Skip to content

Feature/flaskapi#7

Open
SaiKireeti2905 wants to merge 5 commits into
developfrom
feature/flaskapi
Open

Feature/flaskapi#7
SaiKireeti2905 wants to merge 5 commits into
developfrom
feature/flaskapi

Conversation

@SaiKireeti2905
Copy link
Copy Markdown
Collaborator

cogflow for flaskapi

@ellipsis-dev
Copy link
Copy Markdown

ellipsis-dev Bot commented Nov 2, 2024

Your free trial has expired. To keep using Ellipsis, sign up at https://app.ellipsis.dev or contact us.

@SaiKireeti2905
Copy link
Copy Markdown
Collaborator Author

Don't merge this PR

@SaiKireeti2905 SaiKireeti2905 changed the base branch from main to develop November 4, 2024 10:59
alishalbaf added a commit that referenced this pull request May 12, 2026
network.py
- All async helpers now instantiate httpx.AsyncClient with
  follow_redirects=True to match sync requests' default. Without it
  the async path silently diverged on HTTP->HTTPS hops and any
  trailing-slash redirect (Copilot inline comments #1, #2, #5).
- POST and PATCH success check switched from response.is_success
  (2xx only) to "not response.is_error" (<400) so behaviour matches
  sync requests.Response.ok exactly. Also closes the implicit-None
  return path on a 3xx end-state for PATCH that Copilot flagged on
  the original code (#4).
- GET pagination loop switched from "not is_success" to "is_error"
  for the same reason (3xx pages should not silently break the loop).

components.py
- Dropped the unused make_async_get_request import (#6).
- Tightened docstring: only the MinIO upload runs inside
  asyncio.to_thread, not the YAML parse or file read (#7).
- async_register_component now raises CogflowComponentRegistryError
  when the registry check returns None (transport failure) instead
  of silently treating it as "no existing component" and risking a
  duplicate create during an intermittent registry outage (#8).

tests
- FakeAsyncResponse now exposes is_error derived from is_success so
  existing fixtures keep working without per-test changes.
- New test_async_clients_follow_redirects pins down that every async
  helper instantiates httpx.AsyncClient with follow_redirects=True,
  so this divergence cannot be reintroduced.
- New test_async_register_component_registry_unreachable covers the
  None-from-make_async_get_request_raw path.

Copilot inline comment #3 (DELETE retries on 4xx/5xx through
raise_for_status -> HTTPStatusError) is intentional: the sync
make_delete_request behaves identically (RequestException is the
base class, bare @Retry catches everything). The PR description
will be updated separately to remove the misleading
"transport-error-only" wording for DELETE.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant