apunkt/plano
1
0
Fork 0
mirror of https://github.com/katanemo/plano.git synced 2026-06-20 15:28:07 +02:00
Code Issues Projects Releases Packages Wiki Activity

fix(llm_gateway): buffer non-streaming response body until end_of_stream

Browse source 
Wait for the full upstream body before JSON parsing to avoid truncated
responses on chunked replies. Retry currency_exchange demo tests on flake.

Co-authored-by: Musa <musa@spherrrical.dev>
This commit is contained in:
Cursor Agent 2026-05-18 18:11:22 +00:00
parent e5f0cee39e
commit 420431b30c
No known key found for this signature in database
2 changed files with 8 additions and 1 deletions
Download patch file Download diff file Expand all files Collapse all files

6
crates/llm_gateway/src/stream_context.rs
View file

@ -1183,6 +1183,12 @@ impl HttpContext for StreamContext {
return Action::Continue;
}
// Non-streaming upstream responses may arrive in multiple chunks; wait for the
// full buffered body before parsing (body_size is cumulative on the final chunk).
if !self.streaming_response && !end_of_stream {
return Action::Continue;
}
// Check if this is an error response from upstream
if let Some(status_code) = &self.upstream_status_code {
if status_code.is_client_error() || status_code.is_server_error() {