Say we would refactor the commandAction function to return a Promise.
...we could then use async/await for all functions and directly await them from commandAction.
...we could use the cb() function to resolve that promise.
...we could move try catch logic down to the base command, so we don't need to call this.handleRejectedODataJsonPromise(rawRes, logger, cb)) in every command file.
For example:
public commandAction(logger: Logger, args: any, cb: () => void): void {
const requestOptions: any = { /* ... */ };
request
.get<{ value: any[] }>(requestOptions)
.then((res: { value: any[] }): void => {
logger.log(res.value);
cb();
}, (rawRes: any): void => this.handleRejectedODataJsonPromise(rawRes, logger, cb));
}Could be refactored to
public async commandAction(logger: Logger, args: any, cb: () => void): Promise<void> {
const requestOptions: any = { /* ... */ };
const res = await request.get<{ value: any[] }>(requestOptions);
logger.log(res.value);
cb(); //Not even strictly necessary, but for backwards compatibility it could be used.
}Refactoring can be spread over time, the promise chain that is used in commandAction can just be returned instead of refactoring everything to async/await right away. This would make an initial refactoring rather small.
Say we would refactor the
commandActionfunction to return a Promise....we could then use async/await for all functions and directly await them from
commandAction....we could use the
cb()function to resolve that promise....we could move
try catchlogic down to the base command, so we don't need to callthis.handleRejectedODataJsonPromise(rawRes, logger, cb))in every command file.For example:
Could be refactored to
Refactoring can be spread over time, the promise chain that is used in
commandActioncan just be returned instead of refactoring everything to async/await right away. This would make an initial refactoring rather small.