tree 4d19e432a944074b4787a5b9da6a7ed73d0a1485
parent 8c09eaec8f03d257255929387acc81069d4550cc
author Boris Zbarsky <bzbarsky@apple.com> 1668694903 -0500
committer GitHub <noreply@github.com> 1668694903 -0500
gpgsig -----BEGIN PGP SIGNATURE-----
 
 wsBcBAABCAAQBQJjdkN3CRBK7hj4Ov3rIwAAyoYIAGvZn33zh7Fe41W+O8ggMj4l
 9H7wU41xeUZ4YYDOiAey0OIhuE6suoCMI8MofB6zB4PcxCs5SiquVazDxnybtMzy
 PY8z/OcU9VMK9CoYrD2ne3NXDXtciqLJDO/cf5ueIu664P4635IUJpdfQtUXAQgq
 9o79x6v+EeMx/eX6y9sH8ZSL9iqqBDexUcGwSwvXW2IMQHpD5yl79fkpXLwIMDr7
 PQenQv7ARsuNPDGJnzhLrB7z8DMMhOcxVexHx9ZQQqtYhSpJSYSgP3Yq61GPzv+g
 qNG7wopanzqPDNrhgTJNsR0YKtceTRK98IaNzAvczeFjPuj6fl8T4BtuMzhaNPk=
 =AsPy
 -----END PGP SIGNATURE-----
 

Improve logic for deciding whether to re-establish CASE in ReadClient. (#23627)

* Improve logic for deciding whether to re-establish CASE in ReadClient.

We could end up in a situation where we had a defunct session but did not have the "try to re-establish CASE" flag set.  In that case we would keep trying resubscribe attempts, which would keep failing because we could not actually create an exchange on our session, and we would never recover from that.

The fix is that we try to re-establish CASE (assuming the IM engine
has a CASE Session manager) if we don't have an active session.

Also fixes an incorrect error ("no memory") being reported if we in
fact try to subscribe when our session is not active.

* Address review comments.