[ignore rules] fix for backslash handling

An attempt to re-implement not well documented Git CLI behavior for
patterns with backslashes.

It looks like Git silently ignores all \ characters in ignore rules, if
they are NOT covered by 3 cases described in [1]:

{quote}
1) ... Put a backslash ("\") in front of the first hash for patterns
that begin with a hash.
...
2) Trailing spaces are ignored unless they are quoted with backslash
("\").
...
3) Put a backslash ("\") in front of the first "!" for patterns that
begin with a literal "!", for example, "\!important!.txt".
{quote}

Undocumented also is the fact that backslash itself can be escaped by
backslash.

So \h\e\l\l\o\.t\x\t rule matches hello.txt and a\\\\b a\b in Git CLI.

Additionally, the glob parser [2] knows special meaning of backslash:

{quote}
One can remove the special meaning of '?', '*' and '[' by preceding
them by a backslash, or, in case this is part of a shell command
line, enclosing them in quotes.  Between brackets these characters
stand for themselves.  Thus, "[[?*\]" matches the four characters
'[', '?', '*' and '\'.
{quote}

[1] https://www.kernel.org/pub/software/scm/git/docs/gitignore.html
[2] http://man7.org/linux/man-pages/man7/glob.7.html

Bug: 478065
Change-Id: I3dc973475d1943c5622103701fa8cb3ea0684e3e
Signed-off-by: Andrey Loskutov <loskutov@gmx.de>

7 files changed, 82 insertions, 17 deletions

diff --git a/org.eclipse.jgit.test/tst/org/eclipse/jgit/ignore/IgnoreRuleSpecialCasesTest.java b/org.eclipse.jgit.test/tst/org/eclipse/jgit/ignore/IgnoreRuleSpecialCasesTest.java
index ec4a1f1c67..f8eb126826 100644
--- a/org.eclipse.jgit.test/tst/org/eclipse/jgit/ignore/IgnoreRuleSpecialCasesTest.java
+++ b/org.eclipse.jgit.test/tst/org/eclipse/jgit/ignore/IgnoreRuleSpecialCasesTest.java
@@ -830,21 +830,29 @@ public class IgnoreRuleSpecialCasesTest {
 	}
 
 	@Test
+	public void testIgnoredBackslash() throws Exception {
+		// In Git CLI a\b\c is equal to abc
+		assertMatch("a\\b\\c", "abc", true);
+	}
+
+	@Test
 	public void testEscapedBackslash() throws Exception {
 		// In Git CLI a\\b matches a\b file
 		assertMatch("a\\\\b", "a\\b", true);
+		assertMatch("a\\\\b\\c", "a\\bc", true);
+
 	}
 
 	@Test
 	public void testEscapedExclamationMark() throws Exception {
 		assertMatch("\\!b!.txt", "!b!.txt", true);
-		assertMatch("a\\!b!.txt", "a\\!b!.txt", true);
+		assertMatch("a\\!b!.txt", "a!b!.txt", true);
 	}
 
 	@Test
 	public void testEscapedHash() throws Exception {
 		assertMatch("\\#b", "#b", true);
-		assertMatch("a\\#", "a\\#", true);
+		assertMatch("a\\#", "a#", true);
 	}
 
 	@Test
@@ -855,12 +863,12 @@ public class IgnoreRuleSpecialCasesTest {
 
 	@Test
 	public void testNotEscapingBackslash() throws Exception {
-		assertMatch("\\out", "\\out", true);
-		assertMatch("\\out", "a/\\out", true);
-		assertMatch("c:\\/", "c:\\/", true);
-		assertMatch("c:\\/", "a/c:\\/", true);
-		assertMatch("c:\\tmp", "c:\\tmp", true);
-		assertMatch("c:\\tmp", "a/c:\\tmp", true);
+		assertMatch("\\out", "out", true);
+		assertMatch("\\out", "a/out", true);
+		assertMatch("c:\\/", "c:/", true);
+		assertMatch("c:\\/", "a/c:/", true);
+		assertMatch("c:\\tmp", "c:tmp", true);
+		assertMatch("c:\\tmp", "a/c:tmp", true);
 	}
 
 	@Test
@@ -869,6 +877,22 @@ public class IgnoreRuleSpecialCasesTest {
 	}
 
 	@Test
+	public void testBackslash() throws Exception {
+		assertMatch("a\\", "a", true);
+		assertMatch("\\a", "a", true);
+		assertMatch("a/\\", "a/", true);
+		assertMatch("a/b\\", "a/b", true);
+		assertMatch("\\a/b", "a/b", true);
+		assertMatch("/\\a", "/a", true);
+		assertMatch("\\a\\b\\c\\", "abc", true);
+		assertMatch("/\\a/\\b/\\c\\", "a/b/c", true);
+
+		// empty path segment doesn't match
+		assertMatch("\\/a", "/a", false);
+		assertMatch("\\/a", "a", false);
+	}
+
+	@Test
 	public void testDollar() throws Exception {
 		assertMatch("$", "$", true);
 		assertMatch("$x", "$x", true);
diff --git a/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/LeadingAsteriskMatcher.java b/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/LeadingAsteriskMatcher.java
index f1153d9c69..3d0ad09124 100644
--- a/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/LeadingAsteriskMatcher.java
+++ b/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/LeadingAsteriskMatcher.java
@@ -50,7 +50,7 @@ package org.eclipse.jgit.ignore.internal;
 public class LeadingAsteriskMatcher extends NameMatcher {
 
 	LeadingAsteriskMatcher(String pattern, Character pathSeparator, boolean dirOnly) {
-		super(pattern, pathSeparator, dirOnly);
+		super(pattern, pathSeparator, dirOnly, true);
 
 		if (subPattern.charAt(0) != '*')
 			throw new IllegalArgumentException(
diff --git a/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/NameMatcher.java b/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/NameMatcher.java
index 6c4c8092fe..8beae8379e 100644
--- a/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/NameMatcher.java
+++ b/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/NameMatcher.java
@@ -58,9 +58,13 @@ public class NameMatcher extends AbstractMatcher {
 
 	final String subPattern;
 
-	NameMatcher(String pattern, Character pathSeparator, boolean dirOnly) {
+	NameMatcher(String pattern, Character pathSeparator, boolean dirOnly,
+			boolean deleteBackslash) {
 		super(pattern, dirOnly);
 		slash = getPathSeparator(pathSeparator);
+		if (deleteBackslash) {
+			pattern = Strings.deleteBackslash(pattern);
+		}
 		beginning = pattern.length() == 0 ? false : pattern.charAt(0) == slash;
 		if (!beginning)
 			this.subPattern = pattern;
diff --git a/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/PathMatcher.java b/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/PathMatcher.java
index 6b9b5a806b..c3f6694a7a 100644
--- a/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/PathMatcher.java
+++ b/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/PathMatcher.java
@@ -85,7 +85,8 @@ public class PathMatcher extends AbstractMatcher {
 	}
 
 	private boolean isSimplePathWithSegments(String path) {
-		return !isWildCard(path) && count(path, slash, true) > 0;
+		return !isWildCard(path) && path.indexOf('\\') < 0
+				&& count(path, slash, true) > 0;
 	}
 
 	static private List<IMatcher> createMatchers(List<String> segments,
@@ -167,7 +168,7 @@ public class PathMatcher extends AbstractMatcher {
 		case COMPLEX:
 			return new WildCardMatcher(segment, pathSeparator, dirOnly);
 		default:
-			return new NameMatcher(segment, pathSeparator, dirOnly);
+			return new NameMatcher(segment, pathSeparator, dirOnly, true);
 		}
 	}
 
diff --git a/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/Strings.java b/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/Strings.java
index 2e955d3124..f972828bc6 100644
--- a/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/Strings.java
+++ b/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/Strings.java
@@ -157,9 +157,7 @@ public class Strings {
 					return false;
 				}
 				char nextChar = pattern.charAt(nextIdx);
-				if (nextChar == '?' || nextChar == '*' || nextChar == '['
-				// required to match escaped backslashes '\\\\'
-						|| nextChar == '\\') {
+				if (escapedByBackslash(nextChar)) {
 					return true;
 				} else {
 					return false;
@@ -169,6 +167,10 @@ public class Strings {
 		return false;
 	}
 
+	private static boolean escapedByBackslash(char nextChar) {
+		return nextChar == '?' || nextChar == '*' || nextChar == '[';
+	}
+
 	static PatternState checkWildCards(String pattern) {
 		if (isComplexWildcard(pattern))
 			return PatternState.COMPLEX;
@@ -308,6 +310,14 @@ public class Strings {
 					char lookAhead = lookAhead(pattern, i);
 					if (lookAhead == ']' || lookAhead == '[')
 						ignoreLastBracket = true;
+				} else {
+					//
+					char lookAhead = lookAhead(pattern, i);
+					if (lookAhead != '\\' && lookAhead != '['
+							&& lookAhead != '?' && lookAhead != '*'
+							&& lookAhead != ' ' && lookBehind(sb) != '\\') {
+						break;
+					}
 				}
 				sb.append(c);
 				break;
@@ -445,4 +455,30 @@ public class Strings {
 		return null;
 	}
 
+	static String deleteBackslash(String s) {
+		if (s.indexOf('\\') < 0) {
+			return s;
+		}
+		StringBuilder sb = new StringBuilder(s.length());
+		for (int i = 0; i < s.length(); i++) {
+			char ch = s.charAt(i);
+			if (ch == '\\') {
+				if (i + 1 == s.length()) {
+					continue;
+				}
+				char next = s.charAt(i + 1);
+				if (next == '\\') {
+					sb.append(ch);
+					i++;
+					continue;
+				}
+				if (!escapedByBackslash(next)) {
+					continue;
+				}
+			}
+			sb.append(ch);
+		}
+		return sb.toString();
+	}
+
 }
diff --git a/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/TrailingAsteriskMatcher.java b/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/TrailingAsteriskMatcher.java
index 4a1c780d99..b927d27dbe 100644
--- a/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/TrailingAsteriskMatcher.java
+++ b/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/TrailingAsteriskMatcher.java
@@ -50,7 +50,7 @@ package org.eclipse.jgit.ignore.internal;
 public class TrailingAsteriskMatcher extends NameMatcher {
 
 	TrailingAsteriskMatcher(String pattern, Character pathSeparator, boolean dirOnly) {
-		super(pattern, pathSeparator, dirOnly);
+		super(pattern, pathSeparator, dirOnly, true);
 
 		if (subPattern.charAt(subPattern.length() - 1) != '*')
 			throw new IllegalArgumentException(
diff --git a/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/WildCardMatcher.java b/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/WildCardMatcher.java
index 7d12b0ddce..8f9815283d 100644
--- a/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/WildCardMatcher.java
+++ b/org.eclipse.jgit/src/org/eclipse/jgit/ignore/internal/WildCardMatcher.java
@@ -62,7 +62,7 @@ public class WildCardMatcher extends NameMatcher {
 
 	WildCardMatcher(String pattern, Character pathSeparator, boolean dirOnly)
 			throws InvalidPatternException {
-		super(pattern, pathSeparator, dirOnly);
+		super(pattern, pathSeparator, dirOnly, false);
 		p = convertGlob(subPattern);
 	}