Another fix to really skip comments.

[DictionaryPC.git] / src / com / hughes / android / dictionary / parser / WikiTokenizer.java

diff --git a/src/com/hughes/android/dictionary/parser/WikiTokenizer.java b/src/com/hughes/android/dictionary/parser/WikiTokenizer.java
index 8cf882e7de6933ea77dc19c5f39396146bd3f815..83835d95fad90311872adb26d7927ed655b254de 100644 (file)
--- a/src/com/hughes/android/dictionary/parser/WikiTokenizer.java
+++ b/src/com/hughes/android/dictionary/parser/WikiTokenizer.java
@@ -481,15 +481,18 @@ public final class WikiTokenizer {
         int firstNewline = -1;
         int[] nextMatch = new int[8];
         for (int i = 0; i < 8; ++i) {
-            nextMatch[i] = wikiText.indexOf(patterns[i], start);
-            if (nextMatch[i] == -1) nextMatch[i] = i > 0 ? 0x7fffffff : wikiText.length();
+            nextMatch[i] = -2;
         }
         while (end < wikiText.length()) {
             // Manual replacement for matcher.find(end),
             // because Java regexp is a ridiculously slow implementation.
             // Initialize to always match the end.
             int matchIdx = 0;
-            for (int i = 1; i < 8; ++i) {
+            for (int i = 0; i < 8; ++i) {
+                if (nextMatch[i] <= end) {
+                    nextMatch[i] = wikiText.indexOf(patterns[i], end);
+                    if (nextMatch[i] == -1) nextMatch[i] = i > 0 ? 0x7fffffff : wikiText.length();
+                }
                 if (nextMatch[i] < nextMatch[matchIdx]) {
                     matchIdx = i;
                 }
@@ -498,8 +501,6 @@ public final class WikiTokenizer {
             int matchStart = nextMatch[matchIdx];
             String matchText = patterns[matchIdx];
             int matchEnd = matchStart + matchText.length();
-            nextMatch[matchIdx] = wikiText.indexOf(patterns[matchIdx], matchEnd);
-            if (nextMatch[matchIdx] == -1) nextMatch[matchIdx] = matchIdx > 0 ? 0x7fffffff : wikiText.length();
             if (matchIdx == 0) {
                 matchText = "";
                 matchEnd = matchStart;
@@ -527,14 +528,14 @@ public final class WikiTokenizer {
                 if (tokenStack.size() > 0) {
                     final String removed = tokenStack.remove(tokenStack.size() - 1);
                     if (removed.equals("{{") && !matchText.equals("}}")) {
-                        errors.add("Unmatched {{ error: " + wikiText.substring(start));
+                        errors.add("Unmatched {{ error: " + wikiText.substring(start, matchEnd));
                         return safeIndexOf(wikiText, start, "\n", "\n");
                     } else if (removed.equals("[[") && !matchText.equals("]]")) {
-                        errors.add("Unmatched [[ error: " + wikiText.substring(start));
+                        errors.add("Unmatched [[ error: " + wikiText.substring(start, matchEnd));
                         return safeIndexOf(wikiText, start, "\n", "\n");
                     }
                 } else {
-                    errors.add("Pop too many error: " + wikiText.substring(start).replace("\n", "\\\\n"));
+                    errors.add("Pop too many " + matchText + " error: " + wikiText.substring(start, matchEnd).replace("\n", "\\\\n"));
                     // If we were looking for a newline
                     return safeIndexOf(wikiText, start, "\n", "\n");
                 }
@@ -549,7 +550,7 @@ public final class WikiTokenizer {
                 // Do nothing.  These can match spuriously, and if it's not the thing
                 // we're looking for, keep on going.
             } else if (matchText.equals("<!--")) {
-                end = wikiText.indexOf("-->");
+                end = wikiText.indexOf("-->", matchStart);
                 if (end == -1) {
                     errors.add("Unmatched <!-- error: " + wikiText.substring(start));
                     return safeIndexOf(wikiText, start, "\n", "\n");