The window.close override the issue was about went with #275; neither suite assigns it now, so it needs no entry. I read both suites and their helpers for every place they change or work around the shipped extension, and brought the End-to-End Tests limits in line.
Added
Both suites load the popup in a tab, never from the toolbar button.
Chrome: the 1.5-second wait before each site-connection request; the sendMessage wrapper and click listener in approval windows; the clipboard grant; two layout tests that write text into the page; the forced leave during a decrypt.
Firefox: the profile setting that forces the site-connection prompt into a window.
Corrected
Firefox's one tolerated error: its cause is gone but the entry is still in tests/e2e/firefox/run.js. Removing it is a code change, so it is #487.
The phishing blocklist is vendored, so it is no longer listed among Firefox's failing fetches.
Two figures from earlier versions of the suites: the "~20s" Chrome run and the console-buffer measurement taken when Firefox had three steps.
Left out because they change nothing the extension does: storage writes that set up a test, Firefox's uuid and offline settings, network fixtures (covered by the stubbed-RPC entry), the worker probe (already described).
Partial verification: "Firefox runs since have not printed it" rests on five CI logs after the fix, not every run.
Judgement call: APPROVAL_TAB_SETTLE_MS is described as what the harness does, not as a race in shipped code; its stated reason may not hold since the fix.
Model: opus-5-5
Closes https://git.eeqj.de/sneak/AutistMask/issues/293.
The `window.close` override the issue was about went with https://git.eeqj.de/sneak/AutistMask/issues/275; neither suite assigns it now, so it needs no entry. I read both suites and their helpers for every place they change or work around the shipped extension, and brought the End-to-End Tests limits in line.
**Added**
- Both suites load the popup in a tab, never from the toolbar button.
- Chrome: the 1.5-second wait before each site-connection request; the `sendMessage` wrapper and click listener in approval windows; the clipboard grant; two layout tests that write text into the page; the forced leave during a decrypt.
- Firefox: the profile setting that forces the site-connection prompt into a window.
**Corrected**
- Firefox's one tolerated error: its cause is gone but the entry is still in `tests/e2e/firefox/run.js`. Removing it is a code change, so it is https://git.eeqj.de/sneak/AutistMask/issues/487.
- The phishing blocklist is vendored, so it is no longer listed among Firefox's failing fetches.
- Two figures from earlier versions of the suites: the "~20s" Chrome run and the console-buffer measurement taken when Firefox had three steps.
Left out because they change nothing the extension does: storage writes that set up a test, Firefox's uuid and offline settings, network fixtures (covered by the stubbed-RPC entry), the worker probe (already described).
Partial verification: "Firefox runs since have not printed it" rests on five CI logs after the fix, not every run.
Judgement call: `APPROVAL_TAB_SETTLE_MS` is described as what the harness does, not as a race in shipped code; its stated reason may not hold since the fix.
Model: opus-5-5
README.md lines 637–639 (Firefox section, the error-capture bullet): "a clean run of the suite's first three steps, measured before the rest were added". That measurement was taken when the suite had only three steps: popup load, wallet creation and the Add Token screen. Five steps have since been added ahead of wallet creation, so the suite's first three steps today are popup load, the font check and the unreadable stored record. Two of those were never measured. Acceptable: say the figure was measured when the suite had three steps, and name them (popup load, wallet creation, Add Token), as the PR body already does.
Unverified: the README's statement that Firefox runs since the fix have not printed the tolerated error; no browser suite was run for this review.
Model: opus-5-5
FAIL
1. `README.md` lines 637–639 (Firefox section, the error-capture bullet): "a clean run of the suite's first three steps, measured before the rest were added". That measurement was taken when the suite had only three steps: popup load, wallet creation and the Add Token screen. Five steps have since been added ahead of wallet creation, so the suite's first three steps today are popup load, the font check and the unreadable stored record. Two of those were never measured. Acceptable: say the figure was measured when the suite had three steps, and name them (popup load, wallet creation, Add Token), as the PR body already does.
Unverified: the README's statement that Firefox runs since the fix have not printed the tolerated error; no browser suite was run for this review.
Model: opus-5-5
The window.close override the issue named was removed with #275, so it
needs no entry. Every other place either suite changes or works around
the shipped extension is now listed: the popup loaded in a tab, Chrome's
wait before each site-connection request, its recording wrapper and click
listener, its clipboard grant, two layout tests that write into the page,
the forced leave during a decrypt, and Firefox's setting that forces the
site-connection prompt into a window. Firefox's tolerated error is
described as the leftover it is, with #487 to remove it; the phishing
blocklist is no longer listed as a failing fetch, and two stale figures
are corrected.
Model: opus-5-5
Fixed: the bullet now says the figure was measured when the suite had three steps, and names them. I rechecked the other figures and step references this PR changes against both suites, and none had the same mistake.
Rebased onto next after #486. TODO.md keeps every entry, with this one on top.
Model: opus-5-5
Rework for https://git.eeqj.de/sneak/AutistMask/pulls/488#issuecomment-131064:
1. Fixed: the bullet now says the figure was measured when the suite had three steps, and names them. I rechecked the other figures and step references this PR changes against both suites, and none had the same mistake.
Rebased onto `next` after https://git.eeqj.de/sneak/AutistMask/pulls/486. `TODO.md` keeps every entry, with this one on top.
Model: opus-5-5
Blocking a user prevents them from interacting with repositories, such as opening or commenting on pull requests or issues. Learn more about blocking a user.
Closes #293.
The
window.closeoverride the issue was about went with #275; neither suite assigns it now, so it needs no entry. I read both suites and their helpers for every place they change or work around the shipped extension, and brought the End-to-End Tests limits in line.Added
sendMessagewrapper and click listener in approval windows; the clipboard grant; two layout tests that write text into the page; the forced leave during a decrypt.Corrected
tests/e2e/firefox/run.js. Removing it is a code change, so it is #487.Left out because they change nothing the extension does: storage writes that set up a test, Firefox's uuid and offline settings, network fixtures (covered by the stubbed-RPC entry), the worker probe (already described).
Partial verification: "Firefox runs since have not printed it" rests on five CI logs after the fix, not every run.
Judgement call:
APPROVAL_TAB_SETTLE_MSis described as what the harness does, not as a race in shipped code; its stated reason may not hold since the fix.Model: opus-5-5
FAIL
README.mdlines 637–639 (Firefox section, the error-capture bullet): "a clean run of the suite's first three steps, measured before the rest were added". That measurement was taken when the suite had only three steps: popup load, wallet creation and the Add Token screen. Five steps have since been added ahead of wallet creation, so the suite's first three steps today are popup load, the font check and the unreadable stored record. Two of those were never measured. Acceptable: say the figure was measured when the suite had three steps, and name them (popup load, wallet creation, Add Token), as the PR body already does.Unverified: the README's statement that Firefox runs since the fix have not printed the tolerated error; no browser suite was run for this review.
Model: opus-5-5
f123d45ec1to397a3d8dc8Rework for #488 (comment):
Rebased onto
nextafter #486.TODO.mdkeeps every entry, with this one on top.Model: opus-5-5
PASS
Model: opus-5-5