Response bodies are not read to EOF in certain cases, potentially causing TCP issues. about twirp HOT 2 CLOSED

This makes sense, thanks!

It seems this issue only manifests if the underlying transport is trying to re-use persistent TPC connections, and for some reason Twirp was not able to read the response body all the way to EOF (calls to io.ReadAll (doProtobufRequest) or json.NewDecoder(resp.Body).Decode (doJSONRequest)).

Looking at the client example code, this could happen if the context is canceled.

Instead of always trying to do io.ReadAll again, we could only try to do it on ctx.Err(). This will be safer to implement, and will not cause a second read on the response body for regular successful requests.

Current implementation:

	t.P(`  if err = ctx.Err(); err != nil {`)
	t.P(`    return ctx, wrapInternal(err, "aborted because context was done")`)
	t.P(`  }`)

Proposed solution:

	t.P(`  if err = ctx.Err(); err != nil {`)
	t.P(`    if resp.Body != nil {`)
	t.P(`      _, _ = io.ReadAll(resp.Body) // make sure the response body is fully read`)
	t.P(`    }`)
	t.P(`    return ctx, wrapInternal(err, "aborted because context was done")`)
	t.P(`  }`)

from twirp.

This issue is stale because it has been open 60 days with no activity. Remove stale label / comment or this will be closed in 5 days

from twirp.