This is something I sent out a couple of days ago to a few people, and was asked to repost to this list for discussion. I know that parts of this kit are being worked on. I'd like to volunteer to work on the documentation angle of this. ------------------------------------ When talking to developers now, and reaching beyond those that might work directly on the spec/tests/impl and are simply interested in making lsb packages of their work, I find it difficult to hand them something that they can make use of. I realize that it is still very early for this type of participation, but I think you'll find a lot more developers willing to work with alpha/beta quality development tools than you might think. It might also just be helpful to get developers in the mindset of building lsb apps, and familiar with the process. We'll end up with more beta testers for the test suites as well. I envision some sort of "developer's kit", which would contain an rpm/deb and a short instruction manual. Instructions might include: * Installing the rpm/deb on their system * export CC='cc-lsb' * Make the application. (I'm guessing that cc-lsb will default to headers in /usr/include/lsb or equivalent.) * Use lsbappchk on the finished binary(ies) * -In the beta stages- Please report any unexpected behavior from lsbappchk * Package using rpm. (Should we provide an rpm-lsb which is the right version of rpm?) * Offer their lsb app up for review I think that the overall process isn't _that_ hard, but those not totally familiar with the pieces of lsb may find difficulty unless this is documented well. Also, I'm pretty sure we don't have that rpm/deb I'm speaking of. Andrew, I know you're working on an OS check binary RPM, but it sounded in this morning's call like that a sample implementation/developer's type rpm wasn't yet in the works. Is that correct? I'd be able to work on the documentation side of this, but don't feel comfortable being any authority on building apps for lsb. 1) Does this type of kit seem appropriate eventually? 2) Could we build such a thing soon, or is it too early? (I think that this is a pretty important step to getting feedback on a lot of things, so I'm voting for earlier rather than later with this.) 3) If we built an alpha version of this soon, would it be at all useful, or would it be so alpha that a final version would differ greatly? (Even if the process is right, that will offer a significant help in educating people to building lsb apps.) 4) Is my general bullet outline of the process even close to correct? Thanks. -drew -- M. Drew Streib <dtype@dtype.org> | http://dtype.org/ FSG <dtype@freestandards.org> | Linux International <dtype@li.org> freedb <dtype@freedb.org> | SourceForge <dtype@sourceforge.net>
Attachment:
pgpCQKeCjU923.pgp
Description: PGP signature