When referredBy is not explicitly provided, the code was using
callingNumber (user part only) combined with the SBC's internal IP
from dlg.remote.uri. This produced a Referred-By with an IP address
(e.g. sip:+number@10.x.x.x) that carriers like Verizon reject.
Instead, extract the full sip: name-addr from the P-Asserted-Identity
header, which preserves both the display name and the original FQDN
(e.g. "Cha Van" <sip:+number@ims.lte.wal.verizon.com>). Falls back
to the previous behavior when PAI is not available.
Co-Authored-By: Claude Opus 4.6 <noreply@anthropic.com>
* add null check for eventHook
* move guard to superclass, remove logging that adds no value
---------
Co-authored-by: rhonda hollis <rhonda@jambonz.org>
Co-authored-by: Dave Horton <daveh@beachdognet.com>
Corrects the parameter passed to _lccBoostAudioSignal to use
opts.boostAudioSignal instead of the entire opts object, ensuring
the boostAudioSignal option works correctly.
🤖 Generated with [Claude Code](https://claude.com/claude-code)
Co-authored-by: Claude Sonnet 4.5 <noreply@anthropic.com>
* fix: propagate SIP 488 error to SBC on endpoint allocation failure
When FreeSWITCH returns a SIP 488 'Not Acceptable Here' error during
endpoint allocation (e.g., codec incompatibility), this error was not
being propagated back to the SBC/client. Instead, the call would wait
indefinitely for websocket commands or return a generic 603 response.
Implementation:
- In _evalEndpointPrecondition(), detect SipError by checking
err.type === 'SipError' or err.name === 'SipError'
- Extract the SIP status code (e.g., 488), reason, and the Reason
header from the error response (e.g., Q.850;cause=88;text=INCOMPATIBLE_DESTINATION)
- Send the SIP error response immediately to the SBC with:
- X-Reason header: endpoint allocation failure details
- Reason header: original Q.850 cause from FreeSWITCH
- Notify call status change as Failed with proper SIP status
- Release the call immediately instead of waiting for commands
Also added fallback handling in InboundCallSession._onTasksDone() to
propagate the stored error if immediate send was not possible.
* wip
* Simplify SipError check to only use err.name
* refactor recording
removed the test of `(this.cs.accountInfo.account.record_all_calls || this.cs.application.record_all_calls` from backround-task-manager.jsL138 as this check is already done in call-session.js at Line 3007, also allows us to start the record from update or config verbs
* handle start recording for a call that is not yet answered
* return false if not changing recording state
* different check for status
* set hasRecording flag on callInfo when starting
* update redis on recording start
* lint
* update dependency
oddsix