Rewrite email parser using ImapEngine, harden processing loop

Replace webklex/php-imap with directorytree/imapengine in the ticket
email parser. ImapEngine is pure PHP over sockets.

Parser improvements:
- Wrap per-message processing in try/catch so one malformed email
  can't abort the run; failures are flagged and logged with UID
- Query unseen + unflagged so previously-failed (flagged) messages
  are no longer re-processed on every cron run
- Skip vacation/auto-responder emails (RFC 3834) to prevent mail
  loops with the ticket auto-reply
- Cap messages per run (50) and attachment size (15MB); inline
  images over 2MB are stored as attachments instead of base64-embedded
  in ticket details
- Atomic lock file creation
- preg_quote() the ticket prefix in subject matching
- Dedupe CC watchers and exclude the sender
- Map webklex 'tls' encryption setting to STARTTLS for compatibility

NDR/DSN parsing now walks MIME parts via the underlying
zbateson parser instead of relying on attachment extraction.

plugins/vendor/symfony/mime/Encoder/QpContentEncoder.php

@@ -0,0 +1,55 @@
+<?php
+
+/*
+ * This file is part of the Symfony package.
+ *
+ * (c) Fabien Potencier <fabien@symfony.com>
+ *
+ * For the full copyright and license information, please view the LICENSE
+ * file that was distributed with this source code.
+ */
+
+namespace Symfony\Component\Mime\Encoder;
+
+/**
+ * @author Lars Strojny
+ */
+final class QpContentEncoder implements ContentEncoderInterface
+{
+    public function encodeByteStream($stream, int $maxLineLength = 0): iterable
+    {
+        if (!\is_resource($stream)) {
+            throw new \TypeError(\sprintf('Method "%s" takes a stream as a first argument.', __METHOD__));
+        }
+
+        // we don't use PHP stream filters here as the content should be small enough
+        yield $this->encodeString(stream_get_contents($stream), 'utf-8', 0, $maxLineLength);
+    }
+
+    public function getName(): string
+    {
+        return 'quoted-printable';
+    }
+
+    public function encodeString(string $string, ?string $charset = 'utf-8', int $firstLineOffset = 0, int $maxLineLength = 0): string
+    {
+        return $this->standardize(quoted_printable_encode($string));
+    }
+
+    /**
+     * Make sure CRLF is correct and HT/SPACE are in valid places.
+     */
+    private function standardize(string $string): string
+    {
+        // transform CR or LF to CRLF
+        $string = preg_replace('~=0D(?!=0A)|(?<!=0D)=0A~', '=0D=0A', $string);
+        // transform =0D=0A to CRLF
+        $string = str_replace(["\t=0D=0A", ' =0D=0A', '=0D=0A'], ["=09\r\n", "=20\r\n", "\r\n"], $string);
+
+        return match ($string[-1] ?? '') {
+            "\x09" => substr_replace($string, '=09', -1),
+            "\x20" => substr_replace($string, '=20', -1),
+            default => $string,
+        };
+    }
+}