(build) Upgrade JIB to 3.4.5

Since the crawl tasks are started in two phases, first when generating them in one loop, and then in a second loop that drains the task list; if the first loop contains a long-running crawl task that is triggered late, the rest of the crawl may halt until that task is finish.

Fixed the problem by draining and re-trying also in the first loop.

build.gradle

@@ -5,7 +5,7 @@ plugins {
 
     // This is a workaround for a bug in the Jib plugin that causes it to stall randomly
     // https://github.com/GoogleContainerTools/jib/issues/3347
-    id 'com.google.cloud.tools.jib' version '3.4.4' apply(false)
+    id 'com.google.cloud.tools.jib' version '3.4.5' apply(false)
 }
 
 group 'marginalia'
@@ -47,7 +47,7 @@ ext {
     dockerImageBase='container-registry.oracle.com/graalvm/jdk:24'
     dockerImageTag='latest'
     dockerImageRegistry='marginalia'
-    jibVersion = '3.4.4'
+    jibVersion = '3.4.5'
 }
 
 idea {

code/functions/live-capture/java/nu/marginalia/rss/svc/FeedFetcherService.java

@@ -229,13 +229,15 @@ public class FeedFetcherService {
                     .timeout(Duration.ofSeconds(15))
                     ;
 
-            if (ifModifiedSinceDate != null) {
+            // Set the If-Modified-Since or If-None-Match headers if we have them
+            // though since there are certain idiosyncrasies in server implementations,
+            // we avoid setting both at the same time as that may turn a 304 into a 200.
+            if (ifNoneMatchTag != null) {
+                requestBuilder.header("If-None-Match", ifNoneMatchTag);
+            } else if (ifModifiedSinceDate != null) {
                 requestBuilder.header("If-Modified-Since", ifModifiedSinceDate);
             }
 
-            if (ifNoneMatchTag != null) {
-                requestBuilder.header("If-None-Match", ifNoneMatchTag);
-            }
 
             HttpRequest getRequest = requestBuilder.build();
 

code/processes/crawling-process/java/nu/marginalia/crawl/CrawlerMain.java

@@ -264,17 +264,16 @@ public class CrawlerMain extends ProcessMainClass {
                 if (workLog.isJobFinished(crawlSpec.domain))
                     continue;
 
-                var task = new CrawlTask(
+                var task = new CrawlTask(crawlSpec, anchorTagsSource, outputDir, warcArchiver, domainStateDb, workLog);
-                        crawlSpec,
-                        anchorTagsSource,
-                        outputDir,
-                        warcArchiver,
-                        domainStateDb,
-                        workLog);
 
                 // Try to run immediately, to avoid unnecessarily keeping the entire work set in RAM
                 if (!trySubmitDeferredTask(task)) {
-                    // Otherwise add to the taskList for deferred execution
+
+                    // Drain the retry queue to the taskList, and try to submit any tasks that are in the retry queue
+                    retryQueue.drainTo(taskList);
+                    taskList.removeIf(this::trySubmitDeferredTask);
+
+                    // Then add this new task to the retry queue
                     taskList.add(task);
                 }
             }

code/processes/crawling-process/java/nu/marginalia/crawl/fetcher/ContentTags.java

@@ -19,11 +19,13 @@ public record ContentTags(String etag, String lastMod) {
     /** Paints the tags onto the request builder. */
     public void paint(HttpGet request) {
 
+        // Paint the ETag header if present,
+        // otherwise paint the Last-Modified header
+        // (but not both at the same time due to some servers not liking it)
+
         if (etag != null) {
             request.addHeader("If-None-Match", etag);
-        }
+        } else if (lastMod != null) {
-
-        if (lastMod != null) {
             request.addHeader("If-Modified-Since", lastMod);
         }
     }

code/services-application/api-service/build.gradle

@@ -3,7 +3,7 @@ plugins {
 
     id 'application'
     id 'jvm-test-suite'
-    id 'com.google.cloud.tools.jib' version '3.4.4'
+    id 'com.google.cloud.tools.jib' version '3.4.5'
 }
 
 java {

code/services-application/dating-service/build.gradle

@@ -3,7 +3,7 @@ plugins {
 
     id 'application'
     id 'jvm-test-suite'
-    id 'com.google.cloud.tools.jib' version '3.4.4'
+    id 'com.google.cloud.tools.jib' version '3.4.5'
 }
 
 application {

code/services-application/explorer-service/build.gradle

@@ -3,7 +3,7 @@ plugins {
 
     id 'application'
     id 'jvm-test-suite'
-    id 'com.google.cloud.tools.jib' version '3.4.4'
+    id 'com.google.cloud.tools.jib' version '3.4.5'
 }
 
 application {

code/services-application/search-service-legacy/build.gradle

@@ -5,7 +5,7 @@ plugins {
     id 'application'
     id 'jvm-test-suite'
 
-    id 'com.google.cloud.tools.jib' version '3.4.4'
+    id 'com.google.cloud.tools.jib' version '3.4.5'
 }
 
 application {

code/services-application/search-service/build.gradle

@@ -3,7 +3,7 @@ plugins {
     id 'application'
     id 'jvm-test-suite'
     id 'gg.jte.gradle' version '3.1.15'
-    id 'com.google.cloud.tools.jib' version '3.4.4'
+    id 'com.google.cloud.tools.jib' version '3.4.5'
 }
 
 application {

code/services-application/search-service/resources/jte/part/head.jte

@@ -26,4 +26,10 @@
 
     <link rel="search" type="application/opensearchdescription+xml" href="/opensearch.xml" title="Marginalia">
 
-</head>
+</head>
+<noscript>
+    <h1>Users of text-based browsers</h1>
+    <p>Consider using the old interface at <a href="https://old-search.marginalia.nu/">https://old-search.marginalia.nu/</a>,
+    as it uses fewer modern CSS tricks, and should work better than the new UI.  It's functionally nearly identical, but just renders it using a different layout.</p>
+    <hr>
+</noscript>

code/services-application/search-service/resources/jte/serp/error.jte

@@ -26,7 +26,7 @@
 
     <!-- Main content -->
     <main class="flex-1 p-4 max-w-2xl space-y-4">
-        <div class="border dark:border-gray-600 rounded bg-white text-black dark:bg-gray-800 dark:text-white text-m p-4">
+        <div class="border border-gray-300 dark:border-gray-600 rounded bg-white text-black dark:bg-gray-800 dark:text-white text-m p-4">
             <div class="flex space-x-3  place-items-baseline">
                 <i class="fa fa-circle-exclamation text-red-800"></i>
                 <div class="grow">${model.errorTitle()}</div>

code/services-application/search-service/resources/jte/serp/part/footerHowto.jte

@@ -80,10 +80,6 @@
             <tr><td>rank&gt;50</td><td>The ranking of the website is at least 50 in a span of 1 - 255</td></tr>
             <tr><td>rank&lt;50</td><td>The ranking of the website is at most 50 in a span of 1 - 255</td></tr>
 
-            <tr><td>count&gt;10</td><td> The search term must appear in at least 10 results form the domain</td></tr>
-            <tr><td>count&lt;10</td><td> The search term must appear in at most 10 results from the domain</td></tr>
-
-
             <tr><td>format:html5</td><td>Filter documents using the HTML5 standard. This is typically modern websites.</td></tr>
             <tr><td>format:xhtml</td><td>Filter documents using the XHTML standard</td></tr>
             <tr><td>format:html123</td><td>Filter documents using the HTML standards 1, 2, and 3. This is typically very old websites. </td></tr>

code/services-application/search-service/resources/jte/serp/part/searchform.jte

@@ -7,7 +7,7 @@
 
 <form class="flex-1 max-w-2xl" action="/search">
     <div class="flex">
-        @if (query.isBlank())
+        @if (query != null && query.isBlank())
         <%-- Add autofocus if the query is blank --%>
             <input type="text"
                    class="shadow-inner flex-1 dark:bg-black dark:text-gray-100 bg-gray-50 border dark:border-gray-600 border-gray-300 text-gray-900 text-sm rounded-sm block w-full p-2.5"

code/services-application/status-service/build.gradle

@@ -2,7 +2,7 @@ plugins {
     id 'java'
     id 'application'
     id 'jvm-test-suite'
-    id 'com.google.cloud.tools.jib' version '3.4.4'
+    id 'com.google.cloud.tools.jib' version '3.4.5'
 }
 
 java {

code/services-core/assistant-service/build.gradle

@@ -3,7 +3,7 @@ plugins {
 
     id 'application'
     id 'jvm-test-suite'
-    id 'com.google.cloud.tools.jib' version '3.4.4'
+    id 'com.google.cloud.tools.jib' version '3.4.5'
 }
 
 application {

code/services-core/assistant-service/java/nu/marginalia/assistant/suggest/Suggestions.java

@@ -68,6 +68,12 @@ public class Suggestions {
                     // Remove junk items we may have gotten from link extraction
                     if (word.startsWith("click here"))
                         continue;
+                    if (word.contains("new window"))
+                        continue;
+                    if (word.contains("click to"))
+                        continue;
+                    if (word.startsWith("share "))
+                        continue;
 
                     if (word.length() > 3) {
                         ret.insert(word, cnt);

code/services-core/control-service/build.gradle

@@ -2,7 +2,7 @@ plugins {
     id 'java'
     id 'application'
     id 'jvm-test-suite'
-    id 'com.google.cloud.tools.jib' version '3.4.4'
+    id 'com.google.cloud.tools.jib' version '3.4.5'
 }
 
 java {

code/services-core/executor-service/build.gradle

@@ -3,7 +3,7 @@ plugins {
 
     id 'application'
     id 'jvm-test-suite'
-    id 'com.google.cloud.tools.jib' version '3.4.4'
+    id 'com.google.cloud.tools.jib' version '3.4.5'
 }
 
 application {

code/services-core/index-service/build.gradle

@@ -3,7 +3,7 @@ plugins {
 
     id 'application'
     id 'jvm-test-suite'
-    id 'com.google.cloud.tools.jib' version '3.4.4'
+    id 'com.google.cloud.tools.jib' version '3.4.5'
 }
 
 application {

code/services-core/query-service/build.gradle

@@ -3,7 +3,7 @@ plugins {
 
     id 'application'
     id 'jvm-test-suite'
-    id 'com.google.cloud.tools.jib' version '3.4.4'
+    id 'com.google.cloud.tools.jib' version '3.4.5'
 }
 
 application {

code/tools/screenshot-capture-tool/build.gradle

@@ -3,7 +3,7 @@ plugins {
 
     id 'application'
     id 'jvm-test-suite'
-    id 'com.google.cloud.tools.jib' version '3.4.4'
+    id 'com.google.cloud.tools.jib' version '3.4.5'
 }
 
 java {

deploy.txt

@@ -1,4 +1,6 @@
-## This is a token file for automatic deployment
+## This is a token file for triggering automatic deployment when no commit is made.
 
 2025-01-08:  Deploy executor.
-2025-01-07:  Deploy executor.
+2025-01-07:  Deploy executor.
+2025-04-24:  Deploy executor.
+2025-04-24:  Deploy assistant.

gradle/wrapper/gradle-wrapper.properties

@@ -1,5 +1,5 @@
 distributionBase=GRADLE_USER_HOME
 distributionPath=wrapper/dists
-distributionUrl=https\://services.gradle.org/distributions/gradle-8.10-bin.zip
+distributionUrl=https\://services.gradle.org/distributions/gradle-8.14-bin.zip
 zipStoreBase=GRADLE_USER_HOME
 zipStorePath=wrapper/dists