![]() ## Summary This PR starts the process of adding debug task locators to Zed's debugger system. A task locator is a secondary resolution phase that allows a debug task to run a command before starting a debug session and then uses the output of the run command to configure itself. Locators are most applicable when debugging a compiled language but will be helpful for any language as well. ## Architecture At a high level, this works by adding a debug task queue to `Workspace`. Which add's a debug configuration associated with a `TaskId` whenever a resolved task with a debug config is added to `TaskInventory`'s queue. Then, when the `SpawnInTerminal` task finishes running, it emits its task_id and the result of the ran task. When a ran task exits successfully, `Workspace` tells `Project` to start a debug session using its stored debug config, then `DapStore` queries the `LocatorStore` to configure the debug configuration if it has a valid locator argument. Release Notes: - N/A |
||
---|---|---|
.. | ||
src | ||
test_data | ||
Cargo.toml | ||
LICENSE-GPL | ||
README.md |
This contains the code for Zed's Vim emulation mode.
Vim mode in Zed is supposed to primarily "do what you expect": it mostly tries to copy vim exactly, but will use Zed-specific functionality when available to make things smoother. This means Zed will never be 100% vim compatible, but should be 100% vim familiar!
The backlog is maintained in the #vim
channel notes.
Testing against Neovim
If you are making a change to make Zed's behavior more closely match vim/nvim, you can create a test using the NeovimBackedTestContext
.
For example, the following test checks that Zed and Neovim have the same behavior when running *
in visual mode:
#[gpui::test]
async fn test_visual_star_hash(cx: &mut gpui::TestAppContext) {
let mut cx = NeovimBackedTestContext::new(cx).await;
cx.set_shared_state("ˇa.c. abcd a.c. abcd").await;
cx.simulate_shared_keystrokes(["v", "3", "l", "*"]).await;
cx.assert_shared_state("a.c. abcd ˇa.c. abcd").await;
}
To keep CI runs fast, by default the neovim tests use a cached JSON file that records what neovim did (see crates/vim/test_data), but while developing this test you'll need to run it with the neovim flag enabled:
cargo test -p vim --features neovim test_visual_star_hash
This will run your keystrokes against a headless neovim and cache the results in the test_data directory. Note that neovim must be installed and reachable on your $PATH in order to run the feature.
Testing zed-only behavior
Zed does more than vim/neovim in their default modes. The VimTestContext
can be used instead. This lets you test integration with the language server and other parts of zed's UI that don't have a NeoVim equivalent.