Bug 1483458 - Change HTMLParser to look at .py atom files. r=hsivonen,heycam.

Author: Jonathan Kingston

translator-src/nu/validator/htmlparser/cpptranslate/CppTypes.java

@@ -55,7 +55,9 @@
 
 public class CppTypes {
 
-    private static final Pattern ATOM_DEF = Pattern.compile("^GK_ATOM\\(([^,]+),\\s*\"([^\"]*)\"\\).*$");
+    /* Please note we aren't looking for the following Atom definitions:
+       PseudoElementAtom or NonInheritingAnonBoxAtom or InheritingAnonBoxAtom */
+    private static final Pattern ATOM_DEF = Pattern.compile("^\\s*Atom\\(\"([^,]+)\",\\s*\"([^\"]*)\"\\).*$");
 
     private static Set<String> reservedWords = new HashSet<String>();
 
@@ -122,17 +124,15 @@ public class CppTypes {
 
     private final Writer atomWriter;
 
-    private StringBuilder manualAtoms = new StringBuilder();
-
-    public CppTypes(File atomList) {
+    public CppTypes(File atomList, File generatedAtomFile) {
         if (atomList == null) {
             atomWriter = null;
         } else {
             try {
                 ingestAtoms(atomList);
                 atomWriter = new OutputStreamWriter(new FileOutputStream(
-                        atomList), "utf-8");
-                atomWriter.append(manualAtoms);
+                        generatedAtomFile), "utf-8");
+                this.start();
             } catch (IOException e) {
                 throw new RuntimeException(e);
             }
@@ -145,13 +145,19 @@ private void ingestAtoms(File atomList) throws IOException {
                 new InputStreamReader(new FileInputStream(atomList), "utf-8"));
         try {
             String line;
+            boolean startedParsing = false;
             while ((line = atomReader.readLine()) != null) {
-                manualAtoms.append(line);
-                manualAtoms.append('\n');
-                if (line.startsWith("// BEGIN GENERATED")) {
+                // only start parsing lines after this comment
+                if (line.trim().startsWith("# START ATOMS")) {
+                    startedParsing = true;
+                } else if (!startedParsing) {
+                    continue;
+                }
+                // stop parsing lines after this comment
+                if (line.trim().startsWith("# END ATOMS")) {
                     return;
                 }
-                if (!line.startsWith("GK_ATOM")) {
+                if (!line.trim().startsWith("Atom")) {
                     continue;
                 }
                 Matcher m = ATOM_DEF.matcher(line);
@@ -167,10 +173,23 @@ private void ingestAtoms(File atomList) throws IOException {
         }
     }
 
+    public void start() {
+        try {
+
+            if (atomWriter != null) {
+                atomWriter.write("# THIS FILE IS GENERATED BY THE HTML PARSER TRANSLATOR AND WILL BE OVERWRITTEN!\n");
+                atomWriter.write("from Atom import Atom\n\n");
+                atomWriter.write("HTML_PARSER_ATOMS = [\n");
+            }
+        } catch (IOException e) {
+            throw new RuntimeException(e);
+        }
+    }
+
     public void finished() {
         try {
             if (atomWriter != null) {
-                atomWriter.write("// END GENERATED ATOMS, DO NOT ADD CODE BELOW THIS LINE\n");
+                atomWriter.write("]\n");
                 atomWriter.flush();
                 atomWriter.close();
             }
@@ -303,8 +322,8 @@ public String localForLiteral(String literal) {
             atomMap.put(literal, atom);
             if (atomWriter != null) {
                 try {
-                    atomWriter.write("// ATOM GENERATED BY HTML PARSER TRANSLATOR (WILL BE AUTOMATICALLY OVERWRITTEN):\nGK_ATOM(" + atom + ", \"" + literal
-                            + "\")\n");
+                    atomWriter.write("    # ATOM GENERATED BY HTML PARSER TRANSLATOR (WILL BE AUTOMATICALLY OVERWRITTEN):\n    Atom(\"" + atom + "\", \"" + literal
+                            + "\"),\n");
                 } catch (IOException e) {
                     throw new RuntimeException(e);
                 }

translator-src/nu/validator/htmlparser/cpptranslate/GkAtomParser.java

@@ -47,7 +47,9 @@
 
 public class GkAtomParser {
 
-    private static final Pattern ATOM = Pattern.compile("^GK_ATOM\\(([^,]+),\\s*\"([^\"]*)\"\\).*$");
+    /* Please note we aren't looking for the following Atom definitions:
+       PseudoElementAtom or NonInheritingAnonBoxAtom or InheritingAnonBoxAtom */
+    private static final Pattern ATOM = Pattern.compile("^Atom\\(\"([^,]+)\",\\s*\"([^\"]*)\"\\).*$");
 
     private final BufferedReader reader;
 
@@ -59,7 +61,7 @@ public Map<String, String> parse() throws IOException {
         Map<String, String> map = new HashMap<String, String>();
         String line;
         while((line = reader.readLine()) != null) {
-            Matcher m = ATOM.matcher(line);
+            Matcher m = ATOM.matcher(line.trim());
             if (m.matches()) {
                 map.put(m.group(2), m.group(1));
             }

translator-src/nu/validator/htmlparser/cpptranslate/Main.java

@@ -80,7 +80,7 @@ public class Main {
      * @throws IOException
      */
     public static void main(String[] args) throws ParseException, IOException {
-        CppTypes cppTypes = new CppTypes(new File(args[2]));
+        CppTypes cppTypes = new CppTypes(new File(args[2]), new File(args[3]));
         SymbolTable symbolTable = new SymbolTable();
 
         File javaDirectory = new File(args[0]);

translator-src/nu/validator/htmlparser/generator/GenerateNamedCharactersCpp.java

@@ -141,7 +141,7 @@ public static void main(String[] args) throws IOException {
             }
         }
 
-        CppTypes cppTypes = new CppTypes(null);
+        CppTypes cppTypes = new CppTypes(null, null);
         File targetDirectory = new File(args[1]);
 
         generateH(targetDirectory, cppTypes, entities);