It is easier to explain this fix with an example:
* tsserver and LSPs ask for error information when you want to fix
error. tsserver `ts@getCodeFixes` command needs tsserver error code.
* now let's imagine that user has eslint and tsserver in use. Sometimes
both can report same error in different way.
* Now there is no guarantee which error will come first and if eslint
error comes first then tsserver will not return code fixes as we are
passing wrong error code to it.
This fix will return proper error code based on linter.
Without this, we have one `pyre` process running across different pyre
projects. With this change, files in different projects can be linted
with pyre at the same time.
Co-authored-by: Oliver Ruben Albertini <ora@fb.com>
* Fix truncated echo
In typescript, when putting the cursor on a `>` character of an arrow
function, the displayString body comes back as an empty string, and
means the split operation has 0 items, causing a failure when attempting
to call TruncatedEcho.
Even if there's a better fix, I'd assume this is a good safety since we
are injesting external data.
* Convert to use `empty()`
This User autocommand is trigged immediately after an LSP process is
successfully initialized. This provides a way to perform any additional
initialization work, such as setting up buffer-level mappings.
For some reason CI tests started failing with these errors:
> ale_linters/eruby/erb.vim:1:1: Use scriptencoding when multibyte char exists (see :help :scriptencoding)
> ale_linters/mail/languagetool.vim:1:1: Use scriptencoding when multibyte char exists (see :help :scriptencoding)
Not sure at which point or what changed for this to happen but this MR
fixes it by removing the multibyte chars present on the problem files.
Co-authored-by: Horacio Sanson <horacio@allm.inc>
* feat(deno): move init options to handlers
* feat(deno): add deno lsp support for js files
* feat(deno): use default map option
* feat(docs): add deno import map option
* feat(deno): add tests for importMap option
* fix(deno): use full path in importMap
* feat(deno): remove deno as linter for js, separate PR
* fix(deno): test for executable
* fix(deno-test): include filename to simplify function
* Add poetry support to python linters and black fixer.
* Update python.vim to detect poetry project.
* Update ale.vim, add an option for poetry `g:ale_python_auto_poetry`.
* Update ale-python.txt, add poetry support.
* Add and update poetry related tests.
Co-authored-by: unc0 <unc0@users.noreply.github.com>
* implement vim popups for preview
Details on implementation
-------------------------
- we make use of the |popupwin| api
- we split implementations (Nvim* vs. Vim* prefix) and call the right
one based on has('nvim')
- we follow a similar structure in each function, using the relevant API
- popup_list, win_execute, popup_settext in VimShow
- popup_create in VimCreate
- popup_close in VimClose
Some differences
----------------
- we DON'T have VimPrepareWindowContent because we use arguments to
popup_create for borders, padding, etc., and it also takes care of
buffer creation.
- we follow the protocol of setting and using w:preview for information,
but we only need the ID
- InsertEnter is the only autocommand required, because of
popup_create's moved argument. Any cursor movement with 'any' will
close the popup. This in turns means VimClose is only called from
InsertMode, so no mode-restoration necessary
- we don't tweak too much in the buffer because vim's popup buffers
already have most relevant settings and aren't editable without
calling popup functions.
- I enabled scrollbars, close buttons, dragging, and resizing
- vim popups get as big as they need to by default, so no worrying about
truncating/hiding/size
Note: we might want to consider changing w:preview to w:ale_preview to
avoid clashes if someone else tries to use the same variable
* floating window: document that vim supports it
* lint: fix indent/cont. lines
* Fix 3801 - Add ALEDummySign some width.
Due to changes in NeoVim 0.5 the g:ale_sign_column_always configuration
stopped working.
This PR sets the ALEDummySign to a blank space so when g:
ale_sign_column_always is set we have a sign with 1 width allowing the
configuration to work as before.
https://github.com/neovim/neovim/issues/13635
* Fix visual artifact on dummy sign
* Fix visual artifact on dummy sign (attempt 2)
Co-authored-by: Horacio Sanson <horacio@allm.inc>
* Correct typo in a config file filename.
.tool_versions should be .tool-versions
* Correct typo in config file names.
.tool_versions should be .tool-versions
* racket: support racket-langserver lsp
* racket-langserver: find highest dir with init.rkt
* autoload/ale/racket: re-indent to 4 spaces
* racket: lint: sort supported tools
* racket: lint: function!
This is _not_ needed anymore, but the lint wants it. See :help E127
* racket-langserver: do not use new dict format
* racket: lint: use snake_case
* add tests for racket-langserver
* racket-langserver tests: correct result values
* Also check for asdf-vm's .tool_versions file
A minimal python project may only be specifying a python version using a version management tool like asdf-vm, without providing other common python project configuration files. asdf-vm creates a single .tool_versions file in the managed directory. By checking for .tool_versions in addition to other common python config files we ensure that python linters (whose behaviour typically depends on a particular python version) will run with the same version of python used by the project. This will also be the same python version used by vim itself when it is run from inside the project's directories.
* add .tool_versions to ale-python-root documentation
This reflects the corresponding change to autoload/ale/python.vim
Both '-include' and '-imacros' take a file as an argument that will then
be searched in the include path like a regular '#include "..."'
statement in a source file. As such, they should not have their path
converted to an absolute path.
Signed-off-by: Michael Jeanson <mjeanson@efficios.com>
The ocaml filetype is currently used for several, different file
formats. This causes problems as not all tools support all formats.
New filetypes are introduced to support this separation, this needs some
changes in ale that are fortunately backwards-compatible.
These change add ocamlinterface file support for ocp-indent, merlin,
ocamlformat and ocaml-lsp. For ocaml-lsp I took the liberty to
add all recognised language ids, even if they are not supported.
ols has not been changed as the project has been abandoned since 2019.
cmake-format added support for reading from/outputting to stdin/out as
of v0.3.6, released 2018-04-10 (commit 2e2aff2) [0].
Reading from stdin is preferable over reading from a temporary file
because when given a concrete file cmake-format will look for its config
file (.cmake-format.py or similar) in the parent directories of the
provided file. If the temporary file is off in a tmpdir somewhere (e.g.,
/tmp on *nix), cmake-format will almost certainly not come across the
user's intended format configuration file, making it appear that
cmake-format is ignoring the config file.
If cmake-format reads from stdin, though, it'll look for its config file
in its current working directory and its parent directories, in a
similar manner to clang-format. This has a much higher chance of running
across the intended config file.
[0]: https://github.com/cheshirekow/cmake_format/releases/tag/v0.3.6
Co-authored-by: Alex Wang <ts826848@gmail.com>
* fix: added support for local solhint executable
* feat: added support for matching parse errors
* test: added test for solhint command callback and handler
* chore: removed command callback test
* refactor: made solhint handler structure closer to eslint
Buildifier offers a -path command line option:
> Buildifier's reformatting depends in part on the path to the file
> relative to the workspace directory. Normally buildifier deduces
> that path from the file names given, but the path can be given
> explicitly with the -path argument. This is especially useful when
> reformatting standard input, or in scripts that reformat a temporary
> copy of a file.
This lets us drop our `-type`-guessing logic.
For the moment, we rely on the buffer filename being "relative to the
workspace directory", which it generally is, but in a future change, we
should introduce logic that will locate the WORKSPACE file and adjusts
the filename relative to that directory. This could be complicated based
on the working directory in which `buildifier` is executed, so a little
more research is necessary to implement that logic correctly.
