That's pretty sweet. I'd love to see something similar for the EL family: RHEL, CentOS, Fedora.
There's certainly a lot of opportunity, I think, to improve the Linux package management ecosystem. For example, Spacewalk is kind of terrible and I'd love build some projects implementing smaller parts of it (like checking security errata, SCAP audits), but working with rpm is a real bummer.
Career-wise, I came from a proprietary package management system to a red hat based one, and I don't know what you're talking about. Why do you think that spacewalk is kinda terrible, and why is working with rpm a real bummer?
Personally, I only have good experiences with spacewalk, rpm, yum, dnf, mock, koji, kojira, &c, but I am also aware that there may be other, better systems out there.
in what seems to be a brand new implementation (though i failed to find source links) instead of using the 2 or 3 readily available C apis that can be ingested easily in tons of languages?
cgo has its own set of problems, and there is no single C API. The version comparison isn't in shared library form, and breaks API a lot. The C parser for dependencies isn't the same we use in control files, and there's no C API for parsing a deb (or unpacking control files).
Most of this just has to be done. I take your point, but I couldn't find much. And I'm a debian developer - who knows what most non-Debian people would find. The best APIs are in `libapt`, which is C++, and cgo won't work with that all that well.
So, yes, you're right. This is yet another project. That's uncool, but there's not a ton of other ways out, short of me making a clean C shared library and get everyone to switch to that :)
There's certainly a lot of opportunity, I think, to improve the Linux package management ecosystem. For example, Spacewalk is kind of terrible and I'd love build some projects implementing smaller parts of it (like checking security errata, SCAP audits), but working with rpm is a real bummer.