- Forbids unsafe code
- Adds a "Safety Dance" badge
- Fixes the GitHub Actions status badge
- Fixes up links that changed with the move to `iqlusioninc` org
Converts a bag of constant values (`YKPIV_INS_*`) into an enum
representing APDU instruction codes (a.k.a. `ins`).
Among other things, this makes the `Debug` output for `APDU` more human
meaningful, since it can print a text label for the instruction rather
than a code number, which is helpful in trace debugging.
This commit merges the `apdu` and `response` modules: the responses are
APDU responses, and so the two are related.
This also moves the `trace` logging into the APDU type, which allows it
to display `Debug` output for APDUs and responses, which makes it easier
to understand what's going on (and will be even better once instructions
are converted into an enum so you can actually see what's happening).
This adds an `untested` feature to any functions which have not yet been
tested live against a YubiKey device (which is presently pretty much
everything).
This sets a clear expectation of what is presently supported, and
additionally documents the status in the README (and a series of GitHub
issues).
Adds a `cargo build --all-features` to GitHub Actions' `test` step in
order to make sure that `untested` functionality still compiles.
This commit adds quite a bit of documentation about the current status
of the project, including links to GitHub issues for the different Rust
modules which map to specific pieces of functionality.
The intent is to track the current status of the project in the
README.md as that's more up-to-date than the docs.rs documentation
(which depends on a crate release to get updated).
Adds an extremely basic initial test to ensure that we are able to
connect to a YubiKey.
The test is marked `#[ignore]` in the hope that we can eventually start
adding tests which run in CI, e.g. against a mock card.
This also includes a fix for calculating the APDU size, since the ones
we were sending originally were overly long.
Callers of this function always pad up to `CB_PIN_MAX` with `0xFF`.
The logic being changed here was previously identical to the `_verify`
function in `ykpiv.c`:
https://github.com/Yubico/yubico-piv-tool/blob/8ba243f/lib/ykpiv.c#L1299
...but @str4d noticed this potentially allows a caller to send an
unpadded PIN, which may (or may not) be an issue.
It seems like given we're inside a while loop which also has this
conditional, the original code should've been fine, but this change
makes it closer to the original C code.
This commit contains a "big bang" refactor/rewrite which does the
following:
- Replaces all `SCard*` FFI calls with the `pcsc` crate, which provides
a safe, portable PC/SC API across Windows, macOS, and Linux
- Refactors the `util` module into modules representing the various
device functions and concepts, e.g. `certificate`, `key`, `mgm`
- Replaces all usage of `libc` with `std` functionality, and in many
places rewriting functionality to use safe code.
- Removes `ykpiv_` from all function names, and `Piv*` from type names.
In 20/20 hindsight I wish I had done this commit more incrementally so
as to make it easier to review. Que sera sera.
However, realistically we need to test all functionality on the device
to ensure that it actually works. Going forward I would like to put
pretty much all of the current code behind an `untested` cargo feature,
and then remove it for each bit of functionality we test.
Changes the `APDU` struct into a builder for serialized APDU messages.
This makes APDU construction safer and more idiomatic, and also caught a
few bugs in the process (missing templ from the C translation).
There was originally another `Error` type from the translation. Now that
it's gone, and we don't presently have a type just named `Error`, this
renames the current `ErrorKind` type now that the original was deleted.
Switches all of the previous `state->verbose`-gated `eprintln!` calls to
use macros from the `log` crate, trying to map them onto the previous
verbosity levels, more or less following this mapping:
0. off
1. error/info/warn (depending on context)
2. trace
This additionally includes a bunch of logic/branch reformatting (and
occasional missed constants), since getting rid of all the gating on
verbose provided ample opportunities to clean up the code. Hopefully I
didn't break too much in the process!
Carl Wallace