Surface Guzzle response-creation failures as UserException instead of an internal error

src/HttpExtractor.php

@@ -13,10 +13,18 @@
 use Keboola\Component\UserException;
 use Keboola\HttpExtractor\Exception\EncodingException;
 use Psr\Http\Message\UriInterface;
+use Throwable;
 use function in_array;
 
 class HttpExtractor
 {
+    /**
+     * Message Guzzle uses when it cannot build a PSR-7 response out of the transfer.
+     *
+     * @see \GuzzleHttp\Handler\CurlFactory::createRejection()
+     */
+    private const RESPONSE_CREATION_ERROR_MESSAGE = 'An error was encountered while creating the response';
+
     /** @var Client */
     private $client;
 
@@ -78,12 +86,12 @@ private function sendRequest(UriInterface $httpSource, array $options): void
             ];
             $context = $e->getHandlerContext();
             if (!isset($context['errno'])) {
-                throw $e;
+                throw $this->convertResponseCreationError($e, $httpSource);
             }
 
             $curlErrorNumber = $context['errno'];
             if (!in_array($curlErrorNumber, $userErrors)) {
-                throw $e;
+                throw $this->convertResponseCreationError($e, $httpSource);
             }
             $curlErrorMessage = $context['error'];
             throw new UserException(sprintf(
@@ -102,6 +110,36 @@ private function sendRequest(UriInterface $httpSource, array $options): void
         // will throw exception for HTTP errors, no need to signal back
     }
 
+    /**
+     * When Guzzle cannot build the PSR-7 response out of a transfer - typically because the remote
+     * host answered with a malformed or missing HTTP status line - it wraps the real cause into a
+     * generic RequestException whose own message says nothing. cURL reports that abort as
+     * CURLE_WRITE_ERROR, which is not one of the error numbers recognized above, so the exception used
+     * to leave this class unhandled and the job died with an opaque internal error. Re-raise it as a
+     * UserException instead, so the run still fails but the underlying reason is visible. Anything
+     * else is returned untouched and keeps propagating exactly as before.
+     */
+    private function convertResponseCreationError(RequestException $e, UriInterface $httpSource): Throwable
+    {
+        // Guzzle passes this message verbatim, so match it exactly rather than as a substring -
+        // a response body echoed into some other exception message must not be caught here.
+        if ($e->getMessage() !== self::RESPONSE_CREATION_ERROR_MESSAGE) {
+            return $e;
+        }
+
+        $previous = $e->getPrevious();
+        if ($previous === null || $previous->getMessage() === '') {
+            // Nothing to tell the user beyond the generic message, so leave it exactly as it was.
+            return $e;
+        }
+
+        return new UserException(sprintf(
+            'Error requesting "%s": the response could not be processed: %s',
+            (string) $httpSource,
+            $previous->getMessage()
+        ), 0, $e);
+    }
+
     /**
      * @return mixed[]
      */

tests/phpunit/HttpExtractorTest.php

@@ -4,6 +4,7 @@
 
 namespace Keboola\HttpExtractor\Tests;
 
+use GuzzleHttp\Exception\RequestException;
 use GuzzleHttp\Handler\MockHandler;
 use GuzzleHttp\HandlerStack;
 use GuzzleHttp\Middleware;
@@ -17,6 +18,7 @@
 use Keboola\Temp\Temp;
 use Monolog\Handler\TestHandler;
 use PHPUnit\Framework\TestCase;
+use RuntimeException;
 use function file_get_contents;
 use function sys_get_temp_dir;
 
@@ -68,6 +70,74 @@ public function testTooManyRedirects(): void
         $extractor->extract($resource, $destination);
     }
 
+    /**
+     * Guzzle throws this when it cannot build a PSR-7 response out of the transfer, e.g. the server
+     * sent a malformed HTTP status line. The real cause only lives in the previous exception, and the
+     * generic RequestException used to escape the extractor as an opaque internal error (exit code 2).
+     *
+     * Exactly one response is queued on purpose: cURL error 23 is not a retryable code, so an
+     * unexpected retry must fail loudly on an empty mock queue instead of passing silently.
+     */
+    public function testResponseCreationErrorIsThrownAsUserException(): void
+    {
+        $guzzleException = new RequestException(
+            'An error was encountered while creating the response',
+            new Request('GET', 'http://example.com/result.txt'),
+            null,
+            new RuntimeException('HTTP version missing from header data'),
+            ['errno' => CURLE_WRITE_ERROR, 'error' => 'Failed writing header']
+        );
+
+        $client = $this->getMockedExtractorClient([$guzzleException]);
+        $extractor = new HttpExtractor($client, []);
+        $temp = new Temp();
+
+        try {
+            $extractor->extract(
+                new Uri('http://example.com/result.txt'),
+                $temp->createTmpFile()->getPathname()
+            );
+            $this->fail('Expected a UserException to be thrown');
+        } catch (UserException $e) {
+            $this->assertSame(
+                'Error requesting "http://example.com/result.txt": ' .
+                'the response could not be processed: HTTP version missing from header data',
+                $e->getMessage()
+            );
+            // the original exception must survive so the job log keeps the full context
+            $this->assertSame($guzzleException, $e->getPrevious());
+        }
+    }
+
+    /**
+     * Guard for the above: any other RequestException the extractor does not recognize must keep
+     * propagating as the very same object, exactly as before.
+     */
+    public function testUnrelatedRequestExceptionKeepsPropagating(): void
+    {
+        $guzzleException = new RequestException(
+            'cURL error 23: Failed writing body',
+            new Request('GET', 'http://example.com/result.txt'),
+            null,
+            null,
+            ['errno' => CURLE_WRITE_ERROR, 'error' => 'Failed writing body']
+        );
+
+        $client = $this->getMockedExtractorClient([$guzzleException]);
+        $extractor = new HttpExtractor($client, []);
+        $temp = new Temp();
+
+        try {
+            $extractor->extract(
+                new Uri('http://example.com/result.txt'),
+                $temp->createTmpFile()->getPathname()
+            );
+            $this->fail('Expected the original RequestException to be thrown');
+        } catch (RequestException $e) {
+            $this->assertSame($guzzleException, $e);
+        }
+    }
+
     private function getTestLogger(): Logger
     {
         $this->testHandler = new TestHandler();