engineering/zebra-report
1
0
Fork 0
Code Issues Pull requests Projects Releases Packages Wiki Activity Actions

zebra-spaces: stop subscribe→close→subscribe loop — listeners stay connected

Browse source 
Root cause of the 'late listener can't see screens' symptom: my recent
'rebuild sub PC on failure' commit (6a72e77) triggered on connection
state 'closed' as well as 'failed', AND the guard `if (sfuSubPC === pc)`
only worked because we ASSUMED Chrome would fire the state change
asynchronously. Chrome fires it SYNCHRONOUSLY during pc.close(), at
which point sfuSubPC still points at the closing pc — guard passes,
rebuild fires. Then sfuUnsubscribe closes the new PC, which triggers
another rebuild. SFU log showed listeners cycling
subscribe → 40s → close → subscribe forever.

Fix:
1. onconnectionstatechange now only rebuilds on 'failed' (the actually-
   terminal state). 'closed' = sfuUnsubscribe(), 'disconnected' =
   transient and WebRTC may recover on its own.
2. sfuUnsubscribe nulls sfuSubPC BEFORE pc.close(), so even if the
   handler fired synchronously the === guard would correctly fail.
3. visibilitychange handler also tightened to only fire on 'failed' —
   same reasoning.

The rebuild path for actual ICE failures still works (state goes
'connected' → 'disconnected' → 'failed' → rebuild).

83 FSM tests still green.
This commit is contained in:
Russell Ballestrini 2026-06-02 11:36:05 -04:00
parent 81b237eaa7
commit aa36b84a47
No known key found for this signature in database
1 changed files with 32 additions and 20 deletions
Download patch file Download diff file Expand all files Collapse all files

52
web/zebra-spaces.html
View file

@ -2095,15 +2095,19 @@ async function sfuSubscribe(){
* sfuUnsubscribe + sfuSubscribe. The auto-rejoin code that follows
* a hard refresh handles the WS side; this handles the SFU side. */
pc.onconnectionstatechange = () => {
const s = pc.connectionState;
if (s === 'failed' || s === 'closed'){
logLine('err', 'sfu sub PC ' + s + ' — rebuilding');
if (sfuSubPC === pc){
sfuUnsubscribe().then(() => {
if (wantConnected && roomID) sfuSubscribe().catch(e => logLine('err','sfu re-subscribe: '+e.message));
});
}
}
/* 'failed' is terminal ICE failure (we should rebuild).
* 'closed' is OUR OWN sfuUnsubscribe() — never rebuild on that
* (Chrome fires the state change synchronously before sfuSubPC is
* null'd, which used to trigger a subscribe→close→subscribe loop
* every time anyone deliberately tore down the sub).
* 'disconnected' is transient — let WebRTC try to recover before
* we yank the rug. */
if (pc.connectionState !== 'failed') return;
if (sfuSubPC !== pc) return;
logLine('err', 'sfu sub PC failed — rebuilding');
sfuUnsubscribe().then(() => {
if (wantConnected && roomID) sfuSubscribe().catch(e => logLine('err','sfu re-subscribe: '+e.message));
});
};
/* SSE: server pushes renegotiation offers when publisher set changes.
* We answer each via POST /answer. ping events are keepalive only.
@ -2142,7 +2146,15 @@ async function sfuSubscribe(){
async function sfuUnsubscribe(){
if (sfuSubEvents){ try { sfuSubEvents.close(); } catch(_){} sfuSubEvents = null; }
if (sfuSubPC){ try { sfuSubPC.close(); } catch(_){} sfuSubPC = null; sfuSubPeerID = null; }
/* null out sfuSubPC FIRST, then close — Chrome fires the
* connectionstatechange handler SYNCHRONOUSLY during pc.close(), and
* the handler's `if (sfuSubPC === pc)` guard depends on the global
* already being null. Otherwise the handler thinks the close was a
* 'failed' rebuild trigger and we get a subscribe/close loop. */
const oldPC = sfuSubPC;
sfuSubPC = null;
sfuSubPeerID = null;
if (oldPC){ try { oldPC.close(); } catch(_){} }
sfuStreamsByPubHex.clear();
}
@ -3246,14 +3258,14 @@ refreshMicList();
document.addEventListener('visibilitychange', () => {
if (document.visibilityState !== 'visible') return;
if (!wantConnected || !sfuSubPC) return;
const s = sfuSubPC.connectionState;
if (s === 'failed' || s === 'closed' || s === 'disconnected'){
logLine('err', 'visibility back, sub PC ' + s + ' — rebuilding');
const pc = sfuSubPC;
sfuUnsubscribe().then(() => {
if (wantConnected && roomID) sfuSubscribe().catch(e => logLine('err','sfu re-subscribe: '+e.message));
});
}
/* Only rebuild on terminal failure. 'disconnected' is transient
* (WebRTC tries to recover); 'closed' never happens here because
* sfuUnsubscribe() nulls sfuSubPC before pc.close(). */
if (sfuSubPC.connectionState !== 'failed') return;
logLine('err', 'visibility back, sub PC failed — rebuilding');
sfuUnsubscribe().then(() => {
if (wantConnected && roomID) sfuSubscribe().catch(e => logLine('err','sfu re-subscribe: '+e.message));
});
});
/* ==================================================================
@ -3348,8 +3360,8 @@ logLine('', 'ready — pick a handle, type a rendezvous code, enter the space');
<footer style="margin:2.2rem auto 0;font-size:0.65rem;color:#999;line-height:1.7;word-break:break-all;font-family:monospace">
<span id="pi-seal" style="color:#777;cursor:default;user-select:none" title="">page integrity</span> &nbsp;·&nbsp; built <span class="stamp-date">2026-06-02</span><br>
md5 <span class="stamp-md5">e7c6d045cf7204add40ff3ae905eab0f</span><br>
sha256 <span class="stamp-sha">76761abd84bcfee8d479b69c8d4a68f90bf8133b15c7b2b99b817732f31d2deb</span><br>
md5 <span class="stamp-md5">e889a0a868e23e2be985f28db80ff031</span><br>
sha256 <span class="stamp-sha">e2d9440bd38fc3081eaafe68d972d3ac8a856ace24da919c6eb6b44ec9a0a7d2</span><br>
<span style="color:#bbb">hashes are of this page with these two fields zeroed — to verify, blank them and re-hash</span><br>
<span style="color:#bbb">one self-contained file — <strong>save a copy</strong> and verify against these hashes; point at your own servers with ?signal= and ?turncred=, or <a href="host-your-own.html" style="color:#999">host your own community</a></span>
</footer>