`language_servers` is a `HashMap<String, Arc<Client>>` so the clients
are cheap to clone but not the language server names. The names are
unused in the save future so we can avoid the unnecessary clones by
looking only at the `HashMap::values`.
This change switches out the terminal manipulation library to one I've
been working on: Termina. It's somewhat similar to Crossterm API-wise
but is a bit lower-level. It's also influenced a lot by TermWiz - the
terminal manipulation library in WezTerm which we've considered
switching to a few times.
Termina is more verbose than Crossterm as it has a lower level interface
that exposes escape sequences and pushes handling to the application.
API-wise the important piece is that the equivalents of Crossterm's
`poll_internal` / `read_internal` are exposed. This is used for reading
the cursor position in both Crossterm and Termina, for example, but also
now can be used to detect features like the Kitty keyboard protocol and
synchronized output sequences simultaneously.
This is meant to be minimal for now and is expected to change as the
config system evolves.
Features like word completion should be able to hook into this to
initialize or clear the word index when the toggle for the feature is
turned on or off (respectively).
On Windows for example the behavior of this function typically diverges
from the usual behavior on Unix. Instead of checking that the inserted
string starts with `'\n'` (untrue for for CRLF line endings) we need to
check that the first grapheme cluster in the string is a line ending.
(All line endings are single grapheme clusters.)
This type also exists on `Editor`. This change brings it to the
`Document` as well because the replacement for `Syntax` in the child
commits will eliminate `Syntax`'s copy of `syn_loader`. `Syntax` will
also be responsible for returning the highlighter and query iterators
(which will borrow the loader), so the loader must be separated from
that type.
In the long run, when we make a larger refactor to have
`Document::apply` be a function of the `Editor` instead of the
`Document`, we will be able to drop this field on `Document` - it is
currently only necessary for `Document::apply`. Once we make that
refactor, we will be able to eliminate the surrounding `Arc` in
`Arc<ArcSwap<syntax::Loader>>` and use the `ArcSwap` directly instead.
When closing a document we must wait until all views have been updated
first - either replacing their current document or closing the view -
before we remove the document from the `documents` map. The
`Editor::_refresh` helper is called by `Editor::close`. It accesses each
View's Document to sync changes and ensure that the cursor is in view.
When closing multiple Views at once, `Editor::_refresh` will attempt
to access the closing Document while refreshing a to-be-closed View.
This fixes a regression from 6da1a79d80. `:buffer-close` on an
unmodified document would cause later panics since the document should
not have been removed. Instead of eagerly removing the document on the
first line we need to wait until we've checked that it's unmodified.
Previously the `call` helper (and its related functions) returned a
`serde_json::Value` which was then decoded either later in the client
(see signature help and hover) or by the client's caller. This led to
some unnecessary boilerplate in the client:
let resp = self.call::<MyRequest>(params);
Some(async move { Ok(serde_json::from_value(resp.await?)?) })
and in the caller. It also allowed for mistakes with the types. The
workspace symbol request's calling code for example mistakenly decoded a
`lsp::WorkspaceSymbolResponse` as `Vec<lsp::SymbolInformation>` - one of
the untagged enum members (so it parsed successfully) but not the
correct type.
With this change, the `call` helper eagerly decodes the response to a
request as the `lsp::request::Request::Result` trait item. This is
similar to the old helper `request` (which has become redundant and has
been eliminated) but all work is done within the same async block which
avoids some awkward lifetimes. The return types of functions like
`Client::text_document_range_inlay_hints` are now more verbose but it is
no longer possible to accidentally decode as an incorrect type.
Additionally `Client::resolve_code_action` now uses the `call_with_ref`
helper to avoid an unnecessary clone.
This adds events for:
* a document being opened
* a document being closed
* a language server sending the initialized notification
* a language server exiting
and also moves some handling done for these scenarios into hooks,
generally moving more into helix-view. A hook is also added on
`DocumentDidChange` which sends the `text_document_did_change`
notification - this resolves a TODO in `document`.
This resolves a TODO in the core diagnostic module to refactor this
type. It was originally an alias of `LanguageServerId` for simplicity.
Refactoring as an enum is a necessary step towards introducing
"internal" diagnostics - diagnostics emitted by core features such as
a spell checker. Fully supporting this use-case will require further
larger changes to the diagnostic type, but the change to the provider
can be made first.
Note that `Copy` is not derived for `DiagnosticProvider` (as it was
previously because `LanguageServerId` is `Copy`). In the child commits
we will add the `identifier` used in LSP pull diagnostics which is a
string - not `Copy`.
This is a minor move that will make future refactors of code actions
simpler. We should be able to move nearly all code action functionality
into `helix-view`, save UI stuff like the `menu::Item` implementation
and dealings with the compositor.
