Markus Piotrowski
Actually, I can't remember where I got the data from. It was maybe a table which was supplied with our mass spec, but have to look it up. As shown...
We have it at least since May 30, where you started a new CI run for PR #3934 The same problem 14 days ago (May 27) when @mdehoon merged #3942,...
It's strange. In some CI runs, `test_windows (3.10)` passes or fails with the same error as 3.9: fail: https://github.com/biopython/biopython/actions/runs/2359515702 pass: https://github.com/biopython/biopython/actions/runs/2442081003 So 5 days ago, `test_windows (3.10)` was still running....
Yes, another hint, which makes it clearer: The last working/passing Windows tests used 3.9.12 and 3.10.4. 3.9.13 and 3.10.5 always failed.
Possible reason? > [gh-91810](https://github.com/python/cpython/issues/91810): [ElementTree](https://docs.python.org/release/3.9.13/library/xml.etree.elementtree.html#xml.etree.ElementTree.ElementTree) method [write()](https://docs.python.org/release/3.9.13/library/xml.etree.elementtree.html#xml.etree.ElementTree.ElementTree.write) and function [tostring()](https://docs.python.org/release/3.9.13/library/xml.etree.elementtree.html#xml.etree.ElementTree.tostring) now use the text file’s encoding (“UTF-8” if not available) instead of locale encoding in XML declaration when encoding="unicode" is...
~~So, my Windows notepad.exe says that the `test_PhyloXML.py` file (within my Git `biopython` folder) is encoded in 'Unicode BigEndian' and Notepad++ says it's 'UCS-2 BE BOM', so something like UTF-16...
When I run the test on my Windows 10, the `AssertationError` looks different, indicating that the ElementTree value is incorrect. You can also see from the traceback above that there...
Actually, the test `test_Distribution` itself runs fine, what is failing is the 'round-trip' test where the data files are rewritten by `Bio.PhyloXMLIO` and then all other tests are run with...
We can still apply the patch? I can re-open the issue afterwards. I'm happy to assist with testing on Windows. My problem is that I don't find an obvious place...
Found some interesting discussion [here](https://github.com/python/cpython/pull/91903#issuecomment-1144166476) which is probably related to our problem. It describes a behavior for `ElementTree.write()` that I have also seen during testing (but I was not aware...