Is TagLib alive?

[Scott Wheeler]

> On Jan 17, 2020, at 15:23, Fred Gleason <fredg at paravelsystems.com> wrote:
> 
> On Jan 14, 2020, at 12:07, Scott Wheeler <scott at taglib.org> wrote:
> 
>> My general opinion is that TagLib does need bug fixes, but that it doesn't need a lot of new development.  The project is 15 years old and has a fixed scope, so I see it more as "mature" rather than "dead".  ;-)
> 
> It would be great to see support for additional tagging schemas; CartChunk (aka AES46-2002, http://www.aes.org/publications/standards/search.cfm?docID=41) and Broadcast Wave File (https://tech.ebu.ch/docs/tech/tech3285.pdf) being two that come immediately to mind. I know that that CartChunk support has been discussed here before (see for example https://mail.kde.org/pipermail/taglib-devel/2004-August/000021.html), but my understanding is that it has never happened due to the need to maintain a stable ABI within the larger KDE project.

To be honest, if I end up still being in the maintainer's seat for a 2.0 release, I intend to remove, rather than add additional formats.  The problem with a lot of obscure formats is that they end up being dropped on TagLib, often with code in a questionable state, and then the maintainer ends up having to try to fix bugs for formats they've never even heard of and are not widely used (and more importantly, don't know the spec for at all -- meaning trivial bugfixes can mean a lot of reading in advance).

That said, removing said formats would come with finally fixing the file type resolver API so that third parties can easily add support for formats they're interested in.  Moving those formats out of TagLib would allow their maintenance status to vary independently of TagLib's.

> Is this ABI compatibility requirement still operative? Would a PR that implements support for the two above tagging schemas be considered for inclusion in stock TagLib (perhaps as part of a new major version)?

Yes on API requirements, and honestly, no on a PR for those formats.

-Scott