| Commit message (Collapse) | Author | Age | Files | Lines |
|---|
| ... | |
| |
|
| |
fixes issue [#2959](https://github.com/neovim/nvim-lspconfig/issues/2959#issue-2068768094)
|
| |
|
|
|
|
|
| |
* feat(tailwindcss): add `templ` to filetypes
* feat(html): add `templ` to filetypes
* feat(htmx): add `templ` to filetypes
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
I'm not that well versed in Lua, but the previous check
if fname:sub(1, #item) == item
didn't really work as expected and I was still experiencing high CPU
usage when using the `go to definition` functionality. After adding some
print statements for debugging, I noticed that only a few strings were
removed from the base dir and thus the comparison was failing. Perhaps
it might work differently in other operating systems, but my Linux
machine it did not work correctly.
I replaced this check with `util.path.is_descendant` which now works
correctly, as well as added another path to check when git repositories
are used as cargo dependencies.
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* fix(ltex): fix rnoweb support
Unfortunately language code `sweave` will not be understood by LTeX and
`rsweave` must be used, as can be seen in the source code here:
https://github.com/valentjn/ltex-ls/blob/16.0.0/src/main/kotlin/org/bsplines/ltexls/parsing/CodeFragmentizer.kt#L61
and here:
https://github.com/valentjn/ltex-ls/blob/16.0.0/src/main/kotlin/org/bsplines/ltexls/settings/Settings.kt#L185
LTeX will fall back to plaintext, but this will emit a warning and
provide slightly worse suggestions.
* fix(ltex): disable unnecessary xhtml mapping
This mapping does not do anything, the `get_language_id` function will
just return `xhtml` by default.
* feat(ltex): add plaintext mapping
LTeX will only understand `plaintext`:
https://github.com/valentjn/ltex-ls/blob/16.0.0/src/main/kotlin/org/bsplines/ltexls/parsing/CodeFragmentizer.kt#L70
though it will fall back to plain text when provided `text` as mapping,
but will output a warning when doing so.
Note that we do not enable the `text` filetype by default, as LTeX
itself does not do so. However, this filetype will not be useable
without this mapping anyway, so doing this makes using enabling this
manually slightly easier.
* fix(ltex): set ltex.enabled
LTeX is launched whenever the filetype is on the preset list, but will
then verify the language ID against its list and will refuse to run if
the language ID is not on the list. The default can be seen in
https://github.com/valentjn/ltex-ls/blob/16.0.0/src/main/kotlin/org/bsplines/ltexls/settings/Settings.kt#L175
and somewhat matches the supported markup languages list.
The issue resolved by this commit comes from the fact that LTeX will
accept several aliases for each language, for instance it will handle
`gitcommit` and `git-commit` the same way. However, only `git-commit`
appears on the default enabled list. Since there is no `gitcommit ->
git-commit` mapping in `language_id_mapping`, LTeX will ignore
`gitcommit`. This could also be solved by adding this mapping, however
setting a default value of `ltex.enabled` has smaller change of breaking
existing configuration. Same situation happens for `xhtml`.
* feat(ltex): enable in context, html and xhtml filetypes
ConTeXt was requested in #2663. LTeX is documented to enable it by
default, but this is not consistent with the default in the code:
https://github.com/valentjn/ltex-ls/blob/16.0.0/src/main/kotlin/org/bsplines/ltexls/settings/Settings.kt#L175
Regardless of this we will enable it by ourselves anyway.
We also enable HTML and XHTML as there already has been an attempt to
enable them previously, as there has been a mapping for `xhtml` in the
`language_id_mapping` array. HTML support is enabled in LTeX by default
anyway.
|
| | |
|
| |
|
|
| |
This is required for the upcoming release of helm-ls
see https://github.com/mrjosh/helm-ls/pull/58
|
| |
|
|
| |
This is the preferred filename for defining options for a few versions.
meson_options.txt is legacy, but will continue to stick around.
|
| |
|
|
|
|
| |
fix #2975
followup to: 995603cfaecffe10e0f4bf57f92d187f86c2de39
|
| |
|
|
|
|
|
|
|
|
| |
Problem:
The `BufReadPost` handler may autostart LSP, even after `:LspStop`.
Solution:
Remove the `BufReadPost` handler if there are no clients (assuming that
`:LspStop` stopped them).
Note: The `BufReadPost` handler will be recreated if user runs `:LspStart`.
|
| |
|
| |
Per https://github.com/orgs/mdx-js/discussions/2427 and my own testing, mdx_analyzer needs a similar treatment as volar and astro to pass the typescript library into the lsp. Otherwise you get the error mentioned in that discussion.
|
| |
|
|
|
|
|
| |
Closes #2980
This PR adjusts the configuration for the biome LSP to only attach if there is an existing biome.json found.
This helps prevent conflicts in ESLint projects since biome would attach and start providing diagnostics and would format files which could conflict with the ESLint/Prettier rules that are defined.
|
| | |
|
| |
|
|
| |
as of vale 3.0.0, vale can be used with a system default `.vale.ini`,
which means vale_ls can also be used in single file mode
|
| |
|
| |
Co-authored-by: Markus Tauchnitz <dev@planet-mt.de>
|
| |
|
|
|
| |
* chore: update dcm config
* chore: add more docs around dcmls
|
| | |
|
| |
|
| |
https://dcm.dev/
|
| | |
|
| |
|
|
|
|
|
| |
* feat: add lwc-language-server support
* fix(lwc-language-server): added missing , to end of cmd line
* fix(lwc-language-server): remove trailing whitespace
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| |
* fix(ols): remove single_file_support
It doesn't seem like ols really has support for single_file_support. The
ols application crashes with the following message:
"ols/src/server/requests.odin(1494:28) Index 0 is out of range 0..<0\n"
This happens on workspace/didChangeConfiguration when settings are
provided in the users' ols setup.
To work around this, essentially turning off single_file_support
& adding *.odin as a way to detect a root_dir. This isn't very common
but at least one other language server uses this method, qml_ls.
However, without this, the ols doesn't really work out of the box for
projects that lack a .git directory or ols.json file.
* fix(ols): remove single_file_support that defaults to false
|
| | |
|
| | |
|
| |
|
|
|
| |
* feat(snyk-ls): add support for snyk-ls
* fix(snyk-ls): remove filetypes
|
| | |
|
| | |
|
| | |
|
| |
|
|
|
| |
* feat(denols): follow the latest default value of the vscode extension
* fix(denols): nvim_buf_set_option is deprecated
|
| | |
|
| |
|
|
|
|
|
|
|
|
|
|
|
|
| |
File `build.zig` by convention sits at the root dir and thus can be used
to correctly identify the root directory.
This is particularly useful if `zls` has `enable_build_on_save` set to
`true`, since then the language server uses workspace path (root_dir) to
execute `zig build` command, output of which is then used to publish
diagnostic errors.
Without `build.zig` root directory can incorrectly be deduced some
layers too high at the root of the git repo. In such case
`enable_build_on_save` silently fails and does not report build error
diagnostics.
|
| |
|
|
|
|
|
| |
* feat(kotlin_language_server): add storagePath to initilaztion options
* docs(kotlin_language_server): add storage path
* refactor(psalm): reformat code
|
| |
|
|
| |
* fix(psalm): psalm now uses a --language-server option
|
| | |
|
| |
|
|
|
|
|
| |
* feat: added @sap/cds-lsp
* fix(cds-lsp): update root dir pattern
* fix(cds-lsp): cleanup double configuraton of root_dir copied from csslsp example
|
| | |
|
| |
|
|
|
| |
behavior (#2923)
Ref. #2716. The referenced PR changes got somehow lost when it came to the removal of the `on_new_config` function.
|
| | |
|
| | |
|
| |
|
|
|
| |
* feat: add hydra_lsp
* feat(hydra_lsp): single file support
|
| | |
|
| |
|
|
| |
language server (#2914)
|
| |
|
|
|
|
|
|
|
|
|
|
|
| |
* feat: add typos-lsp server configuration
* fixup!
* fixup!
* fixup!
---------
Co-authored-by: Raphael <glephunter@gmail.com>
|
| |
|
|
|
| |
If we're in a subdirectory of a git repo, and that subdirectory contains
a mix.exs file, it makes sense to assume the subdirectory is the project
root, rather than the git repo root.
|
| |\ |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
| | | |
|
