Recognize explicit UTF-8 strings (fixes #654)

ChangeLog

@@ -1,3 +1,8 @@
+2022-02-22  Jay Berkenbilt  <ejb@ql.org>
+
+	* Recognize PDF strings explicitly marked as UTF-8 as allowed by
+	the PDF 2.0 spec. Fixes #654.
+
 2022-02-18  Jay Berkenbilt  <ejb@ql.org>
 
 	* Bug fix: when generating appearance streams, the font size was

TODO

@@ -10,6 +10,14 @@ Priorities for 11:
 * PointerHolder -> shared_ptr
 * ABI
 
+Misc
+* Get rid of "ugly switch statements" in QUtil.cc -- replace with
+  static map initializers. (Search for "ugly switch statements" below
+  as well.)
+* Consider exposing get_next_utf8_codepoint in QUtil
+* Add QUtil::is_explicit_utf8 that does what QPDF_String::getUTF8Val
+  does to detect UTF-8 encoded strings per PDF 2.0 spec.
+
 Soon: Break ground on "Document-level work"
 
 Code Formatting

libqpdf/QPDF_String.cc

@@ -183,6 +183,15 @@ QPDF_String::getUTF8Val() const
     {
         return QUtil::utf16_to_utf8(this->val);
     }
+    else if ((val.length() >= 3) &&
+             (val[0] == '\xEF') &&
+             (val[1] == '\xBB') &&
+             (val[2] == '\xBF'))
+    {
+        // PDF 2.0 allows UTF-8 strings when explicitly prefixed with
+        // the above bytes, which is just UTF-8 encoding of U+FEFF.
+        return this->val.substr(3);
+    }
     else
     {
         return QUtil::pdf_doc_to_utf8(this->val);

manual/release-notes.rst

@@ -9,6 +9,9 @@ For a detailed list of changes, please see the file
 10.6.3: XXX
   - Bug fixes:
 
+    - Recognize strings explicitly encoded as UTF-8 as allowed by the
+      PDF 2.0 spec.
+
     - Fix edge cases with appearance stream generation for form fields
       whose ``/DA`` field lacks proper font size specification or that
       specifies auto sizing. At this time, qpdf does not support auto

qpdf/qtest/qpdf/numeric-and-string-3.out

@@ -7,8 +7,9 @@ end page 1
 QStrings:
 No Special Characters
 These: ¿÷¢þ and no more
+Explicit utf-8 with π
 πωτατω
-treble clef: 𝄠; sixteenth note: 𝅘𝅥𝅮
+treble clef: 𝄠; sixteenth note: 𝅘𝅥𝅯
 QNumbers:
 1.000
 3.142

qpdf/qtest/qpdf/numeric-and-string-3.pdf

@@ -12,8 +12,9 @@
   /QStrings [
     (No Special Characters)
     (These: ¿÷¢þ and no more)
+    (\357\273\277Explicit utf-8 with \317\200)
     <feff03c003c903c403b103c403c9>
-    <feff0074007200650062006c006500200063006c00650066003a0020d834dd20003b0020007300690078007400650065006e007400680020006e006f00740065003a0020d834dd60>
+    <feff0074007200650062006c006500200063006c00650066003a0020d834dd20003b0020007300690078007400650065006e007400680020006e006f00740065003a0020d834dd61>
   ]
   /Type /Catalog
 >>
@@ -110,19 +111,19 @@ xref
 0 10
 0000000000 65535 f 
 0000000025 00000 n 
-0000000377 00000 n 
+0000000424 00000 n 
-0000000459 00000 n 
+0000000506 00000 n 
-0000000694 00000 n 
+0000000741 00000 n 
-0000000793 00000 n 
+0000000840 00000 n 
-0000000835 00000 n 
+0000000882 00000 n 
-0000000933 00000 n 
+0000000980 00000 n 
-0000000952 00000 n 
+0000000999 00000 n 
-0000001070 00000 n 
+0000001117 00000 n 
 trailer <<
   /Root 1 0 R
   /Size 10
   /ID [<e017d8dc1fe53a81e40aa79bcb43fdec><76269ee0b6579446b731e060af8ef436>]
 >>
 startxref
-1105
+1152
 %%EOF