Merge v6.1.5

src/ocrmypdf/exec/ghostscript.py

@@ -32,6 +32,18 @@ def version():
     return get_version('gs')
 
 
+def jpeg_passthrough_available():
+    """
+    Ghostscript 9.23 introduced JPEG passthrough but it seems to corrupt the
+    last two bytes of certain images, for now we disable it for 9.23 and
+    do not mention it for < 9.23.
+
+    https://bugs.ghostscript.com/show_bug.cgi?id=699216
+
+    """
+    return False
+
+
 def _gs_error_reported(stream):
     return re.search(r'error', stream, flags=re.IGNORECASE)
 
@@ -133,7 +145,8 @@ def generate_pdfa(pdf_pages, output_file, compression, log,
     strategy = 'RGB' if version() >= '9.19' else '/RGB'
 
     if version() == '9.23':
-        # 9.23: JPEG passthrough broken for image masks?
+        # 9.23: new feature JPEG passthrough is broken in some cases, best to
+        # disable it always
         # https://bugs.ghostscript.com/show_bug.cgi?id=699216
         compression_args.append('-dPassThroughJPEGImages=false')
 

src/ocrmypdf/hocrtransform.py

@@ -322,8 +322,11 @@ class HocrTransform():
             dy = baseline_y2 - cursor[1]            
             text.moveCursor(dx, dy)
 
-            text.setHorizScale(100 * box_width / font_width)
-            text.textOut(elemtxt)
+            # If reportlab tells us this word is 0 units wide, our best seems
+            # to be to suppress this text
+            if font_width > 0:
+                text.setHorizScale(100 * box_width / font_width)
+                text.textOut(elemtxt)
         pdf.drawText(text)
 
 

src/ocrmypdf/pdfa.py

@@ -143,6 +143,7 @@ def decode_pdf_date(s: str) -> datetime:
 
     if s.startswith('D:'):
         s = s[2:]
+    s = s.replace("'", "")  # Remove apos from PDF time strings
     for fmt in pdfmark_date_fmts:
         try:
             return datetime.strptime(s, fmt)