pref: remove html tags

Author: wenytang-ms

jdtls.ext/com.microsoft.jdtls.ext.core/src/com/microsoft/jdtls/ext/core/parser/ContextResolver.java

@@ -900,6 +900,7 @@ private static String extractRelevantJavaDocContent(org.eclipse.jdt.core.IType t
 
             // Clean Javadoc comment for processing
             String cleanedJavadoc = cleanJavadocComment(rawJavadoc);
+            cleanedJavadoc = removeHtmlTags(cleanedJavadoc);
             cleanedJavadoc = convertHtmlEntities(cleanedJavadoc);
 
             // === High Priority: Extract class description text (first paragraph) ===
@@ -1036,6 +1037,29 @@ private static String convertHtmlEntities(String text) {
                    .replace("–", "-");
     }
 
+    /**
+     * Remove all HTML tags from text, keeping only plain text content.
+     * Preserves line breaks for block-level tags like <p>, <br>, <div>.
+     */
+    private static String removeHtmlTags(String text) {
+        if (text == null || text.isEmpty()) {
+            return text;
+        }
+        
+        // Replace block-level tags with line breaks
+        text = text.replaceAll("(?i)</(p|div|li)>|<br\\s*/?>|<p[^>]*>", "\n");
+        
+        // Remove all remaining HTML tags
+        text = text.replaceAll("<[^>]+>", "");
+        
+        // Clean up whitespace: collapse spaces, trim lines, limit line breaks
+        text = text.replaceAll("[ \\t]+", " ")
+                   .replaceAll(" *\\n *", "\n")
+                   .replaceAll("\\n{3,}", "\n\n");
+        
+        return text.trim();
+    }
+
     /**
      * Extract method JavaDoc content directly for LLM consumption.
      * Returns cleaned JavaDoc without artificial truncation - let LLM understand the full context.
@@ -1056,6 +1080,7 @@ private static String extractMethodJavaDocSummary(IMethod method) {
 
             // Just clean and return - let LLM understand the full context
             String cleaned = cleanJavadocComment(rawJavadoc);
+            cleaned = removeHtmlTags(cleaned);
             return convertHtmlEntities(cleaned);
 
         } catch (Exception e) {
@@ -1176,6 +1201,7 @@ private static String extractFieldJavaDocSummary(org.eclipse.jdt.core.IField fie
 
             // Just clean and return - let LLM understand the full context
             String cleaned = cleanJavadocComment(rawJavadoc);
+            cleaned = removeHtmlTags(cleaned);
             return convertHtmlEntities(cleaned);
 
         } catch (Exception e) {

src/copilot/contextProvider.ts

@@ -44,6 +44,7 @@ export async function registerCopilotContextProviders(
             "status": "succeeded",
             "installCount": installCount
         });
+        console.log(`Registered Java Copilot context provider on ${installCount} API(s).`);
     }
     catch (error) {
         const errorMessage = (error as Error).message || "unknown_error";
@@ -104,7 +105,7 @@ async function resolveJavaContext(request: ResolveRequest, copilotCancel: vscode
         );
         JavaContextProviderUtils.checkCancellation(copilotCancel);
         items.push(...projectDependencyItems);
-
+        
         JavaContextProviderUtils.checkCancellation(copilotCancel);
 
         // Resolve local imports and convert to context items
@@ -113,6 +114,7 @@ async function resolveJavaContext(request: ResolveRequest, copilotCancel: vscode
             copilotCancel,
             JavaContextProviderUtils.checkCancellation
         );
+        console.dir(localImportItems)
         JavaContextProviderUtils.checkCancellation(copilotCancel);
         items.push(...localImportItems);
     } catch (error: any) {