centml/sdk/api.py (1)

176-189: Add a repeated page-token guard to avoid potential infinite pagination loops.

If the backend ever returns the same non-empty next_page_token repeatedly, this loop will not terminate.

Suggested hardening

         all_events = []
         next_page_token = None
+        seen_page_tokens = set()

         while True:
             response = self._api.get_deployment_logs_v3_deployments_logs_v3_deployment_id_revision_number_get(
                 deployment_id=deployment_id,
                 revision_number=revision_number,
                 start_time=start_time,
                 end_time=end_time,
                 next_page_token=next_page_token,
                 start_from_head=start_from_head,
                 line_count=line_count,
             )
             all_events.extend(response.events)
             next_page_token = response.next_page_token
             if not next_page_token:
                 break
+            if next_page_token in seen_page_tokens:
+                raise RuntimeError("Pagination token repeated; aborting to avoid infinite loop")
+            seen_page_tokens.add(next_page_token)

🤖 Prompt for AI Agents

Verify each finding against the current code and only fix it if needed.

In `@centml/sdk/api.py` around lines 176 - 189, The pagination loop using
next_page_token in the method that calls
_api.get_deployment_logs_v3_deployments_logs_v3_deployment_id_revision_number_get
should guard against repeated non-empty tokens to avoid infinite loops: track
seen tokens (e.g., a set seen_tokens) and after each response check if
next_page_token is non-empty and already in seen_tokens—if so, either break
and/or raise a descriptive exception; otherwise add the token to seen_tokens and
continue. Update the loop that extends all_events and assigns next_page_token to
incorporate this repeated-token check and fail-safe.

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes