regex refactoring 2

* precompiled regex pattern are faster than String.split() (https://shipilev.net/talks/joker-Oct2014-string-catechism.pdf, from page 72) * there are visitors, which use splitting for every ASTNode -> perceptible improvement I used example from SonarOpenCommunity#1415 and could measure some spead-up (10%?) The basic profiling with `-agentlib:hprof=cpu=samples,depth=100` was however not enough to find a root cause
ivangalkin · Apr 29, 2018 · 244c69d · 244c69d
1 parent 64098d6
commit 244c69d
Show file tree

Hide file tree

Showing 12 changed files with 60 additions and 23 deletions.
diff --git a/cxx-checks/src/main/java/org/sonar/cxx/checks/CommentContainsPatternChecker.java b/cxx-checks/src/main/java/org/sonar/cxx/checks/CommentContainsPatternChecker.java
@@ -58,6 +58,8 @@ public class CommentContainsPatternChecker {
 
   private Pattern p;
 
+  private static final Pattern EOLPattern = Pattern.compile("\\R");
+
   /**
    * CommentContainsPatternChecker
    *
@@ -83,7 +85,7 @@ public void visitToken(Token token) {
       String comment = triviaToken.getOriginalValue();
       int line = triviaToken.getLine();
       if (indexOfIgnoreCase(comment) != -1) {
-        String[] lines = comment.split("\r\n?|\n");
+        String[] lines = EOLPattern.split(comment);
 
         for (int i = 0; i < lines.length; i++) {
           int start = indexOfIgnoreCase(lines[i]);

diff --git a/cxx-checks/src/main/java/org/sonar/cxx/checks/MagicNumberCheck.java b/cxx-checks/src/main/java/org/sonar/cxx/checks/MagicNumberCheck.java
@@ -62,7 +62,7 @@ public class MagicNumberCheck extends SquidCheck<Grammar> {
   @Override
   public void init() {
     subscribeTo(CxxTokenType.NUMBER);
-    for (String magicNumber : Arrays.asList(exceptions.split(","))) {
+    for (String magicNumber : exceptions.split(",")) {
       magicNumber = magicNumber.trim();
       if (!magicNumber.isEmpty()) {
         exceptionsSet.add(magicNumber);

diff --git a/cxx-checks/src/main/java/org/sonar/cxx/checks/NoSonarCheck.java b/cxx-checks/src/main/java/org/sonar/cxx/checks/NoSonarCheck.java
@@ -23,6 +23,9 @@
 import com.sonar.sslr.api.Grammar;
 import com.sonar.sslr.api.Token;
 import com.sonar.sslr.api.Trivia;
+
+import java.util.regex.Pattern;
+
 import org.sonar.check.Priority;
 import org.sonar.check.Rule;
 import org.sonar.squidbridge.annotations.ActivatedByDefault;
@@ -40,12 +43,14 @@
 @NoSqale
 public class NoSonarCheck extends SquidCheck<Grammar> implements AstAndTokenVisitor {
 
+  private static final Pattern EOLPattern = Pattern.compile("\\R");
+
   @Override
   public void visitToken(Token token) {
     for (Trivia trivia : token.getTrivia()) {
       if (trivia.isComment()) {
-        String[] commentLines = getContext().getCommentAnalyser()
-          .getContents(trivia.getToken().getOriginalValue()).split("(\r)?\n|\r", -1);
+        String[] commentLines = EOLPattern
+            .split(getContext().getCommentAnalyser().getContents(trivia.getToken().getOriginalValue()), -1);
         int line = trivia.getToken().getLine();
 
         for (String commentLine : commentLines) {

diff --git a/cxx-checks/src/main/java/org/sonar/cxx/checks/ReservedNamesCheck.java b/cxx-checks/src/main/java/org/sonar/cxx/checks/ReservedNamesCheck.java
@@ -26,6 +26,9 @@
 import java.nio.charset.Charset;
 import java.util.List;
 import java.util.Locale;
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
+
 import org.sonar.check.Priority;
 import org.sonar.check.Rule;
 import org.sonar.cxx.api.CxxKeyword;
@@ -50,6 +53,7 @@ public class ReservedNamesCheck extends SquidCheck<Grammar> implements CxxCharse
 
   private static volatile String[] keywords = CxxKeyword.keywordValues();
   private Charset charset = Charset.forName("UTF-8");
+  private static final Pattern defineDeclarationPattern = Pattern.compile("^\\s*#define\\s+([^\\s(]+).*$");
 
   @Override
   public void init() {
@@ -67,9 +71,9 @@ public void visitFile(AstNode astNode) {
     int nr = 0;
     for (String line : lines) {
       nr++;
-      String[] sub = line.split("^\\s*#define\\s+", 2);
-      if (sub.length > 1) {
-        String name = sub[1].split("[\\s(]", 2)[0];
+      Matcher matcher = defineDeclarationPattern.matcher(line);
+      if (matcher.matches()) {
+        String name = matcher.group(1);
         if (name.startsWith("_") && name.length() > 1 && Character.isUpperCase(name.charAt(1))) {
           getContext().createLineViolation(this,
             "Reserved name used for macro (begins with underscore followed by a capital letter)", nr);

diff --git a/cxx-checks/src/main/java/org/sonar/cxx/checks/UnnamedNamespaceInHeaderCheck.java b/cxx-checks/src/main/java/org/sonar/cxx/checks/UnnamedNamespaceInHeaderCheck.java
@@ -39,7 +39,8 @@
 //similar Vera++ rule T017
 public class UnnamedNamespaceInHeaderCheck extends SquidCheck<Grammar> {
 
-  private static final String DEFAULT_NAME_SUFFIX = ".h,.hh,.hpp,.H";
+  private static final String[] DEFAULT_NAME_SUFFIX = new String[] { ".h", ".hh", ".hpp", ".H" };
+  private Boolean isHeader;
 
   @Override
   public void init() {
@@ -48,16 +49,20 @@ public void init() {
     //   settings.getStringArray(CxxPlugin.INCLUDE_DIRECTORIES_KEY)
   }
 
+  @Override
+  public void visitFile(AstNode astNode) {
+    isHeader = isHeader(getContext().getFile().getName());
+  }
+
   @Override
   public void visitNode(AstNode node) {
-    if (isHeader(getContext().getFile().getName())) {
+    if (isHeader) {
       getContext().createFileViolation(this, "Unnamed namespaces are not allowed in header files.", node);
     }
   }
 
   private static boolean isHeader(String name) {
-    String[] suffixes = DEFAULT_NAME_SUFFIX.split(",");
-    for (String suff : suffixes) {
+    for (String suff : DEFAULT_NAME_SUFFIX) {
       if (name.endsWith(suff)) {
         return true;
       }

diff --git a/cxx-checks/src/main/java/org/sonar/cxx/checks/UsingNamespaceInHeaderCheck.java b/cxx-checks/src/main/java/org/sonar/cxx/checks/UsingNamespaceInHeaderCheck.java
@@ -39,24 +39,29 @@
 //similar Vera++ rule T018
 public class UsingNamespaceInHeaderCheck extends SquidCheck<Grammar> {
 
-  private static final String DEFAULT_NAME_SUFFIX = ".h,.hh,.hpp,.H";
+  private static final String[] DEFAULT_NAME_SUFFIX = new String[] { ".h", ".hh", ".hpp", ".H" };
+  private Boolean isHeader;
 
   @Override
   public void init() {
     subscribeTo(CxxGrammarImpl.blockDeclaration);
   }
 
+  @Override
+  public void visitFile(AstNode astNode) {
+    isHeader = isHeader(getContext().getFile().getName());
+  }
+
   @Override
   public void visitNode(AstNode node) {
-    if (isHeader(getContext().getFile().getName())
+    if (isHeader
       && "using".equals(node.getTokenValue()) && node.getFirstChild().getChildren().toString().contains("namespace")) {
       getContext().createLineViolation(this, "Using namespace are not allowed in header files.", node);
     }
   }
 
   private static boolean isHeader(String name) {
-    String[] suffixes = DEFAULT_NAME_SUFFIX.split(",");
-    for (String suff : suffixes) {
+    for (String suff : DEFAULT_NAME_SUFFIX) {
       if (name.endsWith(suff)) {
         return true;
       }

diff --git a/cxx-sensors/src/main/java/org/sonar/cxx/sensors/drmemory/DrMemoryParser.java b/cxx-sensors/src/main/java/org/sonar/cxx/sensors/drmemory/DrMemoryParser.java
@@ -31,6 +31,7 @@
 import org.sonar.api.utils.log.Logger;
 import org.sonar.api.utils.log.Loggers;
 import org.sonar.cxx.sensors.drmemory.DrMemoryParser.DrMemoryError.Location;
+import org.sonar.cxx.sensors.utils.CxxUtils;
 
 public final class DrMemoryParser {
 
@@ -146,7 +147,7 @@ public static List<DrMemoryError> parse(File file, String charset) {
       if (m.find()) {
         DrMemoryError error = new DrMemoryError();
         error.type = extractErrorType(m.group(1));
-        String[] elementSplitted = element.split("\\r?\\n");
+        String[] elementSplitted = CxxUtils.EOLPattern.split(element);
         error.message = elementSplitted[0];
         for (String elementPart : elementSplitted) {
           Matcher locationMatcher = rx_file_finder.matcher(elementPart);

diff --git a/cxx-sensors/src/main/java/org/sonar/cxx/sensors/utils/CxxUtils.java b/cxx-sensors/src/main/java/org/sonar/cxx/sensors/utils/CxxUtils.java
@@ -23,6 +23,8 @@
 import java.io.PrintWriter;
 import java.io.StringWriter;
 import java.util.Optional;
+import java.util.regex.Pattern;
+
 import javax.xml.transform.OutputKeys;
 import javax.xml.transform.Source;
 import javax.xml.transform.Transformer;
@@ -40,6 +42,7 @@
 public final class CxxUtils {
 
   public static final String ERROR_RECOVERY_KEY = "errorRecoveryEnabled";
+  public static final Pattern EOLPattern = Pattern.compile("\\R");
   private static final Logger LOG = Loggers.get(CxxUtils.class);
 
   private CxxUtils() {

diff --git a/cxx-sensors/src/main/java/org/sonar/cxx/sensors/visitors/CxxHighlighterVisitor.java b/cxx-sensors/src/main/java/org/sonar/cxx/sensors/visitors/CxxHighlighterVisitor.java
@@ -35,6 +35,7 @@
 import org.sonar.api.utils.log.Loggers;
 import org.sonar.cxx.api.CxxKeyword;
 import org.sonar.cxx.api.CxxTokenType;
+import org.sonar.cxx.sensors.utils.CxxUtils;
 import org.sonar.squidbridge.SquidAstVisitor;
 
 public class CxxHighlighterVisitor extends SquidAstVisitor<Grammar> implements AstAndTokenVisitor {
@@ -91,7 +92,7 @@ private static class CommentLocation extends TokenLocation {
     public CommentLocation(Token token) {
       super(token);
       String value = token.getValue();
-      String[] lines = value.split("\r\n|\n|\r", -1);
+      String[] lines = CxxUtils.EOLPattern.split(value, -1);
 
       if (lines.length > 1) {
         endLine = token.getLine() + lines.length - 1;
@@ -102,10 +103,11 @@ public CommentLocation(Token token) {
 
   private static class PreprocessorDirectiveLocation extends TokenLocation {
 
+    public static final Pattern preprocessorPattern = Pattern.compile("^[ \t]*#[ \t]*\\w+");
+
     PreprocessorDirectiveLocation(Token token) {
       super(token);
-      Pattern r = Pattern.compile("^[ \t]*#[ \t]*\\w+");
-      Matcher m = r.matcher(token.getValue());
+      Matcher m = preprocessorPattern.matcher(token.getValue());
       if (m.find()) {
         endLineOffset = startLineOffset + (m.end() - m.start());
       } else {

diff --git a/cxx-squid/src/main/java/org/sonar/cxx/CxxLanguage.java b/cxx-squid/src/main/java/org/sonar/cxx/CxxLanguage.java
@@ -24,6 +24,8 @@
 import java.util.List;
 import java.util.Map;
 import java.util.Optional;
+import java.util.regex.Pattern;
+
 import org.sonar.api.config.Configuration;
 import org.sonar.api.measures.Metric;
 import org.sonar.api.resources.AbstractLanguage;
@@ -36,6 +38,7 @@ public abstract class CxxLanguage extends AbstractLanguage {
   public static final String ERROR_RECOVERY_KEY = "errorRecoveryEnabled";
   private final Configuration settings;
   private final Map<String, Metric> MetricsCache;
+  public static final Pattern EOLPattern = Pattern.compile("\\R");
 
   public CxxLanguage(String key, Configuration settings) {
     super(key);
@@ -89,7 +92,7 @@ public Optional<Boolean> IsRecoveryEnabled() {
   public String[] getStringLinesOption(String key) {
     Optional<String> value = this.settings.get(getPluginProperty(key));
     if (value.isPresent()) {
-      return value.get().split("\r?\n|\r", -1);
+      return EOLPattern.split(value.get(), -1);
     }
     return new String[0];
   }

diff --git a/cxx-squid/src/main/java/org/sonar/cxx/visitors/CxxCommentLinesVisitor.java b/cxx-squid/src/main/java/org/sonar/cxx/visitors/CxxCommentLinesVisitor.java
@@ -26,6 +26,8 @@
 import com.sonar.sslr.api.Trivia;
 import java.util.HashSet;
 import java.util.Set;
+import java.util.regex.Pattern;
+
 import org.sonar.cxx.api.CxxMetric;
 import org.sonar.cxx.api.CxxPunctuator;
 import org.sonar.squidbridge.SquidAstVisitor;
@@ -42,6 +44,7 @@ public class CxxCommentLinesVisitor<GRAMMAR extends Grammar> extends SquidAstVis
 
   private final Set<Integer> comments = new HashSet<>();
   private boolean seenFirstToken;
+  public static final Pattern EOLPattern = Pattern.compile("\\R");
 
   @Override
   public void init() {
@@ -59,8 +62,8 @@ public void visitToken(Token token) {
     for (Trivia trivia : token.getTrivia()) {
       if (trivia.isComment()) {
         if (seenFirstToken) {
-          String[] commentLines = getContext().getCommentAnalyser().getContents(trivia.getToken().getOriginalValue())
-            .split("(\r)?\n|\r", -1);
+          String[] commentLines = EOLPattern
+              .split(getContext().getCommentAnalyser().getContents(trivia.getToken().getOriginalValue()), -1);
           int line = trivia.getToken().getLine();
           for (String commentLine : commentLines) {
             if (!commentLine.contains("NOSONAR") && !getContext().getCommentAnalyser().isBlank(commentLine)) {

diff --git a/cxx-squid/src/main/java/org/sonar/cxx/visitors/CxxLinesOfCodeVisitor.java b/cxx-squid/src/main/java/org/sonar/cxx/visitors/CxxLinesOfCodeVisitor.java
@@ -22,6 +22,9 @@
 import com.sonar.sslr.api.AstAndTokenVisitor;
 import com.sonar.sslr.api.AstNode;
 import static com.sonar.sslr.api.GenericTokenType.EOF;
+
+import java.util.regex.Pattern;
+
 import com.sonar.sslr.api.Grammar;
 import com.sonar.sslr.api.Token;
 import org.sonar.squidbridge.SquidAstVisitor;
@@ -37,6 +40,7 @@ public class CxxLinesOfCodeVisitor<GRAMMAR extends Grammar>
 
   private final MetricDef metric;
   private int lastTokenLine;
+  public static final Pattern EOLPattern = Pattern.compile("\\R");
 
   public CxxLinesOfCodeVisitor(MetricDef metric) {
     this.metric = metric;
@@ -57,7 +61,7 @@ public void visitFile(AstNode node) {
   public void visitToken(Token token) {
     if (!token.getType().equals(EOF)) {
       /* Handle all the lines of the token */
-      String[] tokenLines = token.getValue().split("\n", -1);
+      String[] tokenLines = EOLPattern.split(token.getValue(), -1);
 
       int firstLineAlreadyCounted = lastTokenLine == token.getLine() ? 1 : 0;
       getContext().peekSourceCode().add(metric, (double) tokenLines.length - firstLineAlreadyCounted);