Fixes#1077. This was caused by the assumption that a block
cursor is represented as zero width internally and simply
rendered to be a single width selection, where as in reality
a block cursor is an actual single width selection in form and
function.
Behavioural changes:
1. Surround selection no longer works when cursor is _on_ a
surround character that has matching pairs (like `'`
or `"`). This was the intended behaviour from the start
but worked till now because of the cursor position
calculation mismatch.
* Jump to end char of surrounding pair from any cursor pos
* Separate bracket matching into exact and fuzzy search
* Add constants for bracket chars
* Abort early if char under cursor is not a bracket
* Simplify bracket char validation
* Refactor node search and unify find methods
* Remove bracket constants
* Add command to inc/dec number under cursor
With the cursor over a number in normal mode, Ctrl + A will increment the
number and Ctrl + X will decrement the number. It works with binary, octal,
decimal, and hexidecimal numbers. Here are some examples.
0b01110100
0o1734
-24234
0x1F245
If the number isn't over a number it will try to find a number after the
cursor on the same line.
* Move several functions to helix-core
* Change to work based on word under selection
* It no longer finds the next number if the cursor isn't already over
a number.
* It only matches numbers that are part of words with other characters
like "foo123bar".
* It now works with multiple selections.
* Add some unit tests
* Fix for clippy
* Simplify some things
* Keep previous selection after incrementing
* Use short word instead of long word
This change requires us to manually handle minus sign.
* Don't pad decimal numbers if no leading zeros
* Handle numbers with `_` separators
* Refactor and add tests
* Move most of the code into core
* Add tests for the incremented output
* Use correct range
* Formatting
* Rename increment functions
* Make docs more specific
* This is easier to read
* This is clearer
* Type can be inferred
* Add treesitter textobject queries
Only for Go, Python and Rust for now.
* Add tree-sitter textobjects
Only has functions and class objects as of now.
* Fix tests
* Add docs for tree-sitter textobjects
* Add guide for creating new textobject queries
* Add parameter textobject
Only parameter.inside is implemented now, parameter.around
will probably require custom predicates akin to nvim' `make-range`
since we want to select a trailing comma too (a comma will be
an anonymous node and matching against them doesn't work similar
to named nodes)
* Simplify TextObject cell init
* allow language.config (in languages.toml) to be passed in as a toml object
* Change config field for languages from json string to toml object
* remove indents on languages.toml config
* fix: remove patch version from serde_json import in helix-core
* Use same tree-sitter-zig as upstream/master
* feat: merge default languages.toml with user provided languages.toml
* refactor: use catch-all to override all other values for merge toml
* tests: add a test case for merging languages configs
* refactor: change test module name
* Fix around-word text-object selection.
* Text object around-word: select to the left if no whitespace on the right.
Also only select around when there's whitespace at all.
* Make select-word-around select all white space on a side.
* Update commented-out test case.
* Fix unused import warning from rebase.
* Implement `margin` calculation for uncommenting
* Move `margin` calculation to `find_line_comment`
* Fix comment bug with multiple selections on a line
* Fix `find_line_comment` test for new return type
* Generate a single vec of lines for comment toggle
`toggle_line_comments` collects the lines covered by all selections into
a `Vec`, skipping duplicates. `find_line_comment` now returns the lines
to operate on, instead of returning the lines to skip.
* Fix test for `find_line_comment`
* Reserve length of `to_change` instead of `lines`
The length of `lines` includes blank lines which will be skipped, and as
such do not need space for a change reserved for them. `to_change`
includes only the lines which will be changed.
* Use `token.chars().count()` for token char length
* Create `changes` with capacity instead of reserving
* Remove unnecessary clones in `test_find_line_comment`
* Add test case for 0 margin comments
* Add comments explaining `find_line_comment`
* Added option to provide a custom config file to the lsp.
* Simplified lsp loading routine with anyhow
* Moved config to language.toml
* Fixed test case
* Cargo fmt
* Revert now-useless changes
* Renamed custom_config to config
Co-authored-by: Cor <prive@corpeters.nl>
For example when the cursor is _on_ the `'` in `'word'`, the cursor
wouldn't move because the search for a matching pair started _from_ the
position of the cursor and simply found itself.
* Add textobjects for word
* Add textobjects for surround characters
* Apply clippy lints
* Remove ThisWordPrevBound in favor of PrevWordEnd
It's the same as PrevWordEnd except for taking the current char
into account, so use a "flag" to capture that usecase
* Add tests for PrevWordEnd movement
* Remove ThisWord* movements
They did not preserve anchor positions and were only used
for textobject boundary search anyway so replace them with
simple position finding functions
* Rewrite tests of word textobject
* Add tests for surround textobject
* Add textobject docs
* Refactor textobject word position functions
* Apply clippy lints on textobject
* Fix overflow error with textobjects
Surround operations previously ignored other pairs that are
enclosed within which should be skipped. For example if the
cursor is on the `,` in `{{a},{b}}`, doing `md{` previously
would delete the `{` on the left of `a` and `}` on the right
of `b` instead of the outermost braces. This commit corrects
this behavior.
* Fix expansion of `~`, dont use directory relative to cwd.
* Add `expand_tilde`
* Bring back `canonicalize_path`, use `expand_tilde` to `normalize`
* Make `:open ~` completion work
* Fix clippy
* Fold home dir into tilde in Document `realitve_path`
Registers are stored inside `Editor` and accessed without `RwLock`.
To work around ownership, I added a sister method to `Editor::current`:
`Editor::current_with_context`. I tried to modify `Editor::current`
directly but it's used at a lot of places so I reverted into this for
now at least.
- Move char functions into their own module under helix_core.
- Use matches!() macro where appropriate.
- Use a static lifetime on indent_unit() now that we can.
* Disable deleting from an empty buffer which can cause a crash.
* Improve on the fix for deleting from the end of the buffer.
* Clean up leftover log.
* Avoid theoretical underflow.
* Implement :before which accepts a time interval and moves the editor to
the closest history state to the commit of the current time minus that
interval. Current time is now by default, or the commit time if :before
has just been used.
* Add :earlier an :later commands that can move through
the edit history and retrieve changes hidded by undoing
and commiting new changes. The commands accept a number
of steps or a time period relative to the currrent change.
* Fix clippy lint error.
* Remove the dependency on parse_duration, add a custom parser instead.
* Fix clippy errors.
* Make helix_core::history a public module.
* Use the helper for getting the current document and view.
* Handled some PR comments.
* Fix the logic in :later n.
Co-authored-by: Ivan Tham <pickfire@riseup.net>
* Add an alias for :earlier.
Co-authored-by: Ivan Tham <pickfire@riseup.net>
* Add an alias for later.
Co-authored-by: Ivan Tham <pickfire@riseup.net>
* Run cargo fmt.
* Add some tests for earlier and later.
* Add more tests and restore the fix for later that diappeared somehow.
* Use ? instead of a match on an option.
Co-authored-by: Ivan Tham <pickfire@riseup.net>
* Rename to UndoKind.
* Remove the leftover match.
* Handle a bunch of review comments.
* More systemd.time compliant time units and additional description for the new commands.
* A more concise rewrite of the time span parser using ideas from PR discussion.
* Replace a match with map_err().
Co-authored-by: Ivan Tham <pickfire@riseup.net>
Co-authored-by: Jakub Bartodziej <jqb@google.com>
Co-authored-by: Ivan Tham <pickfire@riseup.net>
* Add convenience/clarity wrapper for Range initialization
* Test horizontal moves
* Add column jumping tests
* Add failing movement conditions for multi-word moves
* Refactor skip_over_next
* Add complex forward movement unit tests
* Add strict whitespace checks and edge case tests
* Restore formatting
* Remove unused function
* Add empty test case for deletion and fix nth_prev_word_boundary
* Add tests for backward motion
* Refactor word movement
* Address review comments and finish refactoring backwards move
* Finish unit test suite
* Fmt pass
* Fix lint erors
* Clean up diff restoring bad 'cargo fmt' actions
* Simplify movement closures (thanks Pickfire)
* Fmt pass
* Replace index-based movement with iterator based movement, ensuring that each move incurs a single call to the RopeSlice API
* Break down tuple function
* Extract common logic to all movement functions
* Split iterator helpers away into their own module
* WIP reducing clones
* Operate on spans
* WIP simplifying iterators
* Simplify motion helpers
* Fix iterator
* Fix all unit tests
* Refactor and simplify
* Simplify fold
User can select register to yank into with the " command.
A new state is added to `Editor` and `commands::Context` structs.
This state is managed by leveraging a new struct `RegisterSelection`.
