I looked into what it would take to convert to 3, since 2 is no longer supported.
The short list is:
- Run
2to3 -wn .
- Add
#!/usr/bin/env python3 where appropriate
- change
rfc822 to email.utils in src/feedvalidator/validators.py
- change
new.instgancemethod to types.methodType in src/validtest.py
- explicitly set encoding for the XML file when opening in
src/validtest.py's getDescription
- change
UnicodeError to UnicodeDecodeError or UnicodeEncodeError as appropriate
- explicitly convert the result of
.encode('idna') back to a string (various places)
- in
src/tests/testXmlEncoding.py, change many things to binary literals
- avoid using
xmlEncoding.asUTF8 in src/feedvalidator/__init__.py
Where I got stuck was in how the code uses object inheritance so extensively; thanks to the switch to new-style objects, this doesn't work any more, and the trivial fix (removing things like validatorBase to avoid duplication) seems to mess with the processing model, causing many errors. E.g.,
TypeError: Cannot create a consistent method resolution
order (MRO) for bases text, rfc2396_full
Fixing this looks pretty invasive, and I'm not confident enough in this codebase to make the necessary changes. If someone is interested in helping out, I can do a PR for the things above and work with them on the solution.
I looked into what it would take to convert to 3, since 2 is no longer supported.
The short list is:
2to3 -wn .#!/usr/bin/env python3where appropriaterfc822toemail.utilsinsrc/feedvalidator/validators.pynew.instgancemethodtotypes.methodTypeinsrc/validtest.pysrc/validtest.py'sgetDescriptionUnicodeErrortoUnicodeDecodeErrororUnicodeEncodeErroras appropriate.encode('idna')back to a string (various places)src/tests/testXmlEncoding.py, change many things to binary literalsxmlEncoding.asUTF8insrc/feedvalidator/__init__.pyWhere I got stuck was in how the code uses object inheritance so extensively; thanks to the switch to new-style objects, this doesn't work any more, and the trivial fix (removing things like
validatorBaseto avoid duplication) seems to mess with the processing model, causing many errors. E.g.,Fixing this looks pretty invasive, and I'm not confident enough in this codebase to make the necessary changes. If someone is interested in helping out, I can do a PR for the things above and work with them on the solution.