Used named logger - #106

Coverage increased (+0.02%) to 84.877% when pulling 2c12403 on flopp:named-logger into e28f59c on tkrajina:master.

What if log = mod_logging.getLogger(__name__) is placed inside the relevant function/class instead?

When the call is placed at the module level it will create a new logger during import. If the dev using the library then loads a logger config from a file logging.config.fileConfig(), it will disable previous loggers by default. The call to getLogger() doesn't create a new instance each time, after the first call it just returns the previously instantiated object. https://docs.python.org/3.6/library/logging.config.html#logging.config.fileConfig

I think that would be possible.

BTW, my pull request is based on the logging pattern found in other popular libraries, e.g. https://github.com/shazow/urllib3/blob/master/urllib3/response.py

I think that initializing loggers in every single function or class of a module can really clutter up the code. The red warning box in the python 3.6 documentation makes it pretty clear that the default of disable_existing_loggers is only 'true' for legacy reasons and that likely most people actually want it to be 'false'. I guess the reason is precisely to have module level loggers to behave as expected.

However since gpxpy does not have that much logging, class/function level loggers wouldn't clutter the code too much here. So in this case one could go either way.

edit: Warning box at https://docs.python.org/3.6/howto/logging.html , about 2/3 down the page

Merged in dev at the moment, will be in master for the next release.