Closed naps62 closed 10 months ago
is it possible to make a minimal repro? I've never encountered the hid_error is not implemented yet
error.
by overriding the dependency to force ethers-rs to use it. since it's a minor version bump I assume no intended breaking changes
we're pre-1.0, so minor version bumps are still eligible for breaking changes. For this particular change, if you're using use coins_ledger::Ledger
there should be no migration required. if you're using lower-level types, there may be breakage
@prestwich here you go: https://github.com/naps62/coins_ledger-issue-128/tree/main
after narrowing this down, actually found it to be 2 different issues, though they're probably related. see the comments on main()
for details.
this reproduces both issues consistently on my system (Manjaro Linux, Ledger Nano X)
tested with both 0.9.0 and 0.9.1
thanks! looking now :)
So this update makes it no longer panic. Instead a useful error message is added: https://github.com/summa-tx/coins/pull/129
Can you check that this resolves your issues? I ran it against your repo and it prevents panicking, although i'm not sure I completely understood your Issue 1
ffr, you can use the following in your cargo.toml to patch subdeps:
[patch.crates-io]
coins-ledger = { git = "...", tag = "..." }
https://doc.rust-lang.org/cargo/reference/overriding-dependencies.html
ffr, you can use the following in your cargo.toml to patch subdeps:
was wondering why you pointed this out, since I'm aware. just noticed I didn't properly override this dep on my repo above. was late at night, sorry :smile:
My app originally relied on my proposal (#125 ), which was closed in favor of a larger rewrite in the latest version.
I'm trying to run my app using this rewrite (by overriding the dependency to force ethers-rs to use it. since it's a minor version bump I assume no intended breaking changes?)
But any attempt to use it result in an immediate panic followed by
PoisonError
on every future attempt:One thing I should note, and that already happened even before all these fixes, is that I was already getting sporadic errors of the form
hid_error is not implemented yet
, which seems to come directly fromhid.c
, so unfortunately hard to debug. I had these requests around a retry block, since they would randomly fail. But this new update means any failure leads to a poisoned mutex