Solr 10 support

.github/workflows/run-tests.yml

@@ -15,7 +15,7 @@ jobs:
         strategy:
             matrix:
                 php: [8.1, 8.2, 8.3, 8.4, 8.5]
-                solr: [7, 8, 9]
+                solr: [7, 8, 9, 10]
                 mode: [cloud, server]
 
         name: PHP ${{ matrix.php }}, Solr ${{ matrix.solr }} ${{ matrix.mode }}
@@ -59,6 +59,14 @@ jobs:
                 ref: branch_9_10
                 path: lucene-solr
 
+            - name: Checkout solr 10.0
+              if: matrix.solr == 10
+              uses: actions/checkout@v4
+              with:
+                repository: apache/solr
+                ref: branch_10_0
+                path: lucene-solr
+
             - name: Start Solr ${{ matrix.solr }} in ${{ matrix.mode }} mode
               run: |
                 chmod -R a+w lucene-solr

CHANGELOG.md

@@ -6,6 +6,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## [7.0.0]
 ### Added
+- Solr 10 support
 - PHP 8.5 support
 - Solarium\QueryType\Extract\Query::setStreamType()
 
@@ -15,13 +16,14 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 - Solarium\Plugin\MinimumScoreFilter\Document::__call() unpacks the $arguments array when forwarding a method call to the original document
 
 ### Changed
- - Added `void` return type to `Solarium\Core\Plugin\PluginInterface::initPlugin()` method signature
- - Added `void` return type to `Solarium\Core\Plugin\PluginInterface::deinitPlugin()` method signature
- - Added `void` return type to `Solarium\Core\Plugin\AbstractPlugin::initPluginType()` method signature
- - Changed return type of some Facets methods from `self`to `static`
- - Replaced class property type hints with type declarations
- - Added union type declarations to method signatures
- - Solarium\Core\Query\Helper::formatDate() throws a `TypeError` instead of returning `false` if called with an incompatibly typed parameter
+- Added `void` return type to `Solarium\Core\Plugin\PluginInterface::initPlugin()` method signature
+- Added `void` return type to `Solarium\Core\Plugin\PluginInterface::deinitPlugin()` method signature
+- Added `void` return type to `Solarium\Core\Plugin\AbstractPlugin::initPluginType()` method signature
+- Changed return type of some Facets methods from `self` to `static`
+- Replaced class property type hints with type declarations
+- Added union type declarations to method signatures
+- Solarium\Core\Query\Helper::formatDate() throws a `TypeError` instead of returning `false` if called with an incompatibly typed parameter
+- Managed resources queries no longer work around SOLR-6853 by default. Set the 'useDoubleEncoding' option to `true` if this bug affects you.
 
 ### Removed
 - Solarium\Component\Result\Stats\FacetValue::getFacets(), always returned `null`

docs/getting-started.md

@@ -112,6 +112,11 @@ do have one it must be contravariant.
 Code that calls a method with parameters of an incompatible type will result in a `TypeError` where
 they were previously coerced into a compatible type by PHP's type juggling.
 
+Solarium 7 no longer works around [SOLR-6853](https://issues.apache.org/jira/browse/SOLR-6853) for
+managed resources queries by default. The workaround is no longer necessary for Solr 10, but also
+not forward compatible with this version. If this still affects you, set the 'useDoubleEncoding'
+option to `true` on `ManagedResources\Query\Stopwords` or `ManagedResources\Query\Synonyms`.
+
 ### Pitfall when upgrading to 6.3.6
 
 Using a config object is no longer supported. You have to convert it to an array before passing

docs/queries/managedresources-query/managed-stopwords.md

@@ -245,6 +245,24 @@ htmlFooter();
 
 ```
 
+A note on percent-encoding reserved characters
+----------------------------------------------
+
+If the name of a stopword list or a stopword itself contains characters that are not
+[unreserved characters as defined by RFC 3986](https://www.rfc-editor.org/rfc/rfc3986#section-2.3),
+they must be percent-encoded when appearing as part of a URL. Solarium handles this for you.
+
+However, if you're using a Solr version prior to Solr 10 and list names or stopwords
+that contain [reserved characters](https://datatracker.ietf.org/doc/html/rfc3986#section-2.2),
+you will be affected by [SOLR-6853](https://issues.apache.org/jira/browse/SOLR-6853). You can
+instruct Solarium to double up on the percent-encoding as a workaround.
+
+```php
+$query = $client->createManagedStopwords(['useDoubleEncoding' => true]);
+```
+
+Keep in mind that Solr may not be able to handle some of these reserved characters regardless.
+
 A note on `HEAD` requests
 -------------------------
 

docs/queries/managedresources-query/managed-synonyms.md

@@ -305,6 +305,24 @@ htmlFooter();
 
 ```
 
+A note on percent-encoding reserved characters
+----------------------------------------------
+
+If the name of a synonym map or a synonym itself contains characters that are not
+[unreserved characters as defined by RFC 3986](https://www.rfc-editor.org/rfc/rfc3986#section-2.3),
+they must be percent-encoded when appearing as part of a URL. Solarium handles this for you.
+
+However, if you're using a Solr version prior to Solr 10 and map names or synonyms
+that contain [reserved characters](https://datatracker.ietf.org/doc/html/rfc3986#section-2.2),
+you will be affected by [SOLR-6853](https://issues.apache.org/jira/browse/SOLR-6853). You can
+instruct Solarium to double up on the percent-encoding as a workaround.
+
+```php
+$query = $client->createManagedSynonyms(['useDoubleEncoding' => true]);
+```
+
+Keep in mind that Solr may not be able to handle some of these reserved characters regardless.
+
 A note on `HEAD` requests
 -------------------------
 

examples/execute_all.php

@@ -62,8 +62,8 @@
         $response = $client->coreAdmin($coreAdminQuery);
     }
 
-    // @todo Figure out why this fails on Solr 9.
-    if (9 !== $solrVersion) {
+    // @todo Figure out why this fails starting with Solr 9.
+    if (9 > $solrVersion) {
         // check if /mlt handler exists (it will in the github worklow, but not when running this script on its own)
         $query = $client->createApi([
             'version' => Request::API_V1,
@@ -90,6 +90,23 @@
         }
     }
 
+    // Solr 10 no longer comes with LocalTikaExtractionBackend, the github workflow runs a remote Tika Server instead
+    if (10 <= $solrVersion) {
+        $query = $client->createApi([
+            'version' => Request::API_V1,
+            'handler' => $collection_or_core_name.'/config',
+            'method' => Request::METHOD_POST,
+            'rawdata' => json_encode([
+                'update-requesthandler' => [
+                    'name' => '/update/extract',
+                    'class' => 'solr.extraction.ExtractingRequestHandler',
+                    'tikaserver.url' => 'http://tika:9998',
+                ],
+            ]),
+        ]);
+        $client->execute($query);
+    }
+
     // check if attr_* dynamic field definition exists (it was removed from the techproducts configset in Solr 9.1)
     try {
         $query = $client->createApi([
@@ -172,7 +189,7 @@
     // examples that can't be run against this Solr version
     $skipForSolrVersion = [];
 
-    if (9 === $solrVersion) {
+    if (9 <= $solrVersion) {
         $skipForSolrVersion[] = '2.3.1-mlt-query.php';
         $skipForSolrVersion[] = '2.3.2-mlt-stream.php';
     }

src/QueryType/ManagedResources/Query/AbstractQuery.php

@@ -230,4 +230,32 @@ public function removeCommand(): self
      * @return InitArgsInterface
      */
     abstract public function createInitArgs(?array $initArgs = null): InitArgsInterface;
+
+    /**
+     * Percent-encode names and terms twice as a workaround for SOLR-6853?
+     *
+     * @return bool
+     */
+    public function getUseDoubleEncoding(): bool
+    {
+        return $this->getOption('useDoubleEncoding') ?? false;
+    }
+
+    /**
+     * Percent-encode names and terms twice as a workaround for SOLR-6853?
+     *
+     * Solr versions prior to 10 required reserved characters to be doubly
+     * percent-encoded. Set this to true if your Solr version is affected by
+     * {@link https://issues.apache.org/jira/browse/SOLR-6853 SOLR-6853}.
+     *
+     * @param bool $useDoubleEncoding
+     *
+     * @return self Provides fluent interface
+     */
+    public function setUseDoubleEncoding(bool $useDoubleEncoding): self
+    {
+        $this->setOption('useDoubleEncoding', $useDoubleEncoding);
+
+        return $this;
+    }
 }

src/QueryType/ManagedResources/RequestBuilder/Resource.php

@@ -13,7 +13,6 @@
 use Solarium\Core\Query\AbstractRequestBuilder;
 use Solarium\Core\Query\QueryInterface;
 use Solarium\Exception\RuntimeException;
-use Solarium\QueryType\ManagedResources\Query\AbstractCommand;
 use Solarium\QueryType\ManagedResources\Query\AbstractQuery as BaseQuery;
 
 /**
@@ -36,68 +35,92 @@ public function build(QueryInterface|BaseQuery $query): Request
             throw new RuntimeException('Name of the resource is not set in the query.');
         }
 
+        $name = rawurlencode($query->getName());
+        if ($query->getUseDoubleEncoding()) {
+            $name = rawurlencode($name);
+        }
+
         $request = parent::build($query);
-        // reserved characters in a REST resource name need to be encoded twice to make it through the servlet (SOLR-6853)
-        $request->setHandler($query->getHandler().rawurlencode(rawurlencode($query->getName())));
+        $request->setHandler($query->getHandler().$name);
+
         if (null !== $query->getCommand()) {
             $request->setContentType(Request::CONTENT_TYPE_APPLICATION_JSON);
-            $this->buildCommand($request, $query->getCommand());
+            $this->buildCommand($query, $request);
         } else {
             // Lists one or all items.
             $request->setMethod(Request::METHOD_GET);
 
             if (null !== $term = $query->getTerm()) {
-                // reserved characters in a REST resource name need to be encoded twice to make it through the servlet (SOLR-6853)
-                $request->setHandler($request->getHandler().'/'.rawurlencode(rawurlencode($term)));
+                $term = rawurlencode($term);
+                if ($query->getUseDoubleEncoding()) {
+                    $term = rawurlencode($term);
+                }
+
+                $request->setHandler($request->getHandler().'/'.$term);
             }
         }
 
         return $request;
     }
 
     /**
-     * @param Request         $request
-     * @param AbstractCommand $command
+     * @param QueryInterface|BaseQuery $query
+     * @param Request                  $request
      *
      * @throws RuntimeException
      *
      * @return self Provides fluent interface
      */
-    protected function buildCommand(Request $request, AbstractCommand $command): self
+    protected function buildCommand(QueryInterface|BaseQuery $query, Request $request): self
     {
+        $command = $query->getCommand();
+
         $request->setMethod($command->getRequestMethod());
 
         switch ($command->getType()) {
             case BaseQuery::COMMAND_ADD:
                 if (null === $rawData = $command->getRawData()) {
                     throw new RuntimeException('Missing data for ADD command.');
                 }
+
                 $request->setRawData($rawData);
                 break;
             case BaseQuery::COMMAND_CONFIG:
                 if (null === $rawData = $command->getRawData()) {
                     throw new RuntimeException('Missing initArgs for CONFIG command.');
                 }
+
                 $request->setRawData($rawData);
                 break;
             case BaseQuery::COMMAND_CREATE:
                 if (null === $rawData = $command->getRawData()) {
                     throw new RuntimeException('Missing class for CREATE command.');
                 }
+
                 $request->setRawData($rawData);
                 break;
             case BaseQuery::COMMAND_DELETE:
                 if (null === $term = $command->getTerm()) {
                     throw new RuntimeException('Missing term for DELETE command.');
                 }
-                // reserved characters in a REST resource name need to be encoded twice to make it through the servlet (SOLR-6853)
-                $request->setHandler($request->getHandler().'/'.rawurlencode(rawurlencode($command->getTerm())));
+
+                $term = rawurlencode($term);
+                if ($query->getUseDoubleEncoding()) {
+                    $term = rawurlencode($term);
+                }
+
+                $request->setHandler($request->getHandler().'/'.$term);
                 break;
             case BaseQuery::COMMAND_EXISTS:
                 if (null !== $term = $command->getTerm()) {
-                    // reserved characters in a REST resource name need to be encoded twice to make it through the servlet (SOLR-6853)
-                    $request->setHandler($request->getHandler().'/'.rawurlencode(rawurlencode($command->getTerm())));
+                    $term = rawurlencode($term);
+                    if ($query->getUseDoubleEncoding()) {
+                        $term = rawurlencode($term);
+                    }
+
+                    $request->setHandler($request->getHandler().'/'.$term);
                 }
+
                 break;
             case BaseQuery::COMMAND_REMOVE:
                 break;