aiohttp.HttpClient about aiohttp

[fafhrd91]

i have two use case at the moment for this class:

1. you can configure schema, host, port once and then use only paths to access your service. for example

client = aiohttp.HttpClient(hosts=['...'])
yield from client.request('/path1')
yield from client.request('/path2')

1. if you have multiple servers or load balancers and you want to access all of them (example is riak)

i'd rather see hooks for custom failure detection than removing this class.

[asvetlov]

In private chat with @fafhrd91 we decided to leave HttpClient for now.
Closing the issue.

[ludovic-gasc]

Hi @asvetlov, I'm not agree with you.
For me, aiohttp handles HTTP (Client, Server, Websockets, Gunicorn) for AsyncIO Python world. From my point of view, not a good idea to split this project to several projects because:

1. The source code size is small, easy to find information inside, it isn't a spaghetti source code that need to split.
2. Few people is contributing on this project, easier to contribute if everything is in the same place.

Nevertheless, Client API could be improved to be easier to use, and also error handling.

[asvetlov]

I don't want to split aiohhtp to several projects.
From my perspective the HttpClient is not good citizen for Public API, that's all.
I'm ok for keeping it for a while.

Have you some ideas for improvement Client API and error handling?
Please make an issue if you have some proposal.

[ludovic-gasc]

I don't made enough code with Client API to have a pertinent opinion for now.
My main issue was about headers handling, I had write more code that with Requests because it wasn't a dict but a deque, but this problem is solving (#62), thanks to you ;-)

[lock]

This thread has been automatically locked since there has not been
any recent activity after it was closed. Please open a new issue for
related bugs.

If you feel like there's important points made in this discussion,
please include those exceprts into that new issue.