BUG: Fixed reference of value with str.__new__ in TextStringObject

[thomas-forte]

based on comments from #2949

• Fixing reference of value with str.new output
• updated type annotation on update_page_form_field_values

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 96.36%. Comparing base (5587414) to head (00f075f).
Report is 124 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main    #2952   +/-   ##
=======================================
  Coverage   96.36%   96.36%           
=======================================
  Files          52       52           
  Lines        8745     8747    +2     
  Branches     1589     1591    +2     
=======================================
+ Hits         8427     8429    +2     
  Misses        190      190           
  Partials      128      128           

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[stefan6419846]

Could you please add a test which shows why this change is necessary?

[PJBrs]

I was running into the same issue and this PR fixes it. The code I'm working on sometimes writes an int value into a form, which up until recently worked without problem. I'd help out with writing a test, but my distro is pretty ancient in certain regards...

[stefan6419846]

As long as your distro supports running the latest pypdf version, id est at least Python 3.8, this should be no real issue in theory.

[PJBrs]

... or maybe existing tests can be altered a bit, like this:

diff --git a/tests/test_writer.py b/tests/test_writer.py
index c1078226..b328b7d5 100644
--- a/tests/test_writer.py
+++ b/tests/test_writer.py
@@ -1458,7 +1458,7 @@ def test_update_form_fields(tmp_path):
             "RdoS1": "/",
             "Combo1": "!!monCombo!!",
             "Liste1": "Liste2",
-            "Liste2": ["Lst1", "Lst3"],
+            "Liste2": [int(1), float(2.8), "Lst3"],
             "DropList1": "DropListe3",
         },
         auto_regenerate=False,
@@ -1487,7 +1487,7 @@ def test_update_form_fields(tmp_path):
     assert flds["RadioGroup1"]["/V"] == "/2"
     assert flds["RadioGroup1"]["/Kids"][0].get_object()["/AS"] == "/Off"
     assert flds["RadioGroup1"]["/Kids"][1].get_object()["/AS"] == "/2"
-    assert all(x in flds["Liste2"]["/V"] for x in ["Lst1", "Lst3"])
+    assert all(x in flds["Liste2"]["/V"] for x in ["1", "2.8", "Lst3"])
 
     assert all(x in flds["CheckBox1"]["/_States_"] for x in ["/Off", "/Yes"])
     assert all(x in flds["RadioGroup1"]["/_States_"] for x in ["/1", "/2", "/3"])

@stefan6419846 Does this suffices? If so, @thomas-forte , please feel free to just add this to your PR.

@stefan6419846 I'm on Slackware64-15.0 with python 3.9. Tests failed first with:
pytest3: error: unrecognized arguments: --disable-socket

and if I disable that option, I get:

E pytest.PytestUnknownMarkWarning: Unknown pytest.mark.timeout

Anyway, with more commenting out I was able to cobble up and test the above :-)

[stefan6419846]

I have just added some even simpler tests to the TextStringObject directly as mypy validation is not running on the tests anyway.

Regarding your test issues: Please make sure to install all required dependencies using pip install -r requirements/ci.txt beforehand. We depend on some pytest plugins which are reported as missing in your case.

[thomas-forte]

I'm sorry I wasn't on top of this, was dealing with family and holidays, but thank you for adding that simple test!

For me the type annotations and docstring's were not showing for this libraries context in my IDE. So while I was supplying an int or float sometimes I had no idea that was incorrect. It still worked on the previous releases, my app was using 3.9.1, before updating. Specifically it was introduced on the UTF16_BOM pull request when the new method was added.