Since locals are handled during parsing instead of highlighting with
tree-house, we need to call `helix_core::syntax::Loader::set_scopes`
before parsing any documents. During `:config-reload` we previously
reloaded the `Loader` and re-parsed documents and _then_ updated the
theme. So documents were parsed before `Loader::set_scopes` was called
on the new loader.
With this change the `refresh_language_config` helper is inlined into
`refresh_config`. Updating the `Editor`'s `ArcSwap` of the loader is
done before updating the theme so that the `load_configured_theme`
helper can call `set_scopes` with on the new loader.
This function was never used and will be superseded by
`RopeSliceExt::is_grapheme_boundary` (which accepts a byte index rather
than a character index) once we transition to Ropey v2. In the meantime
any callers should convert to byte index and use the `RopeSliceExt`
extension rather than form new dependencies on this.
This fixes a regression from the switch to tree-house with one of the
custom predicates in indent queries: `#not-kind-eq?`. This predicate
should be allowed to be written multiple times in a pattern. For example
in the Go indents:
; Switches and selects aren't indented, only their case bodies are.
; Outdent all closing braces except those closing switches or selects.
(
(_ "}" @outdent) @outer
(#not-kind-eq? @outer "select_statement")
(#not-kind-eq? @outer "type_switch_statement")
(#not-kind-eq? @outer "expression_switch_statement")
)
So instead of an `Option<T>` of one we need a `Vec<T>` and we need to
check that all of these predicates are individually satisfied (basically
`iter().all(/* node kind is not expected kind for that capture */)`).
Previously `recv` for new messages from the language server or debug
adapter allocated a fresh Vec for each message. Instead we can reuse
the buffer. This resolves TODO comments.
Co-authored-by: Rolo <roloedits@gmail.com>
With a directory with spaces in the name (for example
`mkdir -p 'Temp/Abc Def'`), completing `Temp/Ab` would create a
completion item `'Temp/AbAbc Def'`. Now it correctly completes
`'Temp/Abc Def'`
This avoids using any custom configuration in a user-defined
`languages.toml` config for the syntax test cases. The test cases should
only use the builtin `languages.toml` config.
Also the xtask crate reimplemented `default_lang_loader` and
`default_lang_config`. These functions are replaced with calls into
`helix_core`.
Introduces a new command `expand_selection_around` that expands the
selection to the parent node, like `expand_selection`, except it splits
on the selection you start with and continues expansion around this
initial selection.
This allows using multiple distinct state histories. By default, all
history is also cleared any time a view's selection is set, unless
explicitly told to save the state. This way, we can have control over
when selection history is saved. They are also cleared on any text
edit, since an edit could invalidate the previous selection, potentially
causing a panic.
Additionally, the object selections have been moved into `Document`
so that it is easier to manipulate them when changes to the document
happen. They have been put into a wrapper struct named `ViewData`, where
the intention is that any further fields that we want to add in the
future that must be associated with a view, but are more convenient to
store in a document, can be added here, instead of further polluting the
core `Document` type.
Adds two helper functions to `Selection`:
* `overlaps`: tests whether two `Selection`s contain any ranges which
overlap with each other
* `without`: Computes a new `Selection` that is the set difference
of two `Selection`s, i.e. everything in the first `Selection`
with everything that overlaps in the second `Selection` removed,
potentially leaving holes in the original ranges.
It also fixes a bug with `Selection::contains`: it assumes that if the
second `Selection` has a greater number of ranges than the first, then
the first cannot contain the second; but this is false, since one range
from the first could contain multiple smaller ranges in the second.
This changes the behavior of `shrink_selection` to iterate through child
nodes until it finds one that is contained within the selection, with
at least one of the ends of the selection being exclusively inside the
starting selection (though not necessarily both ends). This produces
more intuitive behavior for selecting the "first logical thing" inside
the selection.