is there any chance of a python3 compatible code base? how about just for DBUtils?
Hi Derek, I have many ideas and plans for a modern version of Webware that supports Python 3 (maybe only Python 3). DBUtils will also be updated, but my idea here is to first update PyGreSQL to support Python 3, since DBUtils has special support for PyGreSQL. The only thing that hinders me is not finding enough idle time to tackle this.
Concerning Webware, I'm still not completely decided where to aim for: 1) Port everything 1:1 to Python 3, as compatible as possible, without adding or removing stuff, or 2) Use this opportunity to modernize everything: Use namespace packages instead of the Webware plugin-system, use WSGI instead of the webware adapter protocol, use modern Python features such as properties instead of getters, use PEP8 naming conventions, make WebKit the core package, remove some of the lesser used outdated Plug-ins and replace or integrate them with more modern helper packages, e.g. for templating or creating HTML via Python (e.g. dominate), add better support for AJAX, rewrite the Documentation with Sphinx, add unit tests for everything, etc.
Personally, I'm leaning towards the second aim which is also much more interesting and fun to do as a project, so there is a better chance that this will happen.
The loss of compatibility could be mitigated by adding a py2to3 fixer to convert Webware servlets (e.g. adapt getters to properties and new PEP8 names).
Which of the two aims is more interesting for you? Are there any plug-ins besides WebKit that are essential for you? Or are you more interested in DBUtils than in Webware itself?
Personally I'm more interested in DBUtils but that's because I constantly have to write code to handle loosing connectivity to databases while minimizing reconnects.
When it comes to porting I tend to prefer modernization rather than being strict about 1:1 conversion. A 2to3 fixer would be a very cool thing to offer. We are getting closer to a world where python3 only could make sense and that option sounds the most interesting to me.
Ok, thanks for the feedback. You can subscribe to https://sourceforge.net/p/webware/mailman/ to get info on new developments or make suggestions. The project is currently dormant, but maybe I will revive it later this year.