BREAKING CHANGE: `Driver.proxyCommand` now returns `Promise<unknown>` by default. In strict mode, this forces an extension calling `this.proxyCommand()` to explicitly type the return value (because it's unknown!).
Otherwise, this concerns the `SingularSessionData` type and the function which returns it (`getSession()`). `Driver` now accepts a type param to add arbitrary properties to the value returned by `getSession()`--as this is in-line with `XCUITestDriver` is doing. This param is provided as the new optional second type param to `SingularSessionData`.
Further, it removes the `ISessionCommands` interface, because I didn't understand how to write it such that an isolated implementation could use the type parameter from the class it belongs to. This is probably doable, but seems icky, and I didn't want to fight with it. It's easier to just move the implementation into the class itself, which is what will happen in `BaseDriver` (see next changeset).
AWS S3 expects file field to be the last field in the POST request.
Currently it is the first field and that prevents AWS from parsing the
other POST fields and request gives 400 error.
With the change, it makes it possible to provide formFields necessary
for http upload to S3 from appium.
Use case:
mobile: stopMediaProjectionRecording can upload screen recording to S3
for further processing
The resulting solution keeps a bunch of weak refs to `Consola` objects and sets the log level on all of them if it ever changes.
This is--at minimum--easier to understand than the broken `Proxy` implementation.
This makes `ls` an alias of `list`; e.g., `appium driver ls`
While doing this I noticed a limitation of `argparser`, which is that an option (not a command) can only have a single (1) alias. I have noted this behavior in a comment.
I also notice that we need to rewrite `ls` to be `list` ourselves, or we will have a bunch of conditionals checking for both values scattered around.
We need to use yargs...
In the case of `appium-xcuitest-driver`, each command (or execute) method is declared like so:
```ts
import {someMixin} from './some-mixins';
class XCUITestDriver {
foo = someMixin.foo;
}
```
The docstring exists on `someMixin.foo`, but not on the `foo` prop of an `XCUITestDriver` instance.
This change makes `@appium/typedoc-plugin-appium` dig into such references and derive the comments.
These can now return `T` and `T[]`, respectively. `T` defaults to `string`, but in `appium-xcuitest-driver`, `T` is either `string` or an object, depending on the settings.
- Doc gen for:
- `@appium/driver-test-support`
- `@appium/opencv`
- `@appium/plugin-test-support`
- `@appium/test-support`
- Added a `dev:docs` script in the root monorepo. The file watcher does _not_ presently rebuild reference docs upon changes to sources (which is a limitation of `@appium/docutils`).
- Changed the `dev:docs:*` scripts in `appium` to use `appium-docs` (so it rebuilds reference docs at least once [because it runs typedoc!!])
This seems to fix the bug in #18379. There must be some sort of bug in TypeScript w/r/t `@implements` or maybe even it was not treating `@typedef` as a proper alias. Perhaps it's user error--omitting optional type args in alias-style typedefs may end up making a new type entirely. idk, but this was easier to deal with.
Closes#18379.
Another handful of fixes.
- remove second "extra" parameter from `Capabilities`, `NSCapabilities` `W3CCapabilities`. we were not using it anywhere to my knowledge. it was an escape hatch anyway.
- never assign `BaseDriverCapConstraints` as the default constraints, because this is confusing as hell. instead, drivers should use the `*DriverCaps` types which automatically pulls the base constraints in. applied across several workspaces
- renamed `DeviceSettings` => `IDeviceSettings` because `DeviceSettings` is a class.
- the session ID param of `Core`'s `sessionExists()` is actually optional (per the implementation)
- allow some modicum of type narrowing at the class level around settings and values returned by `ISessionHandler`'s methods. also, not all values in the raw settings object are of the same type `T`, so fix that
- `server`, `serverHost`, `serverPort` and `serverPath` are not members of `Core` (using `AppiumDriver` as a litmus test--it interacts with `Driver` types but does not have these properties itself)
- changed a few types to interfaces where possible
- moved a lost comment
- further dumbed-down the logdef stuff
- `Driver.driverData` is _always_ defined, since `Core` makes it an empty array.
- `DriverOpts` no longer contains "partial" caps; this it not quite right as the caps are not assigned until a session is created, but there's a conflict between some props which are _both_ server args and defined in constraints. anyway, this is what `initialOpts` is for
- fix some type disagreements in basedriver's "capabilities" module
The first arg of `PluginCommand` is a `NextPluginCallback` which now returns `Promise<unknown>` instead of `Promise<void>`, which was incorrect.
We use `unknown` rather than `any` because it forces us to be explicit about the return value of `await next()` if we are to call it.
Added a type `DriverCommandToPluginCommand` which essentially converts a `DriverCommand` to a `PluginCommand`. It's probably best used as a type guard or decorator, but it can "fill in the blanks" when overloading commands in a plugin.
Also normalized some of the `TArgs`/`TReturn` stuff.
mainly because a "command" is a thing in drivers/plugins, and these classes refer to something else completely unrelated. renamed to `*CliCommand` to avoid confusion.
Christopher Hiller