for #1166, #1151, maybe others. #1151 is the big motivation here: that we're currently generating and sending Bluesky a ton of reposts with blank subject cid/URL, which afaict is the main reason bridging is getting stuck for some accounts and the relay is constantly calling `getRepo`. I'm hoping this will reduce or eliminate that.
this isn't a big behavior change, but it is a kind of big change to routing logic. it's tested pretty well, but I still expect something to break. we'll see.
for #990
this assumes an implementation detail of fediverse servers: that they apply a white-space: pre-wrap style to <p>s inside content. awkward overreach, but useful since white-space: pre-wrap preserves more whitespace (eg leading spaces on a line, multiple spaces, etc) than just converting newlines to <br>s.
the ATProto DID spec says it shouldn't be there: https://atproto.com/specs/did#did-documents
> The serviceEndpoint field must contain an HTTPS URL of server. It should contain only the URI scheme (http or https), hostname, and optional port number, not any "userinfo", path prefix, or other components.
thanks for the nudge @mackuba!
This reverts commit 30debfc8fa.
seemed promising, but broke in production. Saw a lot of `IncompleteRead`s on both GETs and POSTs. Rolled back for now.
```
('Connection broken: IncompleteRead(9172 bytes read, -4586 more expected)', IncompleteRead(9172 bytes read, -4586 more expected))
...
File "oauth_dropins/webutil/util.py", line 1673, in call
resp = getattr((session or requests), fn)(url, *args, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "requests_cache/session.py", line 102, in get
return self.request('GET', url, params=params, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "requests_cache/session.py", line 158, in request
return super().request(method, url, *args, headers=headers, **kwargs) # type: ignore
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "requests/sessions.py", line 589, in request
resp = self.send(prep, **send_kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "requests_cache/session.py", line 205, in send
response = self._send_and_cache(request, actions, cached_response, **kwargs)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "requests_cache/session.py", line 233, in _send_and_cache
self.cache.save_response(response, actions.cache_key, actions.expires)
File "requests_cache/backends/base.py", line 89, in save_response
cached_response = CachedResponse.from_response(response, expires=expires)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "requests_cache/models/response.py", line 102, in from_response
obj.raw = CachedHTTPResponse.from_response(response)
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "requests_cache/models/raw_response.py", line 69, in from_response
_ = response.content # This property reads, decodes, and stores response content
^^^^^^^^^^^^^^^^
File "requests/models.py", line 899, in content
self._content = b"".join(self.iter_content(CONTENT_CHUNK_SIZE)) or b""
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "requests/models.py", line 818, in generate
raise ChunkedEncodingError(e)
```
Ryan Barrett