Seeming to be a BAD error when using streaming on release 6.0.3

I haven't checked for prior releases, but since I haven't seen this before 6.0.3, I suspect this may be tied for the rewrite of monk -- but not sure.

The issue is that code like:

( tasks).find({}, {stream: true})
.each((taskData: TaskData) => {
if (taskData._id) {
mapTasks.set(taskData._id, taskData);
}
})
.catch((err: any) => {
console.error(processErrorMessage(err.toString()));
reject(err);
})
.then(() => {
console.log('tasks #:', mapTasks.size());
resolve(true);
});

doesn't get the full TaskData from the database. It does get all of the items, but ONLY the id's. I was using 'stream: true' because I can have many thousands of items and don't want all of them in memory and the same time.

BUT the code:

( tasks).find({})
.then((docs: TaskData[]) => {
docs.forEach((taskData: TaskData) => {
console.log('listData:', listData)
mapTasks.set(taskData._id, taskData);
});
console.log('docs:', docs.length);
})
.then(() => {
console.log('tasks #:', mapTasks.size());
resolve(true);
});

works properly!

This should be fairly easy to add to your tests, since the mapTasks and all the rest can be ignored.

Start at Monk's find({}, {stream: true}) entry point and compare its result handling with find({}); reproduce the release 6.0.3 behavior using the TaskData example. Add a regression test covering streamed documents with fields beyond _id, and verify the non-streaming comparison still passes.

Written by the indexing model from the issue text.