Auto Detect SqlFlulf executable from Python Virtual Environment

Background

Hard coding executable paths into settings.json makes the config completely unportable. People will have different operating systems, and even different folder structures depending how things are installed.

Defining a path variable would also be cumbersome.

It should be noted that different repos might even use different versions, further complicating this practice.

This all came from Issue 133.

Suggestion

It would be nice if the config could specify a dependencyManager and directory. The extension could load this.

settings.json

"sqlfluff.dependencyManager": "poetry",
"sqlfluff.projectRoot": "${workspace}/path"

Then the linter provider could spawn a child process and get exec from there.

With poetry, you can dynamically fetch the path from this poetry run which sqlfluff.

const { exec } = require('child_process');
return new Promise((resolve, reject) => {
  exec('poetry run which sqlfluff', (error, stdout, stderr) => {
     if (error || stderr) {
        reject(error || stderr);
        return;
     }

    resolve(stdout)
  });
});

In addition, this extension could talk to the Microsoft Python extension and get the virtual environment.

It is defiantly more complex. However, it allows for easy changes to the environment.

https://github.com/microsoft/vscode-python/blob/main/src/client/api.ts#L150C9-L150C21 https://stackoverflow.com/questions/50058517/how-to-communicate-between-vscode-extensions/50068416#50068416

sqlfluff / vscode-sqlfluff

Auto Detect SqlFlulf executable from Python Virtual Environment #146

Background

Suggestion