This might be a bug:

I want to suppress spell checking inside Markdown code words and I've got a feeling that spellright.ignoreRegExps will be the way to go, however, I can't figure out how to write that regexp. Could you provide an example please? Is it possible at all? I've tried:

    "spellright.ignoreRegExps": [
        "/`.*`/g"
    ]

which doesn't seem to work.

Thanks for an awesome extension BTW.

I want to try out this extension, but it current says

This extension is currently limited to 64-bit Windows platform. It is due to problems with distribution of native modules in current VSCode's distribution model. For details see (and up-vote maybe) discussion about VSCode's feature request #20266.

I briefly read the issue and understand that you cannot provide an easy way to install on macOS for now. But is there any manual way? Can I build node-spellchecker myself and install it somehow? Any instructions?

1. Type word with spell error
2. Word is marked as bad
3. Fix word spelling
4. Word is marked ok
5. Pres Ctrl-Z to undo last change
6. Word is marked ok
7. Do some more editing somewhere
8. Word is re-checked and marked as bad

I thing steps 6 and 7 should be eliminated.

When I go into Help -> Toggle Developer Tools, I see some errors.

shell.ts:423 Cannot read property '_documentData' of undefined: TypeError: Cannot read property '_documentData' of undefined
at C:\Users[Username].vscode\extensions\ban.spellright-1.1.4\src\spellright.js:170:35
at e.invoke (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:55038)
at e.fire (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:56829)
at c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:392875
at e.invoke (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:55038)
at e.fire (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:56829)
at t.$acceptDocumentsAndEditorsDelta (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:348628)
at t.e.handle (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:308751)
at s (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:188848)
at h (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:189503)
e.onUnexpectedError @ shell.ts:423
shell.ts:423 Cannot read property '_documentData' of undefined: TypeError: Cannot read property '_documentData' of undefined
at C:\Users[Username].vscode\extensions\ban.spellright-1.1.4\src\spellright.js:170:35
at e.invoke (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:55038)
at e.fire (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:56829)
at c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:392875
at e.invoke (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:55038)
at e.fire (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:56829)
at t.$acceptDocumentsAndEditorsDelta (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:348628)
at t.e.handle (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:308751)
at s (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:188848)
at h (c:\Program Files\Microsoft VS Code\resources\app\out\vs\workbench\node\extensionHostProcess.js:23:189503)
e.onUnexpectedError @ shell.ts:423

When document contains more than 250 spelling errors Spell Right produces list that exceeds the number of allowed diagnostic hints (which is 250). It leads to a situation when during editing of the document misspelled words are not highlighted because they diagnostic hints land at the end of the list, on positions beyond 250, which are not displayed by VSCode.

This is a great plugin but one thing that is bugging me is not having the ability to add a selected word to the dictionary with a command.

Each time I add word to user dictionary language indicated at status bar is reset to [none].
Underlines and suggestion tooltips inside editor, spelling problems in PROBLEMS tab are preserved, but yellow bulb is not shown anymore.

After selecting language again it works... till next word added to dict.

For example, .gitignore

Would you please consider adding support for disabling the extension if the file has the comment block:

/* spellright-disable */

<!-- spellright-disable -->

The correct suggestion is displayed but for the wrong part.

Example 1:

all parts wrong, word with English dictionary: MyyOvnCemalCiseWerd (should be MyOwnCamelCaseWord):

cursor position marked with _ char

Example 2:

one part correct, word with English dictionary: MyyOvnCamelCiseWerd (should be MyOwnCamelCaseWord):

There is no way to display suggestion for the first part of CamelCase word.

Selected suggestion corrects right word Cemal -> Camel etc.

When document is spelled in multiple languages (switched with spellcheck-language CODE) in sections which are not in default language (set using switch or extension settings) suggestions provided in Code Action (not in "problems" list) are in wrong language.

Trying to execute select dictionary with a trailing comma in settings.json leads to a popup error that says:

command 'spellright.selectDictionary' not found

Developer tools's console outputs this error:

Activating extension `ban.spellright` failed: Cannot read property 'spellright.ignoreWords' of null.

I only realized I had a trailing comma when adding some console.logs in the extension src files. The problem is the JSON.parse in line 1384 of spellright.js. It fails when parsing settings.json, returns null as the value and leads to the reading ignoreWords of null error above.

I see 2 solutions:

1. Add better error on getUserSettings(), so that the user is notified the settings.json file has bad syntax, and can realize why the extension is not loading.
2. Ignore the trailing comma. Other extensions seem to use vscode.workspace.getConfiguration('extension'), instead of reading the file manually and parsing it (internally VScode parser seems to ignore trailing commas). Maybe the code can be modified for this?

Which would be the better option? I could make a PR if it would help.

When put cursor in CamelCase word and then add it to user dictionary only one part can be added even if this word is brand or trademark and should be added as a whole.

The same problem with snake_case and digit2inside words.

When node-spellchecker module used by the extension is using Windows native Spell Checking API it reports a multitude of dictionaries of same language and different culture:

It would be convenient to have some way to filter the dictionaries (in positive or negative manner) only displaying main or selected dictionaries (e.g. fr_FR instead of all fr_ dictionaries etc.)

It would lead to cleaner view and quicker selection.

Previous I could set the spelling check into English or Dutch, unfortunately it is currently not possible to switch.

I hope this can be resolved.

Thank you in advance.

Is it possible to make a setting to produce warnings for incorrect spell instead of errors.

It's because Latex errors are often unseen between some spelling errors
Thanks

• VSC 1.11.2
• WIndows 10 Pro 15063.138
• Spell Right 1.0.12

With spell checking enabled on a .mustache template file that is HTML no spelling is checked. Need options to check spelling if it's in code or in strings or just text files.

James Yu suggested in one of comments to issue #225 that Spell Right spells math expressions like:

\forall(i,j)\notin\mathcal{E}

where i gets underlined as spelled incorrectly (suggesting capital).

They should not be spelled.

I've got a question. On my PC, I seem to have custom ignore words in two files:

• spellchecker.json
• settings.json (also in the ../Code/User directory)

When I do Add "customword" to global ignore list, it seems to be added to my user settings.json. How does the spellchecker.json file work? Is there a way to add my custom words there? I would prefer that as I synchronize my user settings.json between two computers and custom words create a lot of noise and possible merge conflicts there. Thanks.

Hello!

Firstly, thank you for your work.

In our projects we use readthedocs.org and as consequence reStructuredText (*.rst files). I tried to configure Spell Right using spellright.json as described below but it didn't give a result.

{
    "language": "en",
    "documentTypes": [
        "restructuredtext",
        "markdown",
        "latex",
        "plaintext"
    ],
    "groupDictionaries": true,
    "statusBarIndicator": true,
    "suggestionsInHints": true,
    "ignoreWords": [],
    "ignoreRegExps": [],
    "ignoreFiles": [
        "**/.gitignore",
        "**/.spellignore"
    ]
}

Originally reported in #24 (comment) but it's worth a separate issue, IMO. Link part of MD link syntax shouldn't be spell checked:

[Spell check this](DO NOT spell check this)

Repro in 1.0.25:

If you have a .tex file in the left editor group and a .txt file in the right editor group, the .tex file seems to be checked with the rules for .txt file.

hello, is there a way to config this plugin to perform only in comment, but not in code?

Hey Bartosz, our PM pointed me to an awesome spell checking extension on the marketplace and the amazing part is we already knew each other :) I played with your extension, it's pretty fast and the results are really reasonable (aka I really like it , even though I need to run node-gyp rebuild myself 😃, tada).

We talked about async interface in microsoft/vscode#20266 (comment) and at that time I think it's not a big deal as we can use setTimeout to make spellchecker's api async. But this doesn't mitigate the performance issue.

• In your current implementation, we call spellcheck.checkspelling directly, which is a synchronous function. It means this onDidTextDocumentChange handler is synchronous, when nodejs is running this function, no other events will be handled in the event loop.
• Then let's say we put it into setTimeout or whatever delaying function, this onDidTextDocumentChange handler is finished immediately, which is good. However when nodejs picks up the spellcheck.checkspelling, it still blocks all other events.

So this time let me describe it more precisely, if the spellchecking api is running in the main thread (the event loop), the main thread is blocked. For small documents, it's okay as spellchecking is pretty fast. But for medium to large files (eg, https://github.com/Microsoft/vscode-docs/blob/master/release-notes/v1_8.md ), it takes several seconds to finish the spellchecking, then none other extension is reactive.

A proper reproduce step:

• Install a Vim extension and Spell Right
• Open https://github.com/Microsoft/vscode-docs/blob/master/release-notes/v1_8.md
• Try to use Vim motions or commands, you can see the cursor doesn't move when pressing hjkl. After the spellchecking is done, then all of events get flushed to Vim extension.

To mitigate this issue and the node-gyp rebuild problem, I tried two ways

• Run spellchecking in a command line tool. Write some C++ glue code and talk with spellchecking api through standard IO. You can see my trail here https://github.com/rebornix/vscode-nsc and the C++ code should be improved in many ways https://github.com/rebornix/vscode-nsc/blob/master/src/native/main.cc . This way it won't break when Electron/Nodejs updates.
• Make spellchecker real async and multi-thread. I tweaked the native module, make it async and multithread by leveraging nan asyncworker https://github.com/rebornix/node-spellchecker/tree/asyncworker . Besides I put it inside VS Code.

You may want to go with the first option or similar solutions to solve the node-gyp problem and even better not block the extension host :)

Lastly I'm thinking about how to ignore ranges in a document, your approach is really inspiring.

Fantastic LaTeX Workshop extension shows linting hints from ChkTeX which suggest use \( and \) braces instead of $ and $ for inline math and \[ and \] instead of $$ and $$ for row math. It seems that apart $ and $ pair Spell Right tries to spell all the other.

1. It seems that the config doesn't work
2. It would be better to remove the default white foreground as it overrides the theme color

Currently SpellRight doesn't work on files with the extension *.asc, *.adoc, and *.asciidoc.

Asciidoc files are just text files. Asciidoc is a lightweight markup language similar to markdown but targeting longer documents like books.

http://asciidoctor.org/docs/asciidoc-recommended-practices/#document-extension

AsciiDoc doesn’t care which extension you use. GitHub supports the extensions .asciidoc, .adoc and .asc.
.adoc seems to be the most often used extension and the one we prefer. It does not conflict with other well known applications, is distinct, and not excessively long (like .asciidoc). It also reads as “a doc”. Our recommendation, therefore, is to use .adoc.

Thank you.

Example: MyyOvnCemalCiseWerd again 😄

1. Whole word is underlined as wrong spelling
2. One part is corrected (Ovn -> Own) using suggestion or manually
3. Whole word is not underlined (marked as OK?)
4. After changing the line containing this word...
5. ...underline is displayed again for wrong parts only

I think steps 3 and 4 should be eliminated.

I want to ignore two expressions,

• \verb|something| in LaTeX
• Chinese words, e.g. 中文词语

I have tried many regexs but can't get them to work.

Finally, I tried abc and /abc/g to ignore abc and found they don't work too...

Could you please have a check whether the regexs are applied to ignore words?

Hello!

Today I tried to exclude code blocks from the analysis using ignoreRegExps and expression "/\\s{4,}.*/g". Here is my configuration:

{
    "language": "en",
    "documentTypes": [
        "restructuredtext",
        "markdown",
        "latex",
        "plaintext"
    ],
    "groupDictionaries": true,
    "statusBarIndicator": true,
    "suggestionsInHints": true,
    "ignoreWords": [],
    "ignoreRegExps": [
        "/\\s{4,}.*/g"
    ],
    "ignoreFiles": [
        "**/.gitignore",
        "**/.spellignore"
    ]
}

After few iterations I haven't gotten success and I need your help again. :) I also tested my expression here https://regex101.com/r/GJXToz/1 and it works fine (I think).

Thanks.

I installed this extension on my Visual Studio Code and I couldn't figure it out how it works or if it is working properly on my computer.

The spell check didn't worked ever on new plain text files. I tried with LaTeX files (with syntax plugins) and it only works if I reselect the language that I want to correct (I already set it by default on settings.json).

Any help would be appreciated. I'm using Windows 10 Home Single Language with Creators Update.

Would be nice if custom words could be added to a system dictionary on Windows. It is located at C:\Users\Me\AppData\Roaming\Microsoft\Spelling\en-US\default.dic for me but also very likely accessible via some API as Caret, a Markdown editor I use, stores custom words there.

When clicking on the "Add ??? to global dictionary" the spell checker adds the word to both global (settings.json) and workspace (spellright.json) ignore lists.

Additionally, in the global options file (settings.json) the ignoreWords list is created with a wrong name "spellchecker.ignoreWords" when the accepted one is "spellright.ignoreWords".

Using Windows 7 64bit, Visual Studio Code latest stable version.

Extension is installed.

Tested on a text file (plain). Doesn't highlight any of the mistakes that exist.

Executing any of the "SpellRight" commands returns "command not found"

Other installed extensions: asciidoc; beautify; html preview; markdownlint; rewrap; write good linter.

I'm frustrated that I can't provide better feedback because I can't find any reason from my part for the failure of the extension to work.

Thank you.

Please compare:

with Hunspell output:

$ echo "żólty jerz żółty jest śliczny i biały" | hunspell 
Hunspell 1.4.0
& żólty 1 0: żółty
& jerz 14 6: jeż, jesz, rejz, jer, jerze, jerez, jera, jarz, jery, perz, jedz, jeru, Perz, jer z
*
*
*
*
*

$ echo $LANG
pl_PL.UTF-8

Why ż in jeż is unknown character? Encoding problem maybe?

The name for the Norwegian Bokmål dictionary displays as in both the status bar and the language selection menu. For reference, I'm running the latest version of VS Code (1.11.1) on Windows 10 x86_64.

Perhaps this issue is not specific to this dictionary, and might apply to any dictionary that has a name using characters outside of the ASCII range?

• VSC 1.11.2
• WIndows 10 Pro 15063.138
• Spell Right 1.0.12

Files listed in spellright.documentType do not auto start spell checking. These files should default to the set spellright.language when VSC starts / loads the file.

I use the REST Client extension and their .http files. I have enabled Spell Right for it but the comments are not spell-checked:

Does this have something to do with the way they define their language? Is this a problem or their end or an issue in Spell Right? Thanks.

With SpellRight configured to not check any document types by default, try the following:

1. Open a new instance of VS Code with no documents open.
2. Hit Ctrl + N to start a new text file.
3. Type some random/nonsense words.
4. Hit Ctrl + N to open a second editor/tab.
5. In the second file, type some random/nonsense words.
6. While still in the second file, enable SpellRight from the status bar and notice that the words get underlined in red.
7. Switch to the first file in the first tab. Notice that the words there get underlined in red too. So far so good.
8. Switch back to the second file.
9. Turn off SpellRight from the status bar (i.e. turn off for document type [plaintext])
10. Go back to the first file, where the words are still underlined in red. Notice that according to the status bar, SpellRight is supposedly disabled.

Could there be an option to ignore code blocks?

Ex:

```cs
int a = 0;
```

If I enable Spell Right for TypeScript and other code files, it seems to spell-check comments and strings. Strings are usually technical stuff I don't want to spell check, for example, ES6 imports like import { env } from '../Dotenv'. I'd like only comments spell-checked in the code files.

I wonder if if the documentTypes could be enhanced like this:

"spellright.documentTypes": {
    "plaintext": true,
    "markdown": true,
    "typescript": "comments-only",
    "javascript": "comments-only"
}

I don't have any custom settings and this is in the default settings:

"spellright.language": "en_US"

But when I open any Markdown document, it defaults to a Czech language (which is one of my two languages I write in; where does Spell Right take the languages from anyway?). I think this might be a bug.

Whenever I start VS Code, Spell Right is enabled for the text documents I currently have open, set to spell check in English.

This is despite me having this setting configured:

"spellright.documentTypes": [],

Even if I then disable Spell Right for the current document type, if I then quit and restart VS Code, Spell Right becomes enabled once again.

I believe this only started happening after the recent update to 1.1.10. Previously Spell Right was always disabled for all file types until I explicitly enabled it.

It did not work at all for C++ code. It did not even show up on the status bar. VSCode X64 version: 1.14.2. I'm using Windows 10 Enterprise Version 1703 OS Build 15063.483

This is the console output

Activating extension `ban.spellright` failed: The module 
'\\?\C:\Users\username\.vscode-insiders\extensions\ban.spellright-1.0.8\node_modules\spellchecker\build\Release\spellchecker.node'
was compiled against a different Node.js version using
NODE_MODULE_VERSION 50. This version of Node.js requires
NODE_MODULE_VERSION 53. Please try re-compiling or re-installing
the module (for instance, using `npm rebuild` or`npm install`)..

Windows 10 & vscode-insider build 2017-04-20

If I have an unknown word such as a technical word like "flyout" at the end of a sentence as shown:

This opens a flyout.

The spell checker includes the punctuation. Even if I add "flyout" to the dictionary, it does not work because the spell checker still considers "flyout." to be an incorrect spelling.

On Linux and Windows 7 Spell Right uses built in Hunspell spell checker library. To use it a pair of Dictionary (.dic) and Affixes (.aff) files have to be downloaded from here (please remember to download RAW files) and placed in Dictionaries subfolder of VSCode's user global configuration directory, located at:

• Windows: %APPDATA%\Code\Dictionaries\
• Linux: $HOME/.config/Code/Dictionaries/

Dictionaries subfolder does not exists there by default and has to be created manually.

On Linux it was sufficient for me to symlink system-wide installed dictionaries:

ln -s /usr/share/hunspell ~/.config/Code/Dictionaries

All dictionaries are detected and working.

Please verify if this is correct solution (this dir is readonly for users on Linux) and update SpellRight documentation.

Hi, just updated to 1.0.10 and found that in markdown for example, phrases with apostrophe do not get recognized correctly.

spellright actually suggests to replace it with doesn't and right after replacing it, you get the same warning for doesn again ;)

Thanks for the great extension btw!

If there are some spell problems in a file such as README.md, when launching debug, a warning will popup and I need to click debug anyway.
For some specific 'spell errors', I don't want to either fix them or add them to .ignore.
But it's not good to have one more debug anyway click since they are not crucial errors.