Commit Graph

2 Commits

Author SHA1 Message Date
lilleman ea42bc6d20 Refuse a single unreadable PDU instead of the whole link (#79)
* Regression tests: a PDU the codec cannot read costs that PDU, not the link

* Refuse a single unreadable PDU instead of the whole link

* smscsim interop asserts the first attempt gets every DLR

* Record the smscsim sequence number defect as fixed

* One framing rule and one response-command lookup, per the architecture review

* Apply the stability review's nits: honest sessionError docs and one link-survives assertion
2026-09-05 21:00:52 +02:00
lilleman 1723a0381c Add the interoperability harness and the smscsim experiment, which confirms the one-bad-PDU link drop 2026-09-05 18:40:56 +02:00