* 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>
* Add vim82 and neovim04 to CI tests.
* Fix test_sign_column_hightlighting test.
In vim82 with verbose=1 the output of highlight command changes breaking
the ale#sign#SetUpDefaultColumnWithoutErrorsHighlight(). This commit
forces verbose=0 when the method starts and restores the previous value
before exiting.
* No return values in vim82 returns a numeric value instead of a empty string.
* Fix test_reek_handler test
The FuzzyJSONDecode() method catches E474 when it fails to parse the
input as JSON but Vim8.2 throws E491 instead. This commit modifies the
function to catch both E474 or E491.
* Fix perl6 handler test.
Perl6 handler catches json parse errors using the E474 error but in
Vim82 it changed to E491. This commit modifies the handler so both
errors are considered.
* Fix list opening tests.
In Vim 8.2 the call `range(1, bufnr('$'))` always returns quickfix
buffers no matter if they are closed or not. Using `ls` does not show
them but the above range will always include them.
This new behavior breaks the ale#list#IsQuickfixOpen() method that in
turn breaks many other things. This commit fixes this by using the
getqflist() and getloclist() methods instead.
* Fix test updates loclist test.
For some reason in Vim 8.2 the sign offset seems to not reset between
tests causing the sign_id to not match in the Assert. When the test is
run individually it passes but when run as part of the whole suite the
sign_id is off by one.
Forcing the offset in the test setup seems to fix the issue.
* Fix omnifunc completion test.
For unknown reasons the SetCompletionResponse tests fail in Neovim 0.2
and 0.4. Unfortunatelly the only solution I found is to disable them
for neovim.
* Fix linter warnings
* Fix smoker test.
Add vim 8.2 to the list of versions that need some retires due to
randomly failing tests.
* Add docker image build job.
Trying some clever trick to build the docker image if not available
locally or in Docker hub. It uses the Dockerfile md5 checksum as tag so
only when changes on that file occur will the image be downloaded or
build.
* Add labels to Docker image
* Remove tests for middle versions 8.1 and 0.3.5
* Use same vader commit as appveyor
* Implement image push to Docker Hub
Co-authored-by: Horacio Sanson <horacio@allm.inc>
* Use sign-group only on supported vim versions.
The sign-group feature is only available in nvim 0.4.0 and vim 8.1.614.
* Add priority to ALE signs.
This allows users to set a priority to ALE signs to take precedence over
other plugin signs.
I've noticed that signs weren't unplaced and, learning that this was an issue depending on locale and :sign place, I've fixed the regular expression used to match those messages in my locale (it_IT.UTF-8).
* First pass at optimizing ale to autoload
First off, the structure/function names should be revised a bit,
but I will wait for @w0rp's input before unifying the naming style.
Second off, the docs probably need some more work, I just did some
simple find-and-replace work.
With that said, this pull brings major performance gains for ale. On my
slowest system, fully loading ale and all its code takes around 150ms.
I have moved all of ale's autoload-able code to autoload/, and in
addition, implemented lazy-loading of linters. This brings load time on
that same system down to 5ms.
The only downside of lazy loading is that `g:ale_linters` cannot be
changed at runtime; however, it also speeds up performance at runtime by
simplfying the logic greatly.
Please let me know what you think!
Closes#59
* Address Travis/Vint errors
For some reason, ale isn't running vint for me...
* Incorporate feedback, make fixes
Lazy-loading logic is much improved.
* Add header comments; remove incorrect workaround
* Remove unneeded plugin guards
* Fix lazy-loading linter logic
Set the wrong variable....
* Fix capitialization