[NFC][GlobPattern] Add GlobPattern::longest_substr() (llvm#164512)

Finds longest (almost) plain substring in the pattern.

Implementation is conservative to avoid false positives.

The result is not used to optimize
`GlobPattern::match()` so it's calculated on
request.

For
* llvm#164545

---------

Co-authored-by: Luke Lau <[email protected]>

Author: 2 people

llvm/include/llvm/Support/GlobPattern.h

@@ -79,6 +79,9 @@ class GlobPattern {
   StringRef prefix() const { return Pattern.take_front(PrefixSize); }
   // Returns plain suffix of the pattern.
   StringRef suffix() const { return Pattern.take_back(SuffixSize); }
+  // Returns the longest plain substring of the pattern between prefix and
+  // suffix.
+  StringRef longest_substr() const;
 
 private:
   StringRef Pattern;

llvm/lib/Support/GlobPattern.cpp

@@ -132,6 +132,49 @@ parseBraceExpansions(StringRef S, std::optional<size_t> MaxSubPatterns) {
   return std::move(SubPatterns);
 }
 
+static StringRef maxPlainSubstring(StringRef S) {
+  StringRef Best;
+  while (!S.empty()) {
+    size_t PrefixSize = S.find_first_of("?*[{\\");
+    if (PrefixSize == std::string::npos)
+      PrefixSize = S.size();
+
+    if (Best.size() < PrefixSize)
+      Best = S.take_front(PrefixSize);
+
+    S = S.drop_front(PrefixSize);
+
+    // It's impossible, as the first and last characters of the input string
+    // must be Glob special characters, otherwise they would be parts of
+    // the prefix or the suffix.
+    assert(!S.empty());
+
+    switch (S.front()) {
+    case '\\':
+      S = S.drop_front(2);
+      break;
+    case '[': {
+      // Drop '[' and the first character which can be ']'.
+      S = S.drop_front(2);
+      size_t EndBracket = S.find_first_of("]");
+      // Should not be possible, SubGlobPattern::create should fail on invalid
+      // pattern before we get here.
+      assert(EndBracket != std::string::npos);
+      S = S.drop_front(EndBracket + 1);
+      break;
+    }
+    case '{':
+      // TODO: implement.
+      // Fallback to whatever is best for now.
+      return Best;
+    default:
+      S = S.drop_front(1);
+    }
+  }
+
+  return Best;
+}
+
 Expected<GlobPattern>
 GlobPattern::create(StringRef S, std::optional<size_t> MaxSubPatterns) {
   GlobPattern Pat;
@@ -202,6 +245,11 @@ GlobPattern::SubGlobPattern::create(StringRef S) {
   return Pat;
 }
 
+StringRef GlobPattern::longest_substr() const {
+  return maxPlainSubstring(
+      Pattern.drop_front(PrefixSize).drop_back(SuffixSize));
+}
+
 bool GlobPattern::match(StringRef S) const {
   if (!S.consume_front(prefix()))
     return false;

llvm/unittests/Support/GlobPatternTest.cpp

@@ -329,6 +329,72 @@ TEST_F(GlobPatternTest, PrefixSuffix) {
   EXPECT_EQ("cd", Pat->suffix());
 }
 
+TEST_F(GlobPatternTest, Substr) {
+  auto Pat = GlobPattern::create("");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("", Pat->longest_substr());
+
+  Pat = GlobPattern::create("abcd");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("", Pat->longest_substr());
+
+  Pat = GlobPattern::create("a*bcd");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("", Pat->longest_substr());
+
+  Pat = GlobPattern::create("*abcd");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("", Pat->longest_substr());
+
+  Pat = GlobPattern::create("abcd*");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("", Pat->longest_substr());
+
+  Pat = GlobPattern::create("a*bc*d");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("bc", Pat->longest_substr());
+
+  Pat = GlobPattern::create("a*bc*def*g");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("def", Pat->longest_substr());
+
+  Pat = GlobPattern::create("a*bcd*ef*g");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("bcd", Pat->longest_substr());
+
+  Pat = GlobPattern::create("a*bcd*efg*h");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("bcd", Pat->longest_substr());
+
+  Pat = GlobPattern::create("a*bcd[ef]g*h");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("bcd", Pat->longest_substr());
+
+  Pat = GlobPattern::create("a*bc[d]efg*h");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("efg", Pat->longest_substr());
+
+  Pat = GlobPattern::create("a*bc[]]efg*h");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("efg", Pat->longest_substr());
+
+  Pat = GlobPattern::create("a*bcde\\fg*h");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("bcde", Pat->longest_substr());
+
+  Pat = GlobPattern::create("a*bcde\\[fg*h");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("bcde", Pat->longest_substr());
+
+  Pat = GlobPattern::create("a*bcde?fg*h");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("bcde", Pat->longest_substr());
+
+  Pat = GlobPattern::create("a*bcdef{g}*h");
+  ASSERT_TRUE((bool)Pat);
+  EXPECT_EQ("bcdef", Pat->longest_substr());
+}
+
 TEST_F(GlobPatternTest, Pathological) {
   std::string P, S(40, 'a');
   StringRef Pieces[] = {"a*", "[ba]*", "{b*,a*}*"};