Reduce the number of API calls performed by DatasetFSProvider.readFile()

[benjamin-t-santos]

Is your feature request related to a problem? Please describe.

Extenders can use vscode.workspace.fs.readFile(dsUri) to fetch a dataset's contents. It will also create an entry for the dataset/member in the FileSystemProvider if it does not already exist.

Currently, readFile() makes two MVS API calls. The first call is in remoteLookupForResource() (it calls fetchDataset() which calls allMembers() or dataSet() depending on the provided URI). The second call is in fetchDatasetAtUri() (calls getContents()). I believe we only need to make one API call with getContents().

Describe the solution you'd like

Rework fetchDatasetAtUri() to only make one API call by doing the following:

1. Call getContents()
2. If the API call is successful, update the entry with the contents (create a new entry if one does not already exist) and return it.
3. If getContents() fails, the requested resource does not exist on remote and we returnnull.

That is a rough approach we could take. I have prototyped this in my own fork and there is a performance improvement, which is quite relevant when reading the contents of 100s or 1000s of members. I tested reading 600 members in parallel and with this optimization, it is twice as fast.

Describe alternatives you've considered

Using the DatasetFSProvider, I do not see a way to get the contents of a dataset that does not already have an entry without making two API calls.

Additional context

Here is a PR prototyping that I described above: https://github.com/zowe/zowe-explorer-vscode/pull/3279. I hope the team will consider this change and I am open to feedback.

[github-actions[bot]]

Thank you for raising this enhancement request. The community has 90 days to vote on it. If the enhancement receives at least 10 upvotes, it is added to our development backlog. If it receives fewer votes, the issue is closed.