uvloop breaks aiohttp's capability of serving static files

[diogobaeder]

• uvloop version: 0.5.3
• python version: 3.5.2
• platform: Arch Linux / 4.7.1-1-ARCH / x86_64

uvloop breaks aiohttp's capability of serving static files. It makes the server send the response headers after the body.

To run this, have aiohttp and uvloop installed, then run the gist ( https://gist.github.com/diogobaeder/0558c43d785cd82a1ecff553bc583867 ):
$ python uvhttp.py

Then make a request to the app, making sure to get verbose output:
$ curl -v localhost:8765/jquery.js

These are the results I'm getting when serving jQuery:

1. Without uvloop (correct response): http://pastebin.com/e6ep3g22
2. With uvloop (bad response): http://pastebin.com/Kg37DXB8

Notice how with uvloop the response headers are sent after the body.

I'm not acquainted with the internals of aiohttp or uvloop, but it seems that the former relies on some parser from asyncio and uvloop seems to replace it by another one that behaves differently.

Any ideas of why this happens? At the moment this is preventing me to use uvloop, unfortunately...

Thanks!

[1st1]

What version of aiohttp are you using? AFAIK there was an issue with how aiohttp implemented sendfile, which was fixed not long ago.

cc @asvetlov

[diogobaeder]

Hi @1st1 ,

Unfortunately I'm already using the latest version, which is 0.22.5. Mind trying it as well, and see if it works fine for you?

Thanks for the quick reply :-)

[asvetlov]

There is aio-libs/aiohttp#1093 issue.
It's still not fixed but very reproducible.
I think aiohttp sends headers into transport's buffer but sendfile sends a content directly through the wire, that's why headers are sent too late.

[diogobaeder]

Hmmm... makes sense...

Should we close this issue as a duplicate? Want me to refer it in the other one?

[EDIT] Already done, sorry. You can close this one if you want.

[asvetlov]

@1st1 please close the issue as aio-libs/aiohttp#1093 is resolved.

[1st1]

@asvetlov When will you release aiohttp with this thing fixed?

[asvetlov]

@1st1 just after fixing bugs with cookiejar.
In a few days I hope.

[1st1]

Closing the issue since it has been resolved in aiohttp.