Merge pull request #2188 from guwirth/optimize-regex

guwirth · web-flow · commit 3e8b176a23f1 · 2021-06-10T19:00:13.000+02:00
optimize regular expressions
diff --git a/cxx-sensors/src/main/java/org/sonar/cxx/sensors/clangtidy/ClangTidyParser.java b/cxx-sensors/src/main/java/org/sonar/cxx/sensors/clangtidy/ClangTidyParser.java
@@ -33,7 +33,7 @@ public class ClangTidyParser {
 
   private static final Logger LOG = Loggers.get(ClangTidyParser.class);
 
-  private static final String REGEX = "(.+|[a-zA-Z]:\\\\.+):([0-9]+):([0-9]+): ([^:]+): (.+)";
+  private static final String REGEX = "((?>[a-zA-Z]:\\\\)??[^:]++):(\\d{1,5}):(\\d{1,5}): ([^:]++): (.+)";
   private static final Pattern PATTERN = Pattern.compile(REGEX);
 
   private final CxxClangTidySensor sensor;
@@ -44,7 +44,7 @@ public ClangTidyParser(CxxClangTidySensor sensor) {
   }
 
   public void parse(File report, String defaultEncoding) throws IOException {
-    try ( var scanner = new TextScanner(report, defaultEncoding)) {
+    try (var scanner = new TextScanner(report, defaultEncoding)) {
       LOG.debug("Encoding='{}'", scanner.encoding());
 
       CxxReportIssue currentIssue = null;
diff --git a/cxx-sensors/src/main/java/org/sonar/cxx/sensors/compiler/gcc/CxxCompilerGccSensor.java b/cxx-sensors/src/main/java/org/sonar/cxx/sensors/compiler/gcc/CxxCompilerGccSensor.java
@@ -42,7 +42,7 @@ public class CxxCompilerGccSensor extends CxxCompilerSensor {
    */
   public static final String DEFAULT_ID = "default";
   public static final String DEFAULT_REGEX_DEF
-                               = "(?<file>.*):(?<line>[0-9]+):[0-9]+:\\x20warning:\\x20"
+                               = "(?<file>[^:]*+):(?<line>\\d{1,5}):\\d{1,5}:\\x20warning:\\x20"
                                    + "(?<message>.*?)(\\x20\\[(?<id>.*)\\])?\\s*$";
 
   public static List<PropertyDefinition> properties() {
diff --git a/cxx-sensors/src/main/java/org/sonar/cxx/sensors/compiler/vc/CxxCompilerVcSensor.java b/cxx-sensors/src/main/java/org/sonar/cxx/sensors/compiler/vc/CxxCompilerVcSensor.java
@@ -36,8 +36,9 @@ public class CxxCompilerVcSensor extends CxxCompilerSensor {
   public static final String REPORT_ENCODING_DEF = "sonar.cxx.vc.encoding";
   public static final String DEFAULT_ENCODING_DEF = StandardCharsets.UTF_8.name();
   public static final String DEFAULT_REGEX_DEF
-                               = "(.*>)?(?<file>.*)\\((?<line>\\d+)\\)\\x20:\\x20warning\\x20"
-                                   + "(?<id>C\\d+):(?<message>.*)";
+                               = "(?>[^>]*+>)?(?<file>(?>[^\\\\]{1,260}\\\\)*[^\\\\]{1,260})"
+                                   + "\\((?<line>\\d{1,5})\\)\\x20?:\\x20warning\\x20(?<id>C\\d{4,5}):"
+                                   + "\\x20?(?<message>.*)";
 
   public static List<PropertyDefinition> properties() {
     var category = "CXX External Analyzers";
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
@@ -35,8 +35,9 @@
 public final class DrMemoryParser {
 
   private static final Logger LOG = Loggers.get(DrMemoryParser.class);
-  private static final Pattern RX_MESSAGE_FINDER = Pattern.compile("^Error #\\d+:(.*)");
-  private static final Pattern RX_FILE_FINDER = Pattern.compile("^.*\\[(.*):(\\d+)\\]$");
+  private static final Pattern RX_MESSAGE_FINDER = Pattern.compile("^Error #\\d{1,6}:(.*)");
+  private static final Pattern RX_FILE_FINDER = Pattern.compile(
+    "\\s*+#[^\\[]++\\[((?>[a-zA-Z]:\\\\)??[^:]++):(\\d{1,5})\\]");
   private static final int TOP_COUNT = 4;
 
   private DrMemoryParser() {
@@ -90,7 +91,7 @@ public static List<DrMemoryError> parse(File file, String encoding) {
   public static List<String> getElements(File file, String encoding) {
 
     var list = new ArrayList<String>();
-    try ( var br = new BufferedReader(
+    try (var br = new BufferedReader(
       new InputStreamReader(java.nio.file.Files.newInputStream(file.toPath()), encoding))) {
       var sb = new StringBuilder(4096);
       String line;
diff --git a/cxx-sensors/src/main/java/org/sonar/cxx/sensors/pclint/CxxPCLintSensor.java b/cxx-sensors/src/main/java/org/sonar/cxx/sensors/pclint/CxxPCLintSensor.java
@@ -60,8 +60,8 @@ public class CxxPCLintSensor extends CxxIssuesReportSensor {
 
   private static final String PREFIX_DURING_SPECIFIC_WALK_MSG = "during specific walk";
 
-  private static final Pattern SUPPLEMENTAL_MSG_PATTERN
-                                 = Pattern.compile(PREFIX_DURING_SPECIFIC_WALK_MSG + "\\s+(.+):(\\d+):(\\d+)\\s+.+");
+  private static final Pattern SUPPLEMENTAL_MSG_PATTERN = Pattern.compile(
+    PREFIX_DURING_SPECIFIC_WALK_MSG + "\\s+([^:]++):(\\d{1,5}):(\\d{1,5}).+");
 
   public static List<PropertyDefinition> properties() {
     return Collections.unmodifiableList(Arrays.asList(
diff --git a/cxx-sensors/src/main/java/org/sonar/cxx/sensors/tests/dotnet/VisualStudioTestResultsFileParser.java b/cxx-sensors/src/main/java/org/sonar/cxx/sensors/tests/dotnet/VisualStudioTestResultsFileParser.java
@@ -48,7 +48,7 @@ private static class Parser {
     private final File file;
     private final UnitTestResults unitTestResults;
     private final DateFormat dateFormat = new SimpleDateFormat("yyyy-MM-dd'T'HH:mm:ss.SSSXXX");
-    private final Pattern millisecondsPattern = Pattern.compile("(\\.([0-9]{0,3}))[0-9]*+");
+    private final Pattern millisecondsPattern = Pattern.compile("(\\.(\\d{0,3}))\\d*+");
 
     private boolean foundCounters;
 
@@ -62,7 +62,7 @@ private static void checkRootTag(XmlParserHelper xmlParserHelper) {
     }
 
     public void parse() {
-      try ( var xmlParserHelper = new XmlParserHelper(file)) {
+      try (var xmlParserHelper = new XmlParserHelper(file)) {
         checkRootTag(xmlParserHelper);
         dispatchTags(xmlParserHelper);
         if (!foundCounters) {
diff --git a/cxx-sensors/src/test/java/org/sonar/cxx/sensors/compiler/vc/CxxCompilerVcSensorTest.java b/cxx-sensors/src/test/java/org/sonar/cxx/sensors/compiler/vc/CxxCompilerVcSensorTest.java
@@ -77,7 +77,7 @@ public void shouldReportBCorrectVcViolations() {
     settings.setProperty(CxxCompilerVcSensor.REPORT_PATH_KEY, "compiler-reports/VC-report.vclog");
     settings.setProperty(CxxCompilerVcSensor.REPORT_ENCODING_DEF, StandardCharsets.UTF_8.name());
     settings.setProperty(CxxCompilerVcSensor.REPORT_REGEX_DEF,
-                         ".*>(?<file>.*)\\((?<line>\\d+)\\):\\x20warning\\x20(?<id>C\\d+):(?<message>.*)");
+                         "[^>]*+>(?<file>.*)\\((?<line>\\d{1,5})\\):\\x20warning\\x20(?<id>C\\d{4,5}):(?<message>.*)");
     context.setSettings(settings);
 
     context.fileSystem().add(TestInputFileBuilder.create("ProjectKey", "Server/source/zip/zipmanager.cpp")
diff --git a/cxx-squid/src/main/java/org/sonar/cxx/config/MsBuild.java b/cxx-squid/src/main/java/org/sonar/cxx/config/MsBuild.java
@@ -38,7 +38,7 @@
  */
 public class MsBuild {
 
- /**
+  /**
    * the following settings are in use by the feature to read configuration settings from the VC compiler report
    */
   public static final String REPORT_PATH_KEY = "sonar.cxx.msbuild.reportPaths";
@@ -54,18 +54,22 @@ public class MsBuild {
   private static final String CPPVERSION = "__cplusplus=199711L";
 
   private static final Pattern[] INCLUDE_PATTERNS = {Pattern.compile("/I\"(.*?)\""),
-                                                     Pattern.compile("/I([^\\s\"]+) ")};
-  private static final Pattern[] DEFINE_PATTERNS = {Pattern.compile("[/-]D\\s([^\\s]+)"),
-                                                    Pattern.compile("[/-]D([^\\s]+)")};
-  private static final Pattern PATH_TO_CL_PATTERN = Pattern.compile("^.*\\\\bin\\\\.*CL.exe\\x20.*$");
+                                                     Pattern.compile("/I([^\\s\"]++) ")};
+  private static final Pattern[] DEFINE_PATTERNS = {Pattern.compile("[/-]D\\s([^\\s]++)"),
+                                                    Pattern.compile("[/-]D([^\\s]++)")};
+  private static final Pattern PATH_TO_CL_PATTERN = Pattern.compile(
+    "^(?>[^\\\\]{0,260}\\\\)+bin\\\\(?>[^\\\\]{1,260}\\\\)*CL.exe\\x20.*$");
   private static final Pattern PLATFORM_X86_PATTERN = Pattern.compile("Building solution configuration \".*\\|x64\".");
-  private static final Pattern TOOLSET_V141_PATTERN = Pattern
-    .compile("^.*VC\\\\Tools\\\\MSVC\\\\14\\.1\\d\\.\\d+\\\\bin\\\\HostX(86|64)\\\\x(86|64)\\\\CL.exe.*$");
-  private static final Pattern TOOLSET_V142_PATTERN = Pattern
-    .compile("^.*VC\\\\Tools\\\\MSVC\\\\14\\.2\\d\\.\\d+\\\\bin\\\\HostX(86|64)\\\\x(86|64)\\\\CL.exe.*$");
+  private static final Pattern TOOLSET_V141_PATTERN = Pattern.compile(
+    "^(?>[^\\\\]{0,260}\\\\)+VC\\\\Tools\\\\MSVC\\\\14\\.1\\d\\.\\d{1,6}"
+      + "\\\\bin\\\\HostX(86|64)\\\\x(86|64)\\\\CL.exe.*$");
+  private static final Pattern TOOLSET_V142_PATTERN = Pattern.compile(
+    "^(?>[^\\\\]{0,260}\\\\)+VC\\\\Tools\\\\MSVC\\\\14\\.2\\d\\.\\d{1,6}"
+      + "\\\\bin\\\\HostX(86|64)\\\\x(86|64)\\\\CL.exe.*$");
 
   // It seems that the required line in any language has these elements: "ClCompile" and (*.vcxproj)
-  private static final Pattern PATH_TO_VCXPROJ = Pattern.compile("^(?:\\S+)\\s(?:\"ClCompile\").*\"(.+vcxproj)\".*$");
+  private static final Pattern PATH_TO_VCXPROJ = Pattern.compile(
+    "^\\S+\\s\\\"ClCompile\\\".+\\\"((?>[^\\\\]{1,260}\\\\)*[^\\\\]{1,260}\\.vcxproj)\\\".*$");
 
   private String platformToolset = "V120";
   private String platform = "Win32";
@@ -75,7 +79,7 @@ public class MsBuild {
   /**
    * CxxVCppBuildLogParser (ctor)
    *
-   * @param db
+   * @param squidConfig
    */
   public MsBuild(CxxSquidConfiguration squidConfig) {
     this.squidConfig = squidConfig;
@@ -120,7 +124,7 @@ public void parse(String line, String projectPath, String compilationFile) {
    * @param encodingName
    */
   public void parse(File buildLog, String baseDir, String encodingName) {
-     LOG.info("Processing MsBuild log '{}', Encoding= '{}'", buildLog.getName(), encodingName);
+    LOG.info("Processing MsBuild log '{}', Encoding= '{}'", buildLog.getName(), encodingName);
 
     var detectedPlatform = false;
     try (var br = new BufferedReader(new InputStreamReader(java.nio.file.Files.newInputStream(buildLog.toPath()),
@@ -133,7 +137,7 @@ public void parse(File buildLog, String baseDir, String encodingName) {
         if (line.trim().startsWith("INCLUDE=")) { // handle environment includes
           String[] includes = line.split("=")[1].split(";");
           for (var include : includes) {
-            squidConfig.add(CxxSquidConfiguration.GLOBAL, CxxSquidConfiguration.INCLUDE_DIRECTORIES,include);
+            squidConfig.add(CxxSquidConfiguration.GLOBAL, CxxSquidConfiguration.INCLUDE_DIRECTORIES, include);
           }
         }
 
@@ -296,7 +300,7 @@ private void parseInclude(String element, String project, String fileElement) {
           includeRoot = new File(project, includeRoot.getPath());
         }
       }
-      squidConfig.add(fileElement, CxxSquidConfiguration.INCLUDE_DIRECTORIES,includeRoot.getCanonicalPath());
+      squidConfig.add(fileElement, CxxSquidConfiguration.INCLUDE_DIRECTORIES, includeRoot.getCanonicalPath());
     } catch (IOException e) {
       LOG.error("Cannot parse include path using element '{}' : '{}'", element, e.getMessage());
     }
diff --git a/cxx-squid/src/main/java/org/sonar/cxx/parser/CxxLexer.java b/cxx-squid/src/main/java/org/sonar/cxx/parser/CxxLexer.java
@@ -47,12 +47,12 @@ public final class CxxLexer {
   private static final String HEX_PREFIX = "0[xX]";
   private static final String BIN_PREFIX = "0[bB]";
   private static final String EXPONENT = "[Ee][+-]?+[0-9_]([']?+[0-9_]++)*+";
-  private static final String BINARY_EXPONENT = "[pP][+-]?+[0-9]([']?+[0-9]++)*+"; // since C++17
+  private static final String BINARY_EXPONENT = "[pP][+-]?+\\d([']?+\\d++)*+"; // since C++17
   //private static final String INTEGER_SUFFIX = "(((U|u)(i64|LL|ll|L|l)?)|((i64|LL|ll|L|l)(u|U)?))";
   //private static final String FLOAT_SUFFIX = "(f|l|F|L)";
   // ud-suffix: identifier (including INTEGER_SUFFIX, FLOAT_SUFFIX)
-  private static final String UD_SUFFIX = "[_a-zA-Z][_a-zA-Z0-9]*+";
-  private static final String DECDIGIT_SEQUENCE = "[0-9]([']?+[0-9]++)*+";
+  private static final String UD_SUFFIX = "[_a-zA-Z]\\w*+";
+  private static final String DECDIGIT_SEQUENCE = "\\d([']?+\\d++)*+";
   private static final String HEXDIGIT_SEQUENCE = "[0-9a-fA-F]([']?+[0-9a-fA-F]++)*+";
   private static final String BINDIGIT_SEQUENCE = "[01]([']?+[01]++)*+";
   private static final String POINT = "\\.";
diff --git a/cxx-squid/src/main/java/org/sonar/cxx/preprocessor/CppLexer.java b/cxx-squid/src/main/java/org/sonar/cxx/preprocessor/CppLexer.java
@@ -43,12 +43,12 @@ public final class CppLexer {
   private static final String HEX_PREFIX = "0[xX]";
   private static final String BIN_PREFIX = "0[bB]";
   private static final String EXPONENT = "[eE][+-]?+[0-9_]([']?+[0-9_]++)*+";
-  private static final String BINARY_EXPONENT = "[pP][+-]?+[0-9]([']?+[0-9]++)*+"; // since C++17
+  private static final String BINARY_EXPONENT = "[pP][+-]?+\\d([']?+\\d++)*+"; // since C++17
   //private static final String INTEGER_SUFFIX = "(((U|u)(LL|ll|L|l)?)|((LL|ll|L|l)(u|U)?))";
   //private static final String FLOAT_SUFFIX = "(f|l|F|L)";
   // ud-suffix: identifier (including INTEGER_SUFFIX, FLOAT_SUFFIX)
-  private static final String UD_SUFFIX = "[_a-zA-Z][_a-zA-Z0-9]*+";
-  private static final String DECDIGIT_SEQUENCE = "[0-9]([']?+[0-9]++)*+";
+  private static final String UD_SUFFIX = "[_a-zA-Z]\\w*+";
+  private static final String DECDIGIT_SEQUENCE = "\\d([']?+\\d++)*+";
   private static final String HEXDIGIT_SEQUENCE = "[0-9a-fA-F]([']?+[0-9a-fA-F]++)*+";
   private static final String BINDIGIT_SEQUENCE = "[01]([']?+[01]++)*+";
   private static final String POINT = "\\.";
