We currently initialize the `ArtiNativeKeyStore` with a dummy root dir,
so when `ArtiNativeKeyStore` starts validating directories, this code
will start to fail. Let's preemptively ignore any errors coming from
`ArtiNativeKeyStore::new`. This is temporary and will be removed when we
introduce the key store config (and a real default value for the
keystore root dir).
(Taken from tpo/core/arti!1113 and squashed by Ian Jackson, with
conflicting hunk in StreamPrefs struct skipped.
The setter name is wrong, the cfg feature is wrong, there are no
docs, and the TODO is still there. These will be fixed in a moment.)
We need to allow some lints in the dummy key manager because its
implementation needs to mirror that of `tor_keymgr::KeyMgr` (so we can't
apply the API changes suggested by clippy).
TorClient now only uses the tor_keymgr::KeyMgr implementation if the
keymgr experimental feature is enabled. If the feature is disabled, a
dummy key manager implementation is used.
The new `keymgr` feature depends on `onion-client`, because the key
manager is only used for HS client auth.
This simplifies usage quite a bit and will enable us to implement a
dummy `KeyMgr` that doesn't depend on the error types from tor-keymgr
(which will replace the "real" `KeyMgr` if the keymgr feature is
disabled).
The HS `HsClientSpecifier` and `HsClientSecretKeySpecifier` are moved to
`tor-hsclient`. The HS service secret key specifier stubs are moved to
`tor-hsservice`.
The `KeyMgr` is now initialized with an `ArtiNativeKeyStore` built from
an invalid key store root dir (this is alright for the purposes of this
proof-of-concept, since `ArtiNativeKeyStore::new` won't fail as it
doesn't currently validate the keystore root dir).
This means `KeyMgr` users don't need to specify the underlying key type
(e.g. `ed25519::Keypair`) when retrieving keys. Instead, they can just
specify the type required (as long as it implements `ToEncodableKey`),
e.g. `HsClientIntroAuthKeypair`.
This is a rather tricky piece of functionality. It works as
follows.
We introduce a `CastTable` type. Each `CastTable` tells us how to
downcast `dyn Object` for objects of a single concrete type.
The `Object` type now has a `get_casttable` method that returns
an empty `CastTable` by default.
`CastTable` is, internally, a map from the `TypeId` of the target
dyn Trait reference type to a function
`fn(&dyn Object) -> &dyn Trait`. These functions are stored as
`Box<dyn Any + ...>`. (They are Boxed because they may refer to
generic functions, which you can't get a static reference to,
and they're Any because the functions have different types.)
The `decl_object!` macro now implements `get_casttable` as
appropriate. (The syntax is a bit janky, but that's what we get
for not using derive_adhoc.) For non-generic types, `get_casttable`
uses a Lazy<CastTable>`. to initialize a CastTable exactly once.
For generic types, it use a `Lazy<RwLock<HashMap<..>>` to
build one CastTable per instantiation of the generic type.
This could probably be optimized a bit more, the yaks could be
shaved in a more scintillating hairstyle, and the syntax for
generic `decl_object` could definitely be improved.
Now ClientCirc is no longer `Clone`, and the things that need it
to be `Clone` instead return and use an Arc<ClientCirc>
We're doing this so that ClientCirc can participate in the RPC
system, and so that its semantics are more obvious.
Closes#846.
Thanks to the type system, this was a much simpler refactoring than
I had feared it would be.
This does the following:
- Gives every crate a `full`.
- Cause every `full` to depend on `full` from the lower-level
crates.
- Makes every feature listed _directly_ in `experimental` depend
on `__is_experimental`.
One of the tests run as part of `cargo test -p arti-client` fails
because the `pt-client` feature is not enabled by default:
```
failures:
---- src/config.rs - config::BridgesConfig (line 221) stdout ----
error[E0432]: unresolved import `arti_client::config::pt`
--> src/config.rs:225:26
|
7 | use arti_client::config::pt::ManagedTransportConfigBuilder;
| ^^ could not find `pt` in `config`
error[E0599]: no method named `bridges` found for mutable reference `&mut BridgesConfigBuilder` in the current scope
--> src/config.rs:233:19
|
15 | builder.bridges().bridges().push(bridge_1);
| ^^^^^^^ private field, not a method
error[E0599]: no function or associated item named `default` found for struct `BridgeConfigBuilder` in the current scope
--> src/config.rs:236:48
|
18 | let mut bridge2_builder = BridgeConfigBuilder::default();
| ^^^^^^^ function or associated item not found in `BridgeConfigBuilder`
error[E0599]: no method named `bridges` found for mutable reference `&mut BridgesConfigBuilder` in the current scope
--> src/config.rs:247:19
|
29 | builder.bridges().bridges().push(bridge2_builder);
| ^^^^^^^ private field, not a method
error[E0599]: no method named `transports` found for mutable reference `&mut BridgesConfigBuilder` in the current scope
--> src/config.rs:255:19
|
37 | builder.bridges().transports().push(transport);
| ^^^^^^^^^^ method not found in `&mut BridgesConfigBuilder`
error: aborting due to 5 previous errors
Some errors have detailed explanations: E0432, E0599.
For more information about an error, try `rustc --explain E0432`.
Couldn't compile the test.
failures:
src/config.rs - config::BridgesConfig (line 221)
test result: FAILED. 5 passed; 1 failed; 3 ignored; 0 measured; 0 filtered out; finished in 2.10s
```
This disables the `BridgesConfig` doc test if the `pt-client` feature is not enabled.
Closes#843
Signed-off-by: Gabriela Moldovan <gabi@torproject.org>
These crates had no changes until just a moment ago. But since
we updated the versions on some of their dependents, they have now
changed themselves. Thus they get patchlevel bumps.
```
tor-rtmock
tor-protover
tor-socksproto
tor-consdiff
tor-chanmgr
tor-dirclient
tor-hsservice
```
For these crates, the changes are nontrivial, so we
_do_ bump the versions on which their dependent crates depend.
Fortunately, since they are all pre-1.0, we don't need to
distinguish semver-additions from other changes. (Except for arti,
which _is_ post-1.0, but gets a patchlevel bump anyway.)
These are unstable crates with breaking changes:
```
tor-hscrypto
tor-hsclient
```
These have new or extended APIs:
```
safelog
tor-bytes
tor-cell
tor-linkspec
tor-llcrypto
tor-proto
tor-cert
arti-client
```
These have new unstable APIs or features:
```
tor-netdoc
tor-circmgr (also broke some unstable APIs)
arti (is post-1.0)
```
These have bugfixes only:
```
caret
tor-dirmgr
```
I found that I had the bug where I forgot to call this function, and
reached
bad_api_usage!("The circuit launcher wasn't initialized")
The possibility of such a bug is a hazard of this API pattern.
We're going to want to reuse this.
Also: rename StreamPrefs::isolation and update the docs, to better
reflect its rather limited functionality. The new
TorClient::isolation is the only call site.
When I was trying to add HS support to these layers, I found I could
add a new variant to the `Host` enum but everything would still
compile even though I hadn't written the necessary implementation!
This method is a liability: when using it, one inevitably writes such
latent bugs.
Doing so doesn't seem like a good idea. It might even be some kind of
leak?
Found because I added a variant to `address::Host` for hidden
services, and noticed that the resolve code still compiled.
These crates have had small code changes, but no API additions:
tor-config
tor-socksproto
tor-cert
tor-chanmgr
tor-ptmgr
tor-guardmgr
tor-circmgr
tor-dirclient
tor-dirmgr
arti
tor-hsservice
tor-congestion
These crates have had API extensions:
fs-mistrust
tor-llcrypto
tor-bytes
tor-checkable
tor-linkspec
tor-netdoc
tor-persist
arti-client
These are API breaks, but the crates themselves are currently
100% experimental, so there's no need to bump the minor versions
according to our semver rules.
In !948 we renamed a couple of accessor functions, which is a
breaking change in `tor-cell`'s API.
In retrospect, perhaps we should have deprecated the old names and
added the new ones, so we wouldn't have to break the API. (This is
the only API break AFAICT since 1.1.0.)
These changes influence behavior, but not effect compatibility.
(If I messed up, and any crate except for `arti` has non-breaking
API changes, that's still fine, since they are all version
0.x.)
Change ErrorHint so that, internally, it just holds an enum with a
lightweight reference to whatever parts of the error it needs to
generate a hint. Then we can move the formatting logic into a
Display function for ErrorHint, and do away with ErrorDetail entirely.
Move the "hint" function into Error, and use Option rather than Result.
(I'm using Option here because it's not really an error case not to have
a hint; we just either have a hint, or we don't.)
This warning kind of snuck up on us! (See #748) For now, let's
disable it. (I've cleaned it up in a couple of examples, since
those are meant to be more idiomatic and user-facing.)
Closes#748.
There's no harm in doing this, since we'll try to take ownership of
the state soon as soon as we use it: We just want to try early, so
that we'll get it before we decide that we're using bridges.
Fixes bug #669; bug introduced by !889.
This makes a `pt_state` directory inside .local/share/arti (or the
local equivalent), right next to our existing `state` dir.
Ideally we would use a separate directory for each PT, but we have a
very fuzzy "what is a specific PT" notion.
Closes#667
If support is available at compile-time, then we construct a PtMgr
and register it with the ChanMgr. We keep a handle to it ourself so
that we can reconfigure it as needed.
Closes#659.
We now parse the `bridges.enabled` BoolOrAuto, and the
`bridges.bridges` list.
The `bridges.bridges` list is Vec<()> in the builder, and Vec<Void> in
the built config. Ie, it is simply a count, and vanishes in the built
config.
But this count triggers us to try to call build(), to try to parse
bridges, and to try to set and honour the enablement boolean.
The result is that the type system now ensures that if bridges are
disabled, but specified (either by listing them in the config, or
writing `enabled=true`), we inevitably try to insist that we have a
non-empty Vec<Void>, which is of course impossible.
There will be a test case too for those who think this too abstract a
way to guarantee this property :-).
Ticket #285 is closed and most of this is stable now we think.
(There are still a couple of stability warnings for specific types in
tor-config, which aren't exposed at the arti-client level.)