Allow arbitrary ON values for CheckBoxes

We previously insisted that the ON-values for CheckBoxes are defined as "Yes" in the PDF (the published recommended value).

This fix introduces lifts this restriction and supports arbitrary values as ON for  check boxes.

So a check box turns to OFF if setting its field value to `False` or "Off".
It is set to ON if the field value is set to any of 'True`, "Yes" or the respective value coded in the PDF.

When checking a check box's value, the return values invariably are either "Off" or the ON-value in the PDF. Which value this is, can be checked via method `on_state()`.

Author: JorjMcKie

src/__init__.py

@@ -7269,8 +7269,6 @@ def on_state(self):
         """
         if self.field_type not in (2, 5):
             return None  # no checkbox or radio button
-        if self.field_type == 2:
-            return "Yes"
         bstate = self.button_states()
         if bstate is None:
             bstate = dict()
@@ -17681,13 +17679,13 @@ def GETATTR(name):
                 mupdf.pdf_dict_put_name(annot_obj, PDF_NAME('AS'), on)
             elif text:
                 mupdf.pdf_dict_put_name(annot_obj, PDF_NAME('AS'), text)
-    elif field_type == mupdf.PDF_WIDGET_TYPE_CHECKBOX:    # will always be "Yes" or "Off"
-        if value is True or text == 'Yes':
-            onstate = mupdf.pdf_button_field_on_state(annot_obj)
-            on = mupdf.pdf_to_name(onstate)
+    elif field_type == mupdf.PDF_WIDGET_TYPE_CHECKBOX:
+        onstate = mupdf.pdf_button_field_on_state(annot_obj)
+        on = onstate.pdf_to_name()
+        if value in (True, on) or text == 'Yes':
             mupdf.pdf_set_field_value(pdf, annot_obj, on, 1)
-            mupdf.pdf_dict_put_name(annot_obj, PDF_NAME('AS'), 'Yes')
-            mupdf.pdf_dict_put_name(annot_obj, PDF_NAME('V'), 'Yes')
+            mupdf.pdf_dict_put_name(annot_obj, PDF_NAME('AS'), on)
+            mupdf.pdf_dict_put_name(annot_obj, PDF_NAME('V'), on)
         else:
             mupdf.pdf_dict_put_name( annot_obj, PDF_NAME('AS'), 'Off')
             mupdf.pdf_dict_put_name( annot_obj, PDF_NAME('V'), 'Off')

tests/resources/test-4055.pdf

@@ -8,6 +8,7 @@
 scriptdir = os.path.abspath(os.path.dirname(__file__))
 filename = os.path.join(scriptdir, "resources", "widgettest.pdf")
 file_2333 = os.path.join(scriptdir, "resources", "test-2333.pdf")
+file_4055 = os.path.join(scriptdir, "resources", "test-4055.pdf")
 
 
 doc = pymupdf.open()
@@ -379,3 +380,57 @@ def get_widgets_by_name(doc):
         assert str(e) == 'Annot is not bound to a page'
 
     doc.close()
+
+
+def test_4055():
+    """Check correct setting of CheckBox "Yes" values.
+
+    Test scope:
+    * setting on with any of 'True' / 'Yes' / built-in values works
+    * setting off with any of 'False' or 'Off' works
+    """
+
+    # this PDF has digits as "Yes" values.
+    doc = pymupdf.open(file_4055)
+    page = doc[0]
+
+    # Round 1: confirm all check boxes are off
+    for w in page.widgets(types=[2]):
+        # check that this file doesn't use the "Yes" standard
+        assert w.on_state() != "Yes"
+        assert w.field_value == "Off"  # all check boxes are off
+        w.field_value = w.on_state()
+        w.update()
+
+    page = doc.reload_page(page)  # reload page to make sure we start fresh
+
+    # Round 2: confirm that fields contain the PDF's own on values
+    for w in page.widgets(types=[2]):
+        # confirm each value coincides with the "Yes" value
+        assert w.field_value == w.on_state()
+        w.field_value = False  # switch to "Off" using False
+        w.update()
+
+    page = doc.reload_page(page)
+
+    # Round 3: confirm that 'False' achieved "Off" values
+    for w in page.widgets(types=[2]):
+        assert w.field_value == "Off"
+        w.field_value = True  # use True for the next round
+        w.update()
+
+    page = doc.reload_page(page)
+
+    # Round 4: confirm that setting to True also worked
+    for w in page.widgets(types=[2]):
+        assert w.field_value == w.on_state()
+        w.field_value = "Off"  # set off again
+        w.update()
+        w.field_value = "Yes"
+        w.update()
+
+    page = doc.reload_page(page)
+
+    # Round 5: final check: setting to "Yes" also does work
+    for w in page.widgets(types=[2]):
+        assert w.field_value == w.on_state()