Lulu

The reason was because the callbacks it now stores. But thinking about it they are all boxed so moving alpm into a new thread doesn't move the contents. Something I...

Thinking about it more. Could could store something !Sync + !Send in the context and access it in another thread via the callbacks So it should not be Send. I...

trans prepare (and commit) currently look like this: ```rust pub fn trans_prepare(&mut self) -> Result { PkgInvalidArch(AlpmListMut>), UnsatisfiedDeps(AlpmListMut), Ok, } ``` This can be a bit awkward as you can't...

I'm considering inverting how the structs work for v3. instead of a Pkg struct that contains a pointer it will be &Pkg that contains an alpm_pkg_t. Then like &str, &Pkg...

Unions can be done in a way that doesn't require us to match on them: https://github.com/rust-lang/rfcs/blob/master/text/2195-really-tagged-unions.md

I got around to implementing all of this in https://github.com/archlinux/alpm.rs/tree/v3. Only think not done is transmuting the unions to enums directly. As due to padding that would require putting all...

Right. I don't really know how this ended up like this. I think I have have just been experimenting and accidentally committed the file I don't know.

Sorry bit busy. Also you removed the alpmext trait (even though it was dead code anyway). I think the trait should remain and these methods be on it. Especially if...

I guess that's probably a good idea now that it's an archlinux project and has some use.

Pushed tags for alpm and alpm-sys. Not sure about alpm-utils has I've been thinking of moving that to a separate git.