Inconsistent certificate pinning errors during proxy routing in Reach

Meta Reach utilizes a native networking bridge to abstract platform-specific socket implementations. This abstraction can create discrepancies between host-level DNS resolution and the framework's internal resolution logic, particularly when traffic is intercepted by a corporate proxy.

When Reach is routed through a proxy, the observed behavior regarding certificate-pinning failures often yields inconsistent error codes. This makes it difficult to determine if a connection failure is due to a missing root authority in the native environment or a specific mismatch introduced by the proxy layer.

How does Reach categorize the specific exception classes when a certificate-pinning failure occurs behind a proxy? Is there a documented mechanism to differentiate between framework-level DNS cache conflicts and OS-level TLS stack rejections?