httpx Concurrent streaming of request and response (Fix)

The Fix

Upgrade to version 0.13.0 or later.

Based on closed encode/httpx issue #877 · PR/commit linked

Jump to Verify Open PR/Commit

@@ -46,10 +46,8 @@ async def send(self, request: Request, timeout: TimeoutTypes = None) -> Response
 
         await self._send_request(request, timeout)
-
-        task, args = self._send_request_data, [request.stream(), timeout]
-        async with self.backend.background_manager(task, *args):

repro.py

from typing import Callable

import uvicorn
from starlette.requests import Request
from starlette.responses import PlainTextResponse

async def app(scope: dict, receive: Callable, send: Callable) -> None:
    request = Request(scope, receive=receive)
    size = 0
    async for chunk in request.stream():
        size += len(chunk)
        if size > 1000:
            response = PlainTextResponse("Too large", status_code=413)
            await response(scope, receive, send)
            break

uvicorn.run(app, log_level="trace")

verify

Re-run the minimal reproduction on your broken version, then apply the fix and re-run.

fix.md

Option A — Upgrade to fixed release\nUpgrade to version 0.13.0 or later.\nWhen NOT to use: This fix is not suitable for scenarios requiring concurrent send/read functionality.\n\n

Why This Fix Works in Production

Trigger: ... so I cannot receive anything until once the request is completely sent.
Mechanism: Simplifies the HTTP/1.1 dispatcher by removing the concurrent send/read functionality, which was deemed unnecessarily complex.
Why the fix works: Simplifies the HTTP/1.1 dispatcher by removing the concurrent send/read functionality, which was deemed unnecessarily complex. (first fixed release: 0.13.0).

Why This Breaks in Prod

Production symptom (often without a traceback): ... so I cannot receive anything until once the request is completely sent.

Proof / Evidence

GitHub issue: #877
Fix PR: https://github.com/encode/httpx/pull/569
First fixed release: 0.13.0
Reproduced locally: No (not executed)
Last verified: 2026-02-09
Confidence: 0.85
Did this fix it?: Yes (upstream fix exists)
Own content ratio: 0.48

Discussion

High-signal excerpts from the issue thread (symptoms, repros, edge-cases).

Jump to Sources Open on GitHub

“For now this is solved by speaking HTTP by hand. I'll have a closer look if we need more tests. Thanks!”

@Tronic · 2020-03-26 · source

“Nope, that's not supported”

@lovelydinosaur · 2020-03-26 · source

“Clients such as curl and browsers' fetch API actually do support it, as do many HTTP servers (I specifically need this in tests of Sanic…”

@Tronic · 2020-03-26 · source

“> I specifically need this in tests of Sanic server You might want to double check, since I think the ASGI dispatcher would handle this,…”

@lovelydinosaur · 2020-03-26 · source

Failure Signature (Search String)

... so I cannot receive anything until once the request is completely sent.
You might want to double check, since I think the ASGI dispatcher would handle this, even if the over-the-wire HTTP/1.1 dispatcher doesn't.

Copy-friendly signature

signature.txt

Failure Signature
-----------------
... so I cannot receive anything until once the request is completely sent.
You might want to double check, since I think the ASGI dispatcher would handle this, even if the over-the-wire HTTP/1.1 dispatcher doesn't.

Error Message

Signature-only (no traceback captured)

error.txt

Error Message
-------------
... so I cannot receive anything until once the request is completely sent.
You might want to double check, since I think the ASGI dispatcher would handle this, even if the over-the-wire HTTP/1.1 dispatcher doesn't.

Minimal Reproduction

repro.py

from typing import Callable

import uvicorn
from starlette.requests import Request
from starlette.responses import PlainTextResponse

async def app(scope: dict, receive: Callable, send: Callable) -> None:
    request = Request(scope, receive=receive)
    size = 0
    async for chunk in request.stream():
        size += len(chunk)
        if size > 1000:
            response = PlainTextResponse("Too large", status_code=413)
            await response(scope, receive, send)
            break

uvicorn.run(app, log_level="trace")

What Broke

Users cannot receive response data until the entire request body is sent, causing delays.

Fix Options (Details)

Option A — Upgrade to fixed release Safe default (recommended)

Upgrade to version 0.13.0 or later.

When NOT to use: This fix is not suitable for scenarios requiring concurrent send/read functionality.

Use when you can deploy the upstream fix. It is usually lower-risk than long-lived workarounds.

Fix reference: https://github.com/encode/httpx/pull/569

First fixed release: 0.13.0

Last verified: 2026-02-09. Validate in your environment.

When NOT to Use This Fix

This fix is not suitable for scenarios requiring concurrent send/read functionality.

Verify Fix

verify

Re-run the minimal reproduction on your broken version, then apply the fix and re-run.

Did This Fix Work in Your Case?

Quick signal helps us prioritize which fixes to verify and improve.

Prevention

Add a CI check that diffs key outputs after upgrades (OpenAPI schema snapshots, JSON payload shapes, CLI output).
Upgrade behind a canary and run integration tests against the canary before 100% rollout.

Version Compatibility Table

Version	Status
0.13.0	Fixed

Related Issues

No related fixes found.

Cluster: httpx:other All hubs All clusters

Related clusters: Configuration error Data consistency Timeout

Sources

We don’t republish the full GitHub discussion text. Use the links above for context.