Path completion items always have documentation but future core (i.e.
non-LSP) completions may not always have documentation - for example
word completion from the current buffer.
Completions are not specific to LSP anymore. In the child commits we
will expand on the types in this module so this refactor is done
eagerly to minimize changes later.
Previously LSP notifications were sent within a future and most callers
used a `tokio::spawn` to send the notification, silently discarding any
failures like problems serializing parameters or sending on the channel.
It's possible that tokio could schedule futures out of intended order
though which could cause notifications where order is important, like
document synchronization, to become partially shuffled. This change
removes the future wrapper and logs all internal failures.
Also included in this commit is the same change for `Client::reply`
which was also unnecessarily wrapped in a future.
Co-authored-by: Pascal Kuthe <pascalkuthe@pm.me>
This improves the display of the keymap popup for example, so that if
you bind a key like `C-x = ":buffer-close"` under the `<space>` menu,
the infobox shows "Close the current buffer." rather than `:buffer-close
[]`.
This simplifies the hover component by eagerly converting all
`lsp::Hover` responses into `Markdown`s. Previously we cached the
current `Markdown` and created a new `Markdown` when switching the
active response. Instead we can consume the `lsp::Hover` and avoid some
clones of its inner types.
The Hover component is used as the inner contents of a Popup. The Popup
should be doing calculations based on whether popup_borders is
configured and not Hover. This fixes an issue with hover rendering when
the popup border option is enabled for popups.
Fixes#12742
The language server may return `None` for a definition/reference
request. The parent commits introduced a regression for these commands
when a server did not provide locations. With this change a server may
respond with `null` and its locations will instead not be considered.
Fixes#12732
This covers all goto-definition-like commands: declaration, definition,
type definition and implementation.
Closes#11689
Co-authored-by: j <junglerobba@jngl.one>
<https://github.com/helix-editor/helix/pull/11486> introduced a Location
type in the LSP commands module which unified helpers like
`jump_to_location`. This change moves `OffsetEncoding` onto that type.
`SymbolInformationItem` and `PickerDiagnostic` already had fields for
carrying the offset encoding. We would want a similar setup for goto
definition/references as well (for supporting multiple language servers
with that feature) but those use the `Location` type. By moving
`OffsetEncoding` onto `Location` we make future changes to allow
mulitple language servers possible for goto definition/references
features and also simplify some calls for symbols and diagnostics.
The positions passed to `Transaction::change_by_selection` should be
character indexes. `edit_diff` is meant to track the number of
characters that should be deleted to erase the file name that has been
typed so far (if any). Mistakenly this was using `str::len` which is
the byte count. This fixes a bug that could cause more text to be
deleted than intended or a panic when completing a directory with
multi-byte characters like 'éclair'.
This change also moves the `edit_diff` binding out of the loop since
it's now performing some non-trivial work (counting characters, where
before it was just accessing the pre-computed number of bytes).
Previously you could use `<A-a><A-b>` to jump to a label "ab". We should
not treat characters with modifiers the same as characters without.
With this change the `<A-a>` input exits out of the jumping on-next-key.
Fixes#12695
The new `RopeSliceExt::ceil_char_boundary` from the parent commits can
be used to implement `RopeSliceExt::byte_to_next_char` when used with
`RopeSlice::byte_to_char`. That function had only one caller and that
caller will eventually disappear when we switch to Ropey v2 and drop
character indexing, so we can drop `byte_to_next_char` now and replace
its caller with `byte_to_char` plus `ceil_char_boundary`.
This change keeps the unit tests for `byte_to_next_char` and checks them
against a polyfill of `byte_to_char` plus `ceil_char_boundary` to ensure
that `byte_to_next_char`'s intended behavior is not changed.
This is a good example use-case of the `floor_char_boundary` and
`ceil_char_boundary` functions added in the parent commit. In the
single-width, single-selection case in `goto_file` we cap the search
to either the current line or 1000 bytes before or after the cursor
(whichever case comes earlier). That byte index might not lie on a
character boundary so it needs to be fixed to either the prior or
later boundary.
* fix: panic when pressing `*` at the end of the file
chore: remove incorrect additions
* docs: add info comment
* test: add new syntax to add a selection at the final character
* test: `*` panics when after the last char
* test: move into a more appopriate module
* test: fix failing
* test: account for Windows test suite
* test: choose a different strategy for custom syntax
* test: do not modify the syntax
* style: remove newline
---------
Co-authored-by: Nikita Revenco <154856872+NikitaRevenco@users.noreply.github.com>
This should help debug formatting failures when using external
formatters in the future. Previously we didn't log anything when an
external formatter failed despite having a custom error type for it.
This is partially a style commit:
* Pull more bindings out the `change_by_selection` closure like the
line-ending string and the comment tokens used for continuation.
* Prefer `Editor::config` to `Document`'s config.
The rest is changes to places where `insert_newline` may allocate.
The first is to move `new_text` out of the `change_by_selection`
closure, reusing it between iterations. This is not necessarily always
an improvement as we need to clone the text for the return type of the
closure. `SmartString`'s `From<String>` implementation reuses the
allocation when the string is too long to inline and drops it if it is
short enough to inline though which can be wasteful. `From<&String>`
clones the string's allocation only when it is too long to be inlined,
so we save on allocations for any `new_text` short enough to be inlined.
The rest is changes to `new_text.reserve_exact`. Previously calls to
this function in this block mixed up character and byte indexing by
treating the length of the line-ending as 1. `reserve_exact` takes a
number of bytes to reserve and that may be 2 when `line_ending` is a
CRLF. A call to `reserve_exact` is also added to the branch used when
continuing line comments.
#12177 changed `insert_newline`'s behavior to trim any trailing
whitespace on a line which came before a cursor. `insert_newline` would
previously never delete text. Even the whitespace stripping behavior in
#4854 worked by inserting text - a line ending at the beginning of the
line. `global_offs`, a variable that tracks the number of characters
inserted between iterations over the existing selection ranges, was not
updated to also account for text deleted by the trimming behavior,
causing cursors to be offset by the amount of trailing space deleted
and causing panics in some cases.
To fix this we need to subtract the number of trimmed whitespace
characters from `global_offs`. `global_offs` must become an `isize`
(was a `usize`) because it may become negative in cases where a lot of
trailing whitespace is trimmed. Integration tests have been added for
each of these cases.
Fixes#12461Fixes#12495Fixes#12539
For example `:cd README.md` would say "Not a directory" but would not
print the directory name. Now the error message includes some context
about the operation and requested directory.
Reproduction:
* `hx`
* Open any file in a split (`<space>f` and choose anything with `<C-v>`)
* Close the split with `<C-w>q`
* Open up the buffer picker and look the file you opened previously
Previously the preview was empty in this case because the Document's
`selections` hashmap was empty and we returned early, giving `None`
instead of a FileLocation. Instead when the Document is not currently
open in any view we can show the document but with no range highlighted.
Instantiating EditorView is a lot of machinery which is unnecessary:
the default keymap is exposed through the `default` function in the
keymap module.
Previously we replaced line-endings in pasted text to the document
line-ending for some values in paste commands. We missed the `repeat`
values in paste though and didn't do any replacement in the replace
command.
Along with this change I've refactored the replace command to avoid
intermediary collections. We previously eagerly collected the values
from the input register as a `Vec<String>` but we can avoid both of
those conversions and only allocate for the conversion to a `Tendril`.
We can also switch from `str::repeat` to a manual implementation to
avoid the intermediary conversion to a String - this avoids an extra
allocation in the common case (i.e. no count).
Fixes#12329
* feat: improve information on the amount of files loaded
* refactor: naming consitency Doc and not Buf
* fix: correct name of method
* chore: appease clippy
* feat: more human error information when Helix cannot start
* refatcor: use if guard on match arm
`set_string_truncated` renders the entire string while ignoring
newlines, so if the diagnostic's message contains multiple lines it
produces messages like 'first linesecond line'.
To avoid these run-ons, this commit renders each line separately,
inserting double spaces for disambiguation.
When pressing tab at the edge of the screen, Helix panics in debug mode
subtracting position.col - self.offset.col.
To correctly account for graphemes that are partially visible,
column_in_bounds takes a width and returns whether the whole range is
in bounds.
Co-authored-by: Rose Hogenson <rosehogenson@posteo.net>
* Handle window/showMessage and display it bellow status line
* Enable `editor.lsp.display_messages` by default
---------
Co-authored-by: Michael Davis <mcarsondavis@gmail.com>
The only caller of `from_path_buf` was removed in the parent commit
allowing us to drop owned variant of path's `Cow`. With this change we
never need to allocate in the picker preview callback.
The lsp location type has the lsp's URI type and a range. We can replace
that with a custom type private to the lsp commands module that uses the
core URI type instead.
We can't entirely replace the type with a new Location type in core.
That type might look like:
pub struct Location {
uri: crate::Uri,
range: crate::Range,
}
But we can't convert every `lsp::Location` to this type because for
definitions, references and diagnostics language servers send documents
which we haven't opened yet, so we don't have the information to convert
an `lsp::Range` (line+col) to a `helix_core::Range` (char indexing).
This cleans up the picker definitions in this file so that they can all
use helpers like `jump_to_location` and `location_to_file_location` for
the picker preview. It also removes the only use of the deprecated
`PathOrId::from_path_buf` function, allowing us to drop the owned
variant of that type in the child commit.
This is a temporary limitation because of the way that command sequences
are executed. Each command is currently executed back-to-back
synchronously, but macros are by design queued up for the compositor.
So macros mixed into a command sequence will behave undesirably: they
will be executed after the rest of the static and/or typable commands
in the sequence.
This is pending a larger refactor of how we handle commands.
<https://redirect.github.com/helix-editor/helix/issues/5555> has
further details and <https://redirect.github.com/helix-editor/helix/issues/4508>
discusses a similar problem faced by the command palette.
* refactor(commands): output `stderr` in `:sh` popup
* refactor(commands): switch to `from_utf8_lossy`
This way something is always displayed.
* refactor: no longer log stderr output on failure
* feat: pertty header
* 更新 themes.md
Co-authored-by: Michael Davis <mcarsondavis@gmail.com>
---------
Co-authored-by: Michael Davis <mcarsondavis@gmail.com>
* don't use backup files with hardlinks
* check if the inodes remain the same in the test
* move funcs to faccess and use AsRawHandle
* use a copy as a backup for hardlinks
* delete backup after copy
In the case of network file systems, if the server time is ahead
of the local system time, then helix could annoy with messages
that the file has already been modified by another application.
* replicate t-monaghan's changes
* remove View.offset in favour of Document.view_data.view_position
* improve access patterns for Document.view_data
* better borrow checker wrangling with doc_mut!()
* reintroduce ensure_cursor_in_view in handle_config_events
since we sorted out the borrow checker issues using partial borrows,
there's nothing stopping us from going back to the simpler implementation
* introduce helper functions on Document .view_offset, set_view_offset
* fix rebase breakage
Pastes are probably the last edit one means to make before the query
should run so it doesn't need to be debounced.
This makes global search much snappier for example when accepting the
history suggestion from the '/' register or pasting a pattern from the
clipboard or a register.
Two changes from the parent commit:
* Save only the `Picker::primary_query` - so you don't save other parts
of the query, for example `%path foo.rs` while in `global_search`.
* Move the saving out of the `if let Some(option) = self.selection()`
block. So when you hit enter you save to history whether you have a
selection or not. If you want to close the picker without saving to
the register you can use C-c or Esc instead.
The Prompt is set up to push the current line to history when hitting
Enter but the Picker doesn't pass the Enter event down to the Prompt
(for good reason: we don't want the Prompt's behavior of changing
completions when we hit a path separator). We should save the Prompt's
line to its configured history register when hitting Enter when there
is a selection in the Picker.
This currently only applies to `global_search`'s Picker since it's the
only Picker to use `Picker::with_history_register`.
This commit brings the text decoration API inline with the
LineAnnotation API (so they are consistent) resulting in a single
streamlined API instead of multiple ADHOK callbacks.
We can track the ranges in the input text that correspond to each column
and use this information during rendering to apply a new theme key that
makes the "active column" stand out. This makes it easier to tell at
a glance which column you're entering.
`Picker::new` loops through the input options to inject each of them, so
there's no need to collect into an intermediary Vec. This removes some
unnecessary collections. Also, pickers that start with no initial
options can now pass an empty slice instead of an empty Vec.
Co-authored-by: Luis Useche <useche@gmail.com>
The `FileLocation` and `PathOrId` types can borrow paths rather than
requiring them to be owned. This takes a refactor of the preview
functions and preview internals within `Picker`. With this change we
avoid an unnecessary `PathBuf` clone per render for any picker with a
file preview function (i.e. most pickers).
This refactor is not fully complete. The `PathOrId` is _sometimes_ an
owned `PathBuf`. This is for pragmatic reasons rather than technical
ones. We need a further refactor to introduce more core types like
`Location` in order to eliminate the Cow and only use `&Path`s within
`PathOrId`. This is left for future work as it will be a larger refactor
almost entirely fitting into the LSP commands module and helix-core -
i.e. mostly unrelated to refactoring the `Picker` code itself.
Co-authored-by: Pascal Kuthe <pascalkuthe@pm.me>
Michael Davis
Pascal Kuthe
renshyle
dependabot[bot]
Zi How Poh
kyfanc
Kristoffer Plagborg Bak Sørensen
magackame
Darshan Kumawat
Nikita Revenco
Denys Rybalka
Yomain
TornaxO7
Rolo
Jose Alvarez
janos-r
rhogenson
Jean Abou Samra
Nikita Revenco
Eduardo Rittner Coelho
Christian Schneider
Integral
Aaalibaba
Takumi Matsuura
Ian Hobson
Tomasz Zurkowski
Poliorcetics
Milan Vaško
Philipp Mildenberger
yehor
Alfie Richards
chtenb
Ayoub Benali
Jefta
Blaž Hrastnik
André Carneiro
麦芽糖
Skyler Hawthorne
Kirawi
Nikolay Minaev
Ingrid
Trevor Gross
jyn