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.)
I was doing the tutorial to learn Helix, and it's more sensible to use
`e` than `w`.
If you're using `w`, you will need to add an extra space.
Example, assuming cursor is at |:
This sentence has incorrect words |behind it.
If you use `wc`, then you will get:
This sentence has incorrect words |it.
while `ec` will give you
This sentence has incorrect words | it.
Which enables you to drop removing and adding a space for no reason.
`InactiveQueryCursor::new` might reuse a query cursor from a
thread-local cache if one is available, rather than create a new cursor.
Currently tree-house does not reset cached cursors back to defaults
(i.e. byte range and match limit). For now we can patch around this here
but eventually this should be fixed in `tree-house` upstream. Then this
patch can be reverted.
In practice this caused textobjects like `]f` to get "stuck" trying to
move to the next function if it was out of the current view. This is
because the highlight query cursor sets the range of the cursor to the
current viewport. We can reset the byte range to defaults to fix the
textobject behavior.
The integration tests have recently become a bit flaky in that
individual tests can seem to hang forever. This needs further
investigation. In the meantime we should limit the total allowed time
for the test workflow to something reasonably low. The default timeout
is a very high 360min.
This fixes a regression from the refactor of the highlighters when
switching to tree-house. The old `StyleIter` used `renderer.text_style`
as the base style rather than `Style::default()` for syntax highlights.
The result was that any text not captured by a syntax highlight query
was styled with no foreground or background, defaulting to the
terminal's foreground/background. This could cause text in markdown
files to look off-colored depending on your terminal configuration.
(Though you wouldn't notice if your 'ui.text' theming matches your
terminal's theming.)
Previously the statusline `write` function only accepted a string
and optional Style, so all rendering functions converted text to
strings. Some elements write spans with `&'static str`s, however, making
this unnecessary since `Span<'a>` is a wrapper around `Cow<'a, str>` and
style, and a `Span<'static>` would outlive all required lifetimes.
Moreover many elements could produce `Span<'a>` according to the
lifetime in `RenderContext` in the future, potentially re-borrowing from
the Editor borrow, so this change could save allocations for many
file-type elements (with more future changes). This is not explored in
this patch since the statusline functions currently add bespoke padding
per-element, but with a future refactor to make spacing consistent this
could be possible.
This change refactors the write function to accept a `Span<'a>` and
rewrites some related code to fit the codebase better (preferring `for`
to iterator's `for_each` for example). The new code is more complicated
lifetime-wise but avoids allocations in these cases:
* spacer for mode name when a pane is not focused
* LSP spinner frames
* '●' (workspace) diagnostic indicators
* " W " workspace diagnostic prefix
* file modification indicators
* read-only indicators
* spacer element
... and opens the door to avoid allocation for file name elements in the
future.
If you configure a subset of severities to show in the workspace
diagnostics statusline element you can see the 'W' (and surrounding
space) without any diagnostic indicators. This is the case by default
as it's configured to show warnings and errors only - if you have only
hints in your workspace like if you open `application.rs` in Helix for
example then you would see the 'W' and no indicators.
This change checks that any of the configured diagnostics are non-zero
and bails early if there are none.
The switch to tree-house accidentally dropped some shebang parsing code
from the loader's function to detect by shebang. This change restores
that. The new code is slightly different as it's using a `regex_cursor`
regex on the Rope rather than eagerly converting the text to a
`Cow<str>` and running a regular regex across it.
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.
This is for convenience when testing a few languages or themes, for
example while updating a language's parser and queries. query-check in
particular can take a while since parser initialization and query
analysis can each take some time and there are now many many languages.
Specifying the exact language makes the feedback loop much faster.
Binh Tran
Michael Davis
Nik Revenco
Saheed Adeleye
Christian Fredrik Johnsen
Ricardo Fernández Serrata
Kris Warner
Jérôme Tamba
dependabot[bot]
CalebLarsen
Nguyễn Hồng Quân
Rock Boynton
zoey
Nikita Revenco
Jengamon
RoloEdits
Aitor
Bekobi
omahs
Bryce Berger
Jules Wiriath
bloxx12
David Else
Sean Russell
Remo Senekowitsch