see whatever…

I just upload a new cssutils alpha. While spending most of the last weeks of development (whenever I had some spare time) added more unittests I found quite a few bugs. I added the @media rule in this version and also some minor enhancements in the parser and serializer. I guess I am still not sure if cssutils will become a lib which actually is able to pretty print a given stylesheet without removing any relevant information like various CSS hacks. This was the original intention but then I started implementing the CSS DOM. Something like

a { w\idth: 10px; width: 20px }

is used for crossbrowser development frequently but a DOM implementation should only see “width: 20px”. I’d like to keep both values which is not too difficult but actually clashes sometimes while implementing the official specs. Ideally I’d like to have both, a CSS DOM implementation and something similar to a CSS editor…

Well, quite a few things to do but I guess this alpha will be the final 0.9 release. I intend to put some new stuff in the 0.91 release which is worth the version change.

I know the cssutils documentation is absolutely lacking. Today I started rewriting the cssutils main page to actually reflect the current version 0.9a5 which I released yesterday. During final testing of the provided examples I found quite a few differences to the older versions so the docs had to be adjusted quickly. I hope now at least the basic stuff is explained a bit. I included short examples how to parse, customize logging and serializing (both may change a bit but should generally work like now) and also how to build a stylesheet from scratch.

Sticking more to the W3C DOM API actually changed the API quite a bit. But I think it generally is quite logic and also the W3C DOM spec (css, stylesheets) may (partly) be used as a reference. I added some classes mainly for CSSStyleDeclaration Selector handling which seemed to have changed anyway in the current CSS version but that is not finished yet but still a working draft only.

I still have quite a long list of todos which is only slowly getting shorter (actually growing in between) which is also not nice as I do not have the spare time I would like to. But I think small steps will bring this thing a bit further and the worst parts (I hope are done.

Some people may think that providing a non-conformant (or not complete) implementation of a standard is hardly worth the effort but I guess something is better than nothing. Even browsers do not implement the complete CSS spec (ok, admittedly more than this lib but hey it’s only kind of a hobby) so all in all not so bad, is it not?