In the past days I was more unwillingly than willingly trying to hunt down some bugs in my MP3::Mplib module so that I'm eventually able to parcel an initial release for CPAN. As I've written in an older entry its author has been most helpful and fixed all issues being in his responsibility. Yet, knowing that a particular library that one is going to use behaves as it should is unpleasent. It means that any peculiarities from then on are yours. ;-)
But now I seem to have finished the XS part. What needs to be done is checking the corresponding .pm file for completeness, both API- and documentation-wise. Also, I produced a terrible mess in the t/ directory in my attempts to find and fix the bugs. Subsequently the MANIFEST.SKIP has grown almost as large as the MANIFEST file.
And now the most tricky decision is still ahead of me: How far should I go in explaining the id3tag specifications? Problem is that those for the id3v2 tags are anything but trivial: 74 different frames, they all can show up arbitrarily often. A lot of them have a special structure which in turn requires special functions to parse and create them. So do people rather want simple and fast access or do they prefer to have control over each tiny bit of the specifications? Currently these two are mutually exclusive (till I, if ever, add MP3::Mplib::Lowlevel or so). So I opted out the second alternative, also in the hope of reducing mail-traffic from puzzled users.
Anyway, I need to get back to my work. I really can't await to release my first XS module. I am quite curious as to how well (or even at all) the module will work across different operating systems etc. I am sure the C involved will make module-maintenance a whole new experience. :-) :-)
