* Cover the ids a send gets back where the SMSC named only some of them
* Say undefined where the SMSC named no id for a segment it took
* Record why the merger takes the wider type, and de-duplicate the interop narrowing
* Guard the numbering hole the merger must refuse, and correct the phase record
* Regression tests for the concatenated segment budget and the encoding-name door
* Budget unpacked segments at 134 octets, and export isEncodingName
* Name the length a Latin-1 message starts costing more segments at
* Fold the totality clause into hard rule 1, and log the unrepresentable-alphabet finding
* Make the maxSegments regression discriminate, and drop the narrating comments
* State the per-alphabet capacities accurately, and log the invalid-date path
* Fixtures for the receipt bodies operator documentation publishes
* Read the stat:FAILED several operators write as UNDELIVERABLE
* Cover the 16-bit UDH concatenation element with a fixture
* Record phase 9 of the interoperability plan
* Share the dummy SMSC and tighten the operator fixtures
* Correct what the test conventions claim about the doubles and the tree
* Read a receipt date that carries its century
* Read the stat spelling CM.com publishes
* Say what the receipt status table actually holds
* Regression tests for inbound sar_* concatenation
* Reassemble a concatenated message segmented with the sar_* TLVs
* Record the sar_* fix in the jasmin and Java-client findings
* Name the sar_* TLVs in the refusal, and separate group keys the addresses cannot forge
* Name the field a refused segment got wrong, and key groups on a separator no address holds
* Bound the new session tests, and derive Concat from the UDH fields
* Regression tests for aborting a send queued behind a full window
* Let a signal abort a send queued behind a full send window
* Log a queued send, tighten its regression tests and settle the decision record
* Take the send window's queue off an array so an abort cannot stall the loop
---------
Co-authored-by: Mikael Göransson <mikael.goransson@timewave.se>
* Regression tests for a body in message_payload and for data_sm
* Read a body from message_payload and accept data_sm
* Assert the fixed message_payload and data_sm behaviour against Jasmin
* Record the message_payload and data_sm fixes in the Jasmin findings
* Read an inbound data_sm as the direction it travelled, and export messageOctets
* Refuse a segment with the code its stand-in command defines
* Name the stand-in the refusal status is read from
* Regression tests for answering every inbound segment as it arrives
* Answer every segment of an inbound concatenated message as it arrives
* Record the segment-by-segment answer in AGENTS.md and README
* Regression tests for an empty message_id on deliver_sm_resp
* Answer a deliver_sm with the empty message_id SMPP 3.4 makes it
* Record Jasmin's refusal of a deliver_sm_resp message_id
* Mark the Jasmin multipart deadlock fixed
* Regression tests for the architecture review's findings
* Give the segment id notation an owner, and every segment a status
* Correct what the ids reach and what a lost group tells the application
* Regression tests for a group lost to its own octet overrun
* Report a group lost to its own overrun, and refuse by the command it arrived on
* Keep the docs true about what a segment is answered with
* Count only the answered segments of a group lost to an overrun
* Report only what a lost group cost, and say which cap bit
* Regression tests for telling a refused PDU from a failed session
* Export PduRefusedError so a refusal can be told from a session failure
* Document telling a refused PDU from a session failure
* Share the malformed-PDU fixtures and pin that a dead stream is no refusal
* Say once what a refusal is, and why the header type is published
* Fold the refused-response test into the one that already staged it
* Drop the refusal-rate claim the interop finding does not support
* Build a raw PDU one way, sequence number included
* Follow PduRefusedError into pdu-refusal.ts
* Regression tests for a client that retries its first connect and bind
* Retry the first connect and bind on reconnect: { fromStart: true }
* Pin that the signal bounding the first bind also closes the session
* Name a failed connect once, default the backoff where it lives, and fix the deadline advice
* Carry the failure that started the retries on the abort that ends them
* Pin one retry announcement per attempt when a link drops mid-bind
* Announce a retry when its wait is over, not when it is scheduled
* Say what the process-hold assertion expects, not what a failure would mean
* Regression tests for a receipt body read under a data_coding it is not written in
* Read a delivery receipt's body as the octets that arrived, not by its data_coding
* Assert SMPPSim's UCS2 receipt body parses like every other segment's
* Record the receipt-body defect as fixed
* Retain a multipart segment's octets once, not twice
* Sort the new type import into its file's order
maxDelay bounds the wait and sets the bar a link must clear, and only the
first was documented, so tuning it up quietly lengthened every recovery.
The unreadable PDU the tests write is one named constant now.
A bind that returns is not proof the link works: an unreadable stream is
only found afterwards, so resetting the delay there handed a link that
died on arrival a fresh minDelay every cycle — one connect and bind per
second forever, which is how an account gets blocked for bind flooding.
The loop resets only once a link has outlasted maxDelay.
Also covers the unreadable stream that has no loop to retry it.
A framing or codec error tears the link down rather than the session, so
the loop retries it on a fresh socket with a fresh framer — which is what
a desynced stream needs. Removing that reset failed nothing before; the
reconnect test now leaves half a PDU on the dying link, and does.
`disconnected` counts failed links rather than outages, which the README
now says, and the transport wires its socket as it is built.
Ending a session while the link was down emitted nothing: the retried
drop had already torn it down, so close() found nothing left to do. The
terminal event is now guarded by its own flag rather than by `closed`.
The socket, the framer and the parse move to PduTransport, which is what
takes session.ts back under its line cap; `session.sock` reads through a
getter. The backoff also refuses an inverted pair and an unknown key.
`close` fires only where nothing will bring the session back; a drop the
reconnect loop will retry is `disconnected`, pairing with `reconnected`.
`minDelay: 0` never doubles, so the backoff never started — both delays
are range-checked now, and only `false` spells reconnect off.
Surviving a drop is most of what the session layer is for, and behind an
opt-in an application that never read the options table got none of it.
`reconnect` still takes `{ minDelay, maxDelay }` to retune the backoff;
`false` is the one spelling for off.
lilleman