib/Format/NumericLiteralInfo.cpp
@@ -0,0 +1,67 @@
+//===--- 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;
+
+constexpr auto npos = StringRef::npos;
+
+NumericLiteralInfo::NumericLiteralInfo(StringRef Text, char Separator) {
+  assert(Text.size() > 1);
+
+  bool IsHex = false;
+  BaseLetterPos = npos;
+  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 ? 2 : 0); // e.g. 0x.1 or .1
+
+  // e.g. 1.e2 or 0xFp2
+  const auto Pos = DotPos != npos ? DotPos + 1 : BaseLetterPos + 2;
+
+  ExponentLetterPos =
+      // Trim C++ user-defined suffix as in `1_Pa`.
+      (Separator == '\'' ? Text.substr(0, Text.find('_')) : Text)
+          .find_insensitive(IsHex ? 'p' : 'e', Pos);
+
+  const bool HasExponent = ExponentLetterPos != 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
diff --git a/clang/lib/Format/NumericLiteralInfo.h b/clang/lib/Format/NumericLiteralInfo.h
new file mode 100644
index 0000000000000..8ed5e87e2f410
--- /dev/null
+++ b/clang/lib/Format/NumericLiteralInfo.h
@@ -0,0 +1,41 @@
+//===--- 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 {
+
+extern const size_t npos;
+
+struct NumericLiteralInfo {
+  size_t BaseLetterPos;     // Position of the base letter.
+  size_t DotPos;            // Position of the decimal/hexadecimal point.
+  size_t ExponentLetterPos; // Position of the exponent letter.
+  size_t SuffixPos;         // Starting position of the suffix.
+
+  NumericLiteralInfo(size_t BaseLetterPos = npos, size_t DotPos = npos,
+                     size_t ExponentLetterPos = npos, size_t SuffixPos = npos)
+      : BaseLetterPos(BaseLetterPos), DotPos(DotPos),
+        ExponentLetterPos(ExponentLetterPos), SuffixPos(SuffixPos) {}
+
+  NumericLiteralInfo(llvm::StringRef Text, char Separator);
+
+  bool operator==(const NumericLiteralInfo &R) const {
+    return BaseLetterPos == R.BaseLetterPos && DotPos == R.DotPos &&
+           ExponentLetterPos == R.ExponentLetterPos && SuffixPos == R.SuffixPos;
+  }
+};
+
+} // end namespace format
+} // end namespace clang
+
+#endif
diff --git a/clang/unittests/Format/CMakeLists.txt b/clang/unittests/Format/CMakeLists.txt
index edfc8d7a5beaa..c4c7b483ba68e 100644
--- a/clang/unittests/Format/CMakeLists.txt
+++ b/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
diff --git a/clang/unittests/Format/NumericLiteralInfoTest.cpp b/clang/unittests/Format/NumericLiteralInfoTest.cpp
new file mode 100644
index 0000000000000..6597cf3772de5
--- /dev/null
+++ b/clang/unittests/Format/NumericLiteralInfoTest.cpp
@@ -0,0 +1,64 @@
+//===- 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 {
+
+class NumericLiteralInfoTest : public testing::Test {
+protected:
+  NumericLiteralInfo getInfo(llvm::StringRef Text, char Separator = '\'') {
+    return NumericLiteralInfo(Text, Separator);
+  }
+};
+
+TEST_F(NumericLiteralInfoTest, IntegerLiteral) {
+  // Decimal.
+  EXPECT_EQ(getInfo("90"), NumericLiteralInfo());
+  EXPECT_EQ(getInfo("9L"), NumericLiteralInfo(npos, npos, npos, 1));
+  EXPECT_EQ(getInfo("9'0U"), NumericLiteralInfo(npos, npos, npos, 3));
+
+  // Octal.
+  EXPECT_EQ(getInfo("07"), NumericLiteralInfo());
+  EXPECT_EQ(getInfo("0z"), NumericLiteralInfo(npos, npos, npos, 1));
+  // JavaScript.
+  EXPECT_EQ(getInfo("0o7"), NumericLiteralInfo(1));
+  EXPECT_EQ(getInfo("0O7_0", '_'), NumericLiteralInfo(1));
+
+  // Binary.
+  EXPECT_EQ(getInfo("0b1"), NumericLiteralInfo(1));
+  EXPECT_EQ(getInfo("0B1ul"), NumericLiteralInfo(1, npos, npos, 3));
+
+  // Hexadecimal.
+  EXPECT_EQ(getInfo("0xF"), NumericLiteralInfo(1));
+  EXPECT_EQ(getInfo("0XfZ"), NumericLiteralInfo(1, npos, npos, 3));
+}
+
+TEST_F(NumericLiteralInfoTest, FloatingPointLiteral) {
+  // Decimal.
+  EXPECT_EQ(getInfo(".9"), NumericLiteralInfo(npos, 0));
+  EXPECT_EQ(getInfo("9."), NumericLiteralInfo(npos, 1));
+  EXPECT_EQ(getInfo("9.F"), NumericLiteralInfo(npos, 1, npos, 2));
+  EXPECT_EQ(getInfo("9e9"), NumericLiteralInfo(npos, npos, 1));
+  EXPECT_EQ(getInfo("9E-9f"), NumericLiteralInfo(npos, npos, 1, 4));
+  EXPECT_EQ(getInfo("9.9e+9bf16"), NumericLiteralInfo(npos, 1, 3, 6));
+
+  // Hexadecimal.
+  EXPECT_EQ(getInfo("0X.Fp9"), NumericLiteralInfo(1, 2, 4));
+  EXPECT_EQ(getInfo("0xF.P9"), NumericLiteralInfo(1, 3, 4));
+  EXPECT_EQ(getInfo("0xFp9"), NumericLiteralInfo(1, npos, 3));
+  EXPECT_EQ(getInfo("0xFp+9F128"), NumericLiteralInfo(1, npos, 3, 6));
+  EXPECT_EQ(getInfo("0xF.Fp-9_Pa"), NumericLiteralInfo(1, 3, 5, 8));
+}
+
+} // namespace
+} // namespace format
+} // namespace clang

[owenca]

This is useful for #131510 and #151590.