You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
{{ message }}
Repository navigation
Mark all resolver transport request with "session designator" #2190
I investigated the current Resolver transport architecture for this issue.
The existing HTTP transport implementations obtain configured headers through "HttpTransporterUtils.getHttpHeaders(session, repository)". Since the session is already available at this point, a shared implementation could potentially expose the session ID without modifying each individual transport implementation or the transport-task SPI.
Before starting an implementation, I'd appreciate clarification on a few design points:
Default behavior: Should Resolver automatically attach a session-designator header to HTTP requests, or should it provide an opt-in mechanism for consumers to configure one?
Header configuration: Would you prefer a dedicated session-designator configuration property or a dynamic supplier mechanism integrated with the existing HTTP header configuration?
Header naming: Should the header name be configurable, and is there an existing naming convention you would prefer?
Scope: Is the primary goal to support HTTP transports, or should the design also provide a session designator to non-HTTP transports?
My initial investigation suggests that a shared HTTP-header utility could avoid changes to the individual transport implementations. I have not modified any files yet and will wait for guidance on the intended behavior before proposing a patch.
Activity
harshab054 commentedon Oct 28, 2026
Hi @cstamas,
I investigated the current Resolver transport architecture for this issue.
The existing HTTP transport implementations obtain configured headers through "HttpTransporterUtils.getHttpHeaders(session, repository)". Since the session is already available at this point, a shared implementation could potentially expose the session ID without modifying each individual transport implementation or the transport-task SPI.
Before starting an implementation, I'd appreciate clarification on a few design points:
My initial investigation suggests that a shared HTTP-header utility could avoid changes to the individual transport implementations. I have not modified any files yet and will wait for guidance on the intended behavior before proposing a patch.
Thank you!