hotfix for smalot#359 and smalot#360: fallback for glyphs not in the …

…postscript lookup table

src/Smalot/PdfParser/Encoding/PostScriptGlyphs.php

@@ -1088,6 +1088,12 @@ public static function getGlyphs()
 
     public static function getCodePoint($glyph)
     {
-        return hexdec(static::getGlyphs()[$glyph]);
+        $glyphsMap = static::getGlyphs();
+
+        if (isset($glyphsMap[$glyph])) {
+            return hexdec($glyphsMap[$glyph]);
+        }
+
+        return $glyph;
     }
 }

tests/Integration/ParserTest.php

@@ -210,6 +210,29 @@ public function testIssue334()
 
         $this->assertStringContainsString('This question already has an answer here', $document->getText());
     }
+
+    /**
+     * Test that issue related pdf can now be parsed:
+     * Glyphs not in the Postscript lookup table would cause "Notice: Undefined offset"
+     *
+     * Note that some characters are missing from the parsed text or wrongly decoded,
+     * like the "ł" in przepływu or "ó" in osób. This was already the case before the
+     * PR that caused this issue and is not currently covered by this test case.
+     * However, these issues should be addressed in the future and their fixes can
+     * then be incorporated into this test by uncommenting the two assertions below.
+     *
+     * @see https://github.com/smalot/pdfparser/issues/359
+     */
+    public function testIssue359()
+    {
+        $filename = $this->rootDir.'/samples/bugs/Issue359.pdf';
+
+        $document = $this->fixture->parseFile($filename);
+
+        $this->assertStringContainsString('dnia 10 maja 2018 roku o ochronie danych osobowych', $document->getText());
+        // $this->assertStringContainsString('sprawie ochrony osób fizycznych w związku', $document->getText());
+        // $this->assertStringContainsString('sprawie swobodnego przepływu takich danych oraz uchylenia dyrektywy', $document->getText());
+    }
 }
 
 class ParserSub extends Parser