New arm target missing from cargo manifests

[MagaTailor]

The new armv7 target hasn't been introduced in the crates ecosystem yet, predictably resulting in errors like this:

/home/odroid/.cargo/registry/src/github.com-48ad6e4054423464/libssh2-sys-0.1.34/lib.rs:10:1: 10:26 error: can't find crate for `openssl_sys` [E0463]
/home/odroid/.cargo/registry/src/github.com-48ad6e4054423464/libssh2-sys-0.1.34/lib.rs:10 extern crate openssl_sys;
                                                                                          ^~~~~~~~~~~~~~~~~~~~~~~~~

BTW, @alexcrichton well spotted fixing the .mk file - I was going to open a PR myself.

I was also wondering if on the last line:
CFG_GNU_TRIPLE_armv7-unknown-linux-gnueabihf := armv7-unknown-linux-gnueabihf

it was really necessary to differentiate from arm-?

[alexcrichton]

I don't think this is a particularly actionable bug in Cargo itself, it'll just take time for the target to be found on the stable/beta compilers. If you use a nightly compiler to build the target then everything should work just fine. Cargo itself has no knowledge of targets or target triples, so it's just the underlying compiler that needs to be updated here (or perhaps a stale build directory?)

[MagaTailor]

I was using a bootstrapped compiler - otherwise I wouldn't even have the necessary rlibs to try this.

I'm well aware cargo itself doesn't know about targets, etc; when I jokingly wrote about "telling cargo" I meant stuff like this:

https://github.com/alexcrichton/ssh2-rs/blob/master/libssh2-sys/Cargo.toml#L37

or these:
https://github.com/alexcrichton/git2-rs/blob/master/libgit2-sys/Cargo.toml#L37
https://github.com/tomaka/glutin/blob/master/Cargo.toml#L96

So I meant you probably had the tools to automatically fix the affected crates.

EDIT:
Corrected the first link to point to the pertinent file. (to this issue)

[alexcrichton]

Oh yeah for those you can just send PRs to update them (or wait for the cfg-in-dependencies to get implemented)