Glyph

StubTreq & RequestTraversalAgent end up scheduling timeouts with the global reactor

1

This is related to #161 and #250 .

Treq has a global HTTP client, and maybe that's a good thing for quick experiments, but it makes testing difficult and creates type confusion. Instead: 1. `treq.client.HTTPClient` should be promoted...

support files= in StubTreq

1

Currently if you specify a `files=` argument to `StubTreq`, it complains that this is not supported. The *reason* it's not supported is that `HTTPClient.request`, when given a `files=` argument, will...

We already have `Response.content()` which is what you should use. The main purpose that `treq.content` serves is to trick unsuspecting passers-by into accidentally attempting to call it on an `HTTPClient`...

treq should define an augmented response interface, not document private (???) concrete _Response class

1

The documentation currently documents `_Response`, which raises the question of what that underscore prefix is doing there exactly. Instead, we should define a `treq.itreq.IHelpfulResponse` (better name pending) which documents all...

StubTreq should get its mind right about what interface it's a replacement for

1

Apropos of #242: is `StubTreq` a `treq` or an `HTTPClient` ? It's confused about this. We should deprecate `StubTreq` itself and replace it with `StubClient`.

expose the Twisted-testing internals of `treq.testing` for testing things that are concurrent with HTTP requests

1

Twisted is great because you can do lots of stuff concurrently. `MemoryReactor` is great because you can test the completion of concurrent operations in memory. `StubTreq` is great because it...

☂ objects should document their full, public interface

1

This is an umbrella issue to track track public classes having clear documentation of the form "what can I, as an application programmer, invoke on and depend on from this...

Looks like #449 and #445 are back again. Anything I can do to help get these codesigned so that users aren't being trained to ignore system security features?

Hi Merlin! Thanks for putting this up on Github, it's an interesting document. Reading this document sparks lots of ideas, and in a McLuhanist way, seeing it in the context...