Fix MavenPomDownloader.htmlIndexToVersioning() to handle HTML title attributes (#6745)

timtebeek · Tim te Beek · claude · web-flow · commit 0cee24399102 · 2026-02-15T16:13:35.000+01:00
When parsing HTML directory listings, the method searched for "\"&gt;" to delimit the
href attribute value. This overshoots when anchor tags include additional attributes
like title="...", causing corrupted version strings such as "1.0.0/\" title=\"1.0.0"
instead of "1.0.0/". The fix changes the search to find the closing quote directly,
which correctly terminates at the href attribute boundary regardless of subsequent
attributes. Includes a test case covering this scenario.

Co-authored-by: Tim te Beek &lt;tim@mac.home&gt;
Co-authored-by: Claude Haiku 4.5 &lt;noreply@anthropic.com&gt;
diff --git a/rewrite-maven/src/main/java/org/openrewrite/maven/internal/MavenPomDownloader.java b/rewrite-maven/src/main/java/org/openrewrite/maven/internal/MavenPomDownloader.java
@@ -411,7 +411,7 @@ public MavenMetadata downloadMetadata(GroupArtifactVersion gav, @Nullable Resolv
         int start = responseBody.indexOf("<a href=\"");
         while (start > 0) {
             start += 9;
-            int end = responseBody.indexOf("\">", start);
+            int end = responseBody.indexOf("\"", start);
             if (end < 0) {
                 break;
             }
diff --git a/rewrite-maven/src/test/java/org/openrewrite/maven/internal/MavenPomDownloaderTest.java b/rewrite-maven/src/test/java/org/openrewrite/maven/internal/MavenPomDownloaderTest.java
@@ -626,6 +626,42 @@ void deriveMetaDataFromHtmlBasedRepository() {
               new MavenPomDownloader(emptyMap(), ctx).downloadMetadata(new GroupArtifact("does.definitely.not", "exist"), null, List.of(repository)));
         }
 
+        @Issue("https://github.com/openrewrite/rewrite/issues/6739")
+        @Test
+        void deriveMetaDataFromHtmlWithTitleAttributes() throws Exception {
+            try (MockWebServer server = new MockWebServer()) {
+                server.setDispatcher(new Dispatcher() {
+                    @Override
+                    public MockResponse dispatch(RecordedRequest request) {
+                        if (request.getPath() != null && request.getPath().endsWith("maven-metadata.xml")) {
+                            return new MockResponse().setResponseCode(404);
+                        }
+                        return new MockResponse().setResponseCode(200).setBody(
+                          """
+                            <html><body>
+                            <a href="../" title="../">../</a>
+                            <a href="1.0.0/" title="1.0.0/">1.0.0/</a>
+                            <a href="1.1.0/" title="1.1.0/">1.1.0/</a>
+                            <a href="2.0.0/" title="2.0.0/">2.0.0/</a>
+                            </body></html>
+                            """
+                        );
+                    }
+                });
+                server.start();
+
+                MavenRepository repository = MavenRepository.builder()
+                  .id("html-with-title")
+                  .uri("http://%s:%d/".formatted(server.getHostName(), server.getPort()))
+                  .knownToExist(true)
+                  .deriveMetadataIfMissing(true)
+                  .build();
+                MavenMetadata metaData = new MavenPomDownloader(emptyMap(), ctx)
+                  .downloadMetadata(new GroupArtifact("fred", "fred"), null, List.of(repository));
+                assertThat(metaData.getVersioning().getVersions()).containsExactlyInAnyOrder("1.0.0", "1.1.0", "2.0.0");
+            }
+        }
+
         @SuppressWarnings("ConstantConditions")
         @Test
         void mergeMetadata() throws Exception {

Original file line number	Diff line number	Diff line change
`@@ -411,7 +411,7 @@ public MavenMetadata downloadMetadata(GroupArtifactVersion gav, @Nullable Resolv`
`411`	`411`	`int start = responseBody.indexOf("<a href=\"");`
`412`	`412`	`while (start > 0) {`
`413`	`413`	`start += 9;`
`414`		`- int end = responseBody.indexOf("\">", start);`
	`414`	`+ int end = responseBody.indexOf("\"", start);`
`415`	`415`	`if (end < 0) {`
`416`	`416`	`break;`
`417`	`417`	`}`