Use conda run for conda environments for running python files and possibly installation

[karrtikr]

Seems like this bit was missed:

Feel free to also add it in the iteration planning board. Note we should not be using the OUTPUT_MARKER_SCRIPT for running/debugging files, so add an option in this API

vscode-python/src/client/pythonEnvironments/common/environmentManagers/conda.ts

Lines 408 to 420 in 3b114e3

	public async getRunPythonArgs(env: CondaEnvInfo): Promise<string[] | undefined> {
	const condaVersion = await this.getCondaVersion();
	if (condaVersion && lt(condaVersion, CONDA_RUN_VERSION)) {
	return undefined;
	}
	const args = [];
	if (env.name) {
	args.push('-n', env.name);
	} else {
	args.push('-p', env.prefix);
	}
	return [this.command, 'run', ...args, '--no-capture-output', 'python', OUTPUT_MARKER_SCRIPT];
	}

to skip that.

cc/ @karthiknadig

[karrtikr]

It turns out we cannot use conda run for installing modules. Reason is:

• We already activate the conda environment before we send the installation command.
• conda run then kind of double activates the environment, so the module isn't installed where it's expected.
• When we run, for eg. a linter using conda run, it cannot find the module.

Seems like another bug on conda. (conda/conda#11305)