#32 Bluefish selecting wrong, non-UTF8 encoding based on source

[Miguel Simoes]

Bluefish is selecting a non-UTF8 encoding based on reading Python source code, namely a function call with parameter ' encoding="ISO-8859-1" '. This can be fixed by splitting that string parameter into a plus-concatenation. See attached files.

[Olivier Sessink]

yes I understand why this is happening.

what I don't know: is python sourcecode alsways in UTF-8 ? If so we can force UTF8 for every python file. If not we have to try and guess (which we currently do, resulting in the problem you describe)

[Miguel Simoes]

Python allows specifying the source file encoding at either the first or second line, see https://www.python.org/dev/peps/pep-0263/ . A regular expression is provided there for allowed directive formats. Default is ASCII for Python2 and UTF8 for Python3.
I would suggest a printed warning for this type of situations, which override preferences and locale.
Finally, the "Document/Character Encoding" options seem to do nothing in this case, but I didn't research much.
Cheers and let me know if I can help.

[Olivier Sessink]

I just found https://www.python.org/dev/peps/pep-0263/ which describes the proper python encoding.

[Olivier Sessink]

encoding detection according the pep-0263 is now properly detected in revision 8872

[Dr. Martin Senftleben]

Closing as the issue has been resolved. If not, please open a new ticket.