Quote:
Originally Posted by jhowell
My main concern is due to the large number of changes needed in order to make it Python 3 compatible. Because the file formats involved are all binary data and there are significant differences in that between Python 2 and 3, I suspect that it will take many hours of work to fix the plugin.
Then there is the problem of testing all of the features of the plugin to see if the changes have broken them. There is no test suite or existing collection of all of the file types that are needed for testing.
|
KindleUnpack has been Python3 compatible (and backward-compatible with Python2) for years. The portions that deal with parsing the binary db and the EXTH header would be much the same for both. It shouldn't take all that long since there's already a basic roadmap to follow for dealing with all the binary mobi data structures. At least for the KindleBook side of things anyway.
I think they're making a mistake by dropping support for the standalone tools, myself. It's much easier to debug standalone python scripts. Plus there's no need to get new contributors "up to speed" on calibre's plugin api and plugin dev best-practices just to start offering assistance. Bringing everything under one roof has been great for users, but it's killed the notion of the "casual contributor" to the codebase. But it's their decision. No skin off my back. *shrug*