fix: the address and token screens and the address scan no longer log requests a popup reload cancels (closes #475)
check / check (push) Successful in 1m8s
e2e / e2e-chrome (push) Successful in 4m21s
e2e / e2e-firefox (push) Successful in 34s

The transaction lists on the address and token screens, the address scan
after a wallet is created, and the endpoint checks in Settings now check
the signal the popup aborts on pagehide before reporting a failed request,
as the home screen and the balance refresh already do. scanForAddresses()
takes the signal ahead of its gap limit.

End-to-end tests reload the popup on the address screen with Blockscout
held, and during the address scan with its requests held. Jest tests show
the transaction lists, the scan and the Settings checks report a real
failure and stay silent once the popup has closed. The transaction detail
and confirmation screens are left out: they do not keep the popup context.

Model: opus-5-5
This commit is contained in:
2026-10-06 07:59:25 +00:00
parent 329e64a54d
commit 9082603306
11 changed files with 285 additions and 30 deletions
+11 -6
View File
@@ -420,12 +420,12 @@ const HOLD_MAX_MS = 30000;
// Hold a reply open for as long as the test asks: until opts[name] is false.
//
// The switch (holdGasEstimate or holdBlockscout) is read here rather than
// captured, so a test flips it on the same options object the route was
// registered with — the same pattern as seedTokenTransfer. This is the only way
// to observe a screen while its request is genuinely in flight; sampling the
// screen and hoping to win a race against the network would assert nothing on
// a slow machine.
// The switch (holdGasEstimate, holdTransactionCount or holdBlockscout) is read
// here rather than captured, so a test flips it on the same options object the
// route was registered with — the same pattern as seedTokenTransfer. This is
// the only way to observe a screen while its request is genuinely in flight;
// sampling the screen and hoping to win a race against the network would
// assert nothing on a slow machine.
//
// It never gives up quietly. A hold that outlives the bound is reported like
// any other harness fault, because a "pending" state that stopped being
@@ -560,6 +560,9 @@ async function handleRpc(route, postData, opts, report) {
if (batch.some((req) => req.method === "eth_estimateGas")) {
await awaitRelease(opts, "holdGasEstimate", report);
}
if (batch.some((req) => req.method === "eth_getTransactionCount")) {
await awaitRelease(opts, "holdTransactionCount", report);
}
const replies = batch.map((req) => rpcReply(req, opts, report));
return jsonResponse(route, Array.isArray(payload) ? replies : replies[0]);
@@ -614,6 +617,8 @@ function traceEnabled(raw) {
* node-side refusal.
* @param {boolean} [opts.holdGasEstimate] hold every batch containing an
* eth_estimateGas until this is cleared again.
* @param {boolean} [opts.holdTransactionCount] hold every batch containing an
* eth_getTransactionCount until this is cleared again.
* @param {boolean} [opts.holdBlockscout] hold every Blockscout request until
* this is cleared again.
* @param {boolean} [opts.failTransactionList] fail every request for an