Closes https://github.com/zed-industries/zed/issues/38690 Closes #37353 ### Background On Windows, paths are normally separated by `\`, unlike mac and linux where they are separated by `/`. When editing code in a project that uses a different path style than your local system (e.g. remoting from Windows to Linux, using WSL, and collaboration between windows and unix users), the correct separator for a path may differ from the "native" separator. Previously, to work around this, Zed converted paths' separators in numerous places. This was applied to both absolute and relative paths, leading to incorrect conversions in some cases. ### Solution Many code paths in Zed use paths that are *relative* to either a worktree root or a git repository. This PR introduces a dedicated type for these paths called `RelPath`, which stores the path in the same way regardless of host platform, and offers `Path`-like manipulation APIs. RelPath supports *displaying* the path using either separator, so that we can display paths in a style that is determined at runtime based on the current project. The representation of absolute paths is left untouched, for now. Absolute paths are different from relative paths because (except in contexts where we know that the path refers to the local filesystem) they should generally be treated as opaque strings. Currently we use a mix of types for these paths (std::path::Path, String, SanitizedPath). Release Notes: - N/A --------- Co-authored-by: Cole Miller <cole@zed.dev> Co-authored-by: Piotr Osiewicz <24362066+osiewicz@users.noreply.github.com> Co-authored-by: Peter Tripp <petertripp@gmail.com> Co-authored-by: Smit Barmase <heysmitbarmase@gmail.com> Co-authored-by: Lukas Wirth <me@lukaswirth.dev>
Inspector
This is a tool for inspecting and manipulating rendered elements in Zed. It is only available in debug builds. Use the dev::ToggleInspector action to toggle inspector mode and click on UI elements to inspect them.
Current features
-
Picking of elements via the mouse, with scroll wheel to inspect occluded elements.
-
Temporary manipulation of the selected element.
-
Layout info for
Div. -
Both Rust and JSON-based style manipulation of
Divstyle. The rust style editor only supports argumentlessStyledandStyledExtmethod calls. -
Navigation to code that constructed the element.
Known bugs
JSON style editor undo history doesn't get reset
The JSON style editor appends to its undo stack on every change of the active inspected element.
I attempted to fix it by creating a new buffer and setting the buffer associated with the json_style_buffer entity. Unfortunately this doesn't work because the language server uses the version: clock::Global to figure out the changes, so would need some way to start the new buffer's text at that version.
json_style_buffer.update(cx, |json_style_buffer, cx| {
let language = json_style_buffer.language().cloned();
let file = json_style_buffer.file().cloned();
*json_style_buffer = Buffer::local("", cx);
json_style_buffer.set_language(language, cx);
if let Some(file) = file {
json_style_buffer.file_updated(file, cx);
}
});
Future features
-
Action and keybinding for entering pick mode.
-
Ability to highlight current element after it's been picked.
-
Info and manipulation of element types other than
Div. -
Indicate when the picked element has disappeared.
-
To inspect elements that disappear, it would be helpful to be able to pause the UI.
-
Hierarchy view?
Methods that take arguments in Rust style editor
Could use TreeSitter to parse out the fluent style method chain and arguments. Tricky part of this is completions - ideally the Rust Analyzer already being used by the developer's Zed would be used.
Edit original code in Rust style editor
Two approaches:
-
Open an excerpt of the original file.
-
Communicate with the Zed process that has the repo open - it would send the code for the element. This seems like a lot of work, but would be very nice for rapid development, and it would allow use of rust analyzer.
With both approaches, would need to record the buffer version and use that when referring to source locations, since editing elements can cause code layout shift.
Source location UI improvements
-
Mode to navigate to source code on every element change while picking.
-
Tracking of more source locations - currently the source location is often in a ui component. Ideally this would have a way for the components to indicate that they are probably not the source location the user is looking for.
-
Could have
InspectorElementIdbeVec<(ElementId, Option<Location>)>, but if there are multiple code paths that construct the same element this would cause them to be considered different. -
Probably better to have a separate
Vec<Option<Location>>that uses the same indices asGlobalElementId.
-
Persistent modification
Currently, element modifications disappear when picker mode is started. Handling this well is tricky. Potential features:
-
Support modifying multiple elements at once. This requires a way to specify which elements are modified - possibly wildcards in a match of the
InspectorElementIdpath. This might default to ignoring all numeric parts and just matching on the names. -
Show a list of active modifications in the UI.
-
Support for modifications being partial overrides instead of snapshots. A trickiness here is that multiple modifications may apply to the same element.
-
The code should probably distinguish the data that is provided by the element and the modifications from the inspector. Currently these are conflated in element states.
If support is added for editing original code, then the logical selector in this case would be just matches of the source path.
Code cleanups
Consider removing special side pane rendering
Currently the inspector has special rendering in the UI, but maybe it could just be a workspace item.
Pull more inspector logic out of GPUI
Currently crates/gpui/inspector.rs and crates/inspector_ui/inspector.rs are quite entangled. It seems cleaner to pull as much logic a possible out of GPUI.
Cleaner lifecycle for inspector state viewers / editors
Currently element state inspectors are just called on render. Ideally instead they would be implementors of some trait like:
trait StateInspector: Render {
fn new(cx: &mut App) -> Task<Self>;
fn element_changed(inspector_id: &InspectorElementId, window: &mut Window, cx: &mut App);
}
See div_inspector.rs - it needs to initialize itself, keep track of its own loading state, and keep track of the last inspected ID in its render function.