fix(fsthttp): restore previous limit method return values

[cee-dub]

Follow up to #186, responding to comments.

[cee-dub]

It appears we need to update another integration test since underlyingReaderFrom() can detect the reader inside an io.NopCloser.

[joeshaw]

👍

[joeshaw]

It appears we need to update another integration test since underlyingReaderFrom() can detect the reader inside an io.NopCloser.

Can you say more? When I last looked it was not possible to see inside of io.NopCloser, which is why we have our own implementation of it in fsthttp which can.

[cee-dub]

Can you say more? When I last looked it was not possible to see inside of io.NopCloser, which is why we have our own implementation of it in fsthttp which can.

I may have misread the code in seeking an explanation for the repeated failure (flake?) of the io.NopCloser test that's currently failing in CI.

[cee-dub]

Oh, it is a flake! My GitHub app UI was stale and showed a failure from yesterday. It passed after the latest re-run.