10 Commits
Author SHA1 Message Date
clawbot ca18beb97f docs: put Sepolia in scope and write down the provider flag exception (closes #165)
check / check (push) Waiting to run
e2e / e2e-chrome (push) Waiting to run
e2e / e2e-firefox (push) Waiting to run
The README's 1.0 non-goal said Ethereum mainnet only, while Sepolia
ships as a selectable network. It now puts mainnet and Sepolia in scope
and other chains out of it.

The injected provider's flag named after the other wallet stays, as an
interface-compatibility flag many dApps check. It is now written down
as an exception to the rule against naming competitors, in the README's
Policies section and in RULES.md, and the comment beside it says so.
Neither document spells the name, so script/check-censored, which
already allows the flag only in src/content/inpage.js and its built
copies, is unchanged.

Model: opus-5-5
2026-10-07 23:58:49 +02:00
clawbot bb60b399ec test: tighten two checks in the persisted-field harness (closes #379)
check / check (push) Waiting to run
e2e / e2e-chrome (push) Waiting to run
e2e / e2e-firefox (push) Waiting to run
The polarity check now counts only `hostile` and `falsy` values, not a
`hostileRestore` value, which reaches only the views its entry names.
The stale index 5 moves into `hostile` for `selectedWallet` and
`selectedAddress`, as each field's one value still truthy after the
floor.

Each `hostileRestore` entry declares whether its boot lands on its view
or falls back to Home, and the test checks the one view on screen. A
value driven onto every restorable view lists only the views it falls
back on, so a view added later is driven by default.

The header and the README restate the remaining limits as built and say
which boots are held to where the popup lands.

Model: opus-5-5
2026-10-07 11:59:14 +02:00
clawbot 447d714313 fix: show every password error in its own fixed-height line (closes #493)
check / check (push) Waiting to run
e2e / e2e-chrome (push) Waiting to run
e2e / e2e-firefox (push) Waiting to run
The add wallet screen reported a missing, short or mismatched password in
the flash line at the top of the popup, and the private key export,
recovery phrase and delete wallet screens each wrote to a line of their own
above the password field. All four now use showError() and hideError() with
a fixed-height error line below the field, as the send confirmation and
approval screens do. On the add wallet screen the line sits beside the
Import button, which keeps its place at 360x600. The line clears when the
screen is shown again and when the password is tried again. Other add
wallet messages stay in the flash line.

Model: opus-5-5
2026-10-07 10:59:16 +02:00
clawbot 29ba54d5b6 docs: record the pre-1.0 security review in TODO.md (closes #383)
check / check (push) Waiting to run
e2e / e2e-chrome (push) Waiting to run
e2e / e2e-firefox (push) Waiting to run
The security review moves from Next Step to Completed Steps, saying what it
read (the tree at 99292b9), that its ten findings are filed and fixed on next,
which owner decisions it raised are still open, and what it did not cover.
Next Step takes the one Future Steps item, cutting 1.0.0 once the milestone is
empty. Status drops a dated gate result and links the current milestone PR.

Model: opus-5-5
2026-10-07 09:43:07 +02:00
clawbot e3dd0e44da chore: prune landed remote branches (closes #167)
check / check (push) Waiting to run
e2e / e2e-chrome (push) Waiting to run
e2e / e2e-firefox (push) Waiting to run
Eighteen branches on origin that the issue classifies as landed, or
superseded by merged pull requests, were deleted; the evidence for each
is on the issue. Four whose work is not in next are kept: the three
issue 87 error-display branches, reference for issue 493, and
chore/token-list-enrichment, re-created at f7a2437, pending issue 495.
TODO.md records this.

The branch-pruning Future Step had already left TODO.md in the issue 191
rewrite, so only the Completed Steps entry is added.

Model: opus-5-5
2026-10-07 09:09:07 +02:00
clawbot 860db6034c docs: README says why the wallet never clears the clipboard (closes #492)
check / check (push) Waiting to run
e2e / e2e-chrome (push) Waiting to run
e2e / e2e-firefox (push) Waiting to run
The owner's ruling on #19 is
now in the README, under ExportPrivKey: copying the key leaves it on the
clipboard, because the clipboard is the user's, clearing it would go
against what they expect, and it could destroy something else they
copied since. ShowRecoveryPhrase copies the phrase the same way and
points back to it. Both describe the warning each password screen
shows on next, which does not mention the clipboard.

Model: opus-5-5
2026-10-07 08:09:06 +02:00
clawbot cd8e45ae0c test: the Firefox suite tolerates no extension error (closes #487)
check / check (push) Waiting to run
e2e / e2e-chrome (push) Waiting to run
e2e / e2e-firefox (push) Waiting to run
The one entry in ALLOWED_ERRORS in tests/e2e/firefox/run.js, Firefox
reporting a popup promise that settled after the page unloaded, lost its
cause when the site-connection buttons stopped sending with an unawaited
sendMessage before closing (#275). Left in place it would also hide the
same error from any other popup code that sends and then closes. The
entry goes, with the code that only printed and set aside tolerated
errors, and the README paragraph that described it.

Model: opus-5-5
2026-10-07 07:26:09 +02:00
clawbot eeb10c20ef chore: draw the toolbar icons in-tree with make icons (closes #378)
check / check (push) Waiting to run
e2e / e2e-chrome (push) Waiting to run
e2e / e2e-firefox (push) Waiting to run
script/lib/icons.js draws the mark from geometry read back out of the
committed PNGs, since the coordinates were never recorded, and writes each
PNG with node's own zlib. `make icons` runs it and leaves alone a file that
already holds the drawn image.

tests/icons.test.js requires each committed file to hold exactly that
image: the same IHDR and every pixel. The compressed bytes are not
compared, because node's bundled zlib does not compress as the stock zlib
that made the committed files did; the decision is recorded on the issue.

build.js copies the manifests from MANIFEST_SOURCES instead of naming the
two paths a second time.

Model: opus-5-5
2026-10-07 06:43:08 +02:00
clawbot 0aaa94471f docs: README end-to-end limits match what the browser suites do (closes #293)
check / check (push) Waiting to run
e2e / e2e-chrome (push) Waiting to run
e2e / e2e-firefox (push) Waiting to run
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
2026-10-07 05:59:15 +02:00
clawbot 763b50b0e5 fix: Back from Settings never lands on Settings itself (closes #481)
check / check (push) Waiting to run
e2e / e2e-chrome (push) Waiting to run
e2e / e2e-firefox (push) Waiting to run
The recovery phrase and delete wallet screens take themselves off the
Back stack when left, so Settings, one of those screens, then the
settings gear leaves Settings under the Settings now showing. A reopened
popup restores the same stack, cut at the screen the gear left. Back
then showed Settings again and seemed to do nothing.

goBack() now skips any entry for the screen already showing before it
pops its target. Jest tests drive the gear and then Back for the
recovery phrase screen, once and twice over, for both delete screens,
and after a reopen.

Model: opus-5-5
2026-10-07 05:26:06 +02:00
23 changed files with 1071 additions and 271 deletions
+8 -1
View File
@@ -1,4 +1,4 @@
.PHONY: bootstrap setup install test test-e2e test-e2e-firefox lint fmt fmt-check check check-censored docker hooks build build-debug package vendor-blocklist clean dev
.PHONY: bootstrap setup install test test-e2e test-e2e-firefox lint fmt fmt-check check check-censored docker hooks build build-debug package vendor-blocklist icons clean dev
# Standard targets are thin shims; the implementations live in script/
# per the scripts-to-rule-them-all pattern (see the Entrypoints section
@@ -104,6 +104,13 @@ build-debug:
vendor-blocklist:
@script/vendor-blocklist
# Redraw the toolbar icons in icons/ from script/lib/icons.js, at every size
# manifest/chrome.json declares, leaving alone a file that already holds the
# drawn image. Commit the result with the drawing: tests/icons.test.js fails
# while the two disagree.
icons:
@node script/lib/icons.js
clean:
@rm -rf dist/ release/
+119 -64
View File
@@ -317,6 +317,11 @@ The Makefile shims to those. It also carries a few targets that have no
- `make build-debug` — the same build with `AUTISTMASK_DEBUG=1`, verified as a
debug build, and keeping its `dist/` on failure (see
[Debug Builds](#debug-builds))
- `make icons` — redraw the toolbar icons in `icons/` from
`script/lib/icons.js`, at every size `manifest/chrome.json` declares. A file
that already holds the drawn image, the same IHDR and every pixel, is left
untouched. Commit the files with the drawing: `make check` fails while their
image differs from what it draws
- `make clean` — remove `dist/` and `release/`
- `make dev` — run the build `make build` runs, without the checks that follow
it, then run it again after every change to a file under `src/`, `manifest/`
@@ -334,7 +339,9 @@ There are two suites, one per browser, and they share no code. Chrome runs on
Playwright; Firefox has its own WebDriver client, because Playwright cannot
observe errors on a Firefox extension page at all — see
[Firefox](#firefox-make-test-e2e-firefox) below. Both require docker, and both
are outside `make check`.
are outside `make check`. Neither opens the popup from the toolbar button: both
load its page in an ordinary tab, so what the toolbar popup itself adds, its
size and its closing when it loses focus, is covered by neither.
### Chrome (`make test-e2e`)
@@ -420,14 +427,36 @@ required to be present, so that check cannot pass by observing nothing. That
last one is the standing floor under
[#157](https://git.eeqj.de/sneak/AutistMask/issues/157).
Two limits of that coverage, neither of them papered over. The RPC is stubbed
throughout, so this is **not** a real dApp against a real network with real
funds; that remains a human pass before 1.0.0. The site-connection prompt is
raised through `chrome.action.openPopup()`, and headless Chromium's
browser-action popup is not a page Playwright can see or click, so that one
prompt is driven at the URL the extension itself puts on the action — the same
page and the same approval id, but whether a real toolbar click shows it is not
observable here.
The limits of that coverage and of the rest of the Chrome suite, none of them
papered over:
- The RPC is stubbed throughout, so this is **not** a real dApp against a real
network with real funds; that remains a human pass before 1.0.0.
- The site-connection prompt is raised through `chrome.action.openPopup()`, and
headless Chromium's browser-action popup is not a page Playwright can see or
click, so that one prompt is driven at the URL the extension itself puts on
the action — the same page and the same approval id, but whether a real
toolbar click shows it is not observable here.
- Each site-connection request is made 1.5 seconds after the tab it will be
driven in is opened (`APPROVAL_TAB_SETTLE_MS` in `tests/e2e/run.js`), because
opening that tab closes the previous prompt's toolbar popup; a request made
just after that popup closes is not covered.
- In the tests that approve a signature or a transaction, the approval window
runs with `chrome.runtime.sendMessage` wrapped to record what it sends, and in
the tests that reject a site connection the prompt gets a click listener that
records that Reject was pressed; neither changes what the window does.
- The tap to copy test first grants clipboard permission to every page in the
browser, which the manifest does not ask for, so whether a real popup may
write to the clipboard on a click alone is not covered.
- The layout tests for an over-long flash message and for the password error
lines write the text straight into the page instead of letting the popup's
code put it there, and the second brings each screen up by toggling its
`hidden` class rather than navigating to it; they cover the layout, not the
code that fills it.
- Leaving the recovery phrase or private key screen while its decrypt runs is
forced by clicking Reveal and the settings gear in one page task, which a
person cannot do; the case a person can hit, the first decrypt after the popup
opens while libsodium is still loading, is not driven.
Any test that drives a failure path on purpose declares the `console.error` it
is about to provoke, via `errors.expect()`. That is not a mute: the declaration
@@ -441,8 +470,8 @@ collecting for a fixed grace period after the last test returns
the context. A request whose _first_ dispatch falls after that window is never
seen at all and cannot fail the run. In practice a request a test fires without
awaiting reaches the route handler about 10ms later, and anything on a repeating
timer gets observed on an earlier tick during the ~20s suite — but a one-shot
call deliberately deferred past the window will escape.
timer gets observed on an earlier tick during the suite — but a one-shot call
deliberately deferred past the window will escape.
That interception covers the MV3 background service worker as well as the popup
page, which it does not by default — `script/test-e2e` sets
@@ -574,25 +603,18 @@ without an `await`. Demonstrated, not assumed: a `throw` placed past the first
and on Chrome (`pageerror`), with the rest of the run unaffected because the
approval view had already rendered.
One error is tolerated rather than fatal, listed in `ALLOWED_ERRORS` in
`tests/e2e/firefox/run.js` with the issue that will delete it, and printed on
every occurrence so the concession stays visible in the run output. It is
Firefox reporting the site-approval popup's unawaited `sendMessage` settling
after `window.close()` unloaded the context — the same teardown ordering as
[#275](https://git.eeqj.de/sneak/AutistMask/issues/275), and unsuppressable from
the calling code, because `BaseContext.wrapPromise` reports it whether or not a
handler is attached. Errors are read from the privileged `nsIConsoleService` in
Marionette's chrome context and filtered to non-warning entries whose
`sourceName` is the extension origin. That mechanism is not a stylistic choice.
WebDriver BiDi's `log.entryAdded` delivers **nothing** for extension pages: on a
plain `http://` page it reports uncaught errors with stack traces, and on the
`moz-extension://` popup it reports zero events, because Firefox's remote agent
excludes extension browsing contexts from BiDi observation. Any harness built on
Playwright-BiDi or Puppeteer-BiDi would therefore see nothing and report
success, which is exactly the vacuous check this repo has already shipped twice.
Do not migrate this suite to BiDi.
Errors are read from the privileged `nsIConsoleService` in Marionette's chrome
context and filtered to non-warning entries whose `sourceName` is the extension
origin. That mechanism is not a stylistic choice. WebDriver BiDi's
`log.entryAdded` delivers **nothing** for extension pages: on a plain `http://`
page it reports uncaught errors with stack traces, and on the `moz-extension://`
popup it reports zero events, because Firefox's remote agent excludes extension
browsing contexts from BiDi observation. Any harness built on Playwright-BiDi or
Puppeteer-BiDi would therefore see nothing and report success, which is exactly
the vacuous check this repo has already shipped twice. Do not migrate this suite
to BiDi.
Two limits are worth knowing, both real differences from the Chrome suite:
Three limits are worth knowing, all real differences from the Chrome suite:
- **Error capture is poll-based, not event-streamed.** The console is drained at
each step boundary, so an error is attributed to the step it was drained
@@ -609,24 +631,30 @@ Two limits are worth knowing, both real differences from the Chrome suite:
and silently evicts the oldest, so more than 250 console messages between two
drains destroys the excess unread. 400 throws inside one step are reported as
exactly the newest 250, three runs running. That buffer is shared with
Firefox's own console noise; a clean run peaks at 4 of 250 at the install
drain and 0 at every later drain, so the three steps here have wide headroom,
but a step that logs heavily could evict unread errors. What poll-based costs
is location, not coverage: an error cannot be placed within a step the way the
Chrome suite's `pageerror` events place it.
Firefox's own console noise; a clean run, measured when the suite had three
steps (popup load, wallet creation and Add Token), peaked at 4 of 250 at the
install drain and 0 at every later drain, but a step that logs heavily could
evict unread errors. What poll-based costs is location, not coverage: an error
cannot be placed within a step the way the Chrome suite's `pageerror` events
place it.
- **Almost nothing is stubbed, which inverts the coverage of network-dependent
code.** The container still runs with `--network none`, so the run is offline
and no request can escape. The one thing it can reach is the loopback fixture
in `tests/e2e/firefox/dapp.js`, which serves the dApp page and a JSON-RPC node
and which the extension's `rpcUrl` is pointed at for the dApp steps; a
JSON-RPC method that fixture does not model fails the run rather than
answering `null`. Everything else — Blockscout, the price feed, the phishing
blocklist — has no fixture and simply fails, and the extension swallows its
own fetch failures, so only the _failure_ branches of that code are ever
executed. A `ReferenceError` in the success path of `renderTransactions`, or
of price rendering, passes this suite green. The offline run is also weaker
than the Chrome suite's interception for those calls: it proves nothing got
out, but it cannot report which requests were attempted.
answering `null`. Everything else, Blockscout and the price feed among it, has
no fixture and simply fails, and the extension swallows its own fetch
failures, so only the _failure_ branches of that code are ever executed. A
`ReferenceError` in the success path of `renderTransactions`, or of price
rendering, passes this suite green. The offline run is also weaker than the
Chrome suite's interception for those calls: it proves nothing got out, but it
cannot report which requests were attempted.
- **The site-connection prompt always opens in a window of its own.** The
profile turns off `extensions.openPopupWithoutUserGesture.enabled`, so
`src/background/index.js` falls back from the toolbar popup, which WebDriver
cannot see, to `windows.create()`; the toolbar popup path is not covered on
Firefox.
Neither `make test-e2e` nor `make test-e2e-firefox` is part of `make check` or
`make test`. `REPO_POLICIES.md` caps `make test` at 60 seconds and a browser
@@ -1235,14 +1263,21 @@ path rather than on the home screen. Read the claim narrowly, as that file
states it: what those boots prove is no structural dereference on the code paths
a WHOLLY-CORRUPTED PROFILE takes, which is not every path a stored record takes.
Not driven: any pairing of values the four slots do not produce, a view only
forward navigation opens, anything behind a click, and everything a healthy
profile reaches. Within that boundary the verdict is unconditional — if one of
those boots leaves the popup unhealthy or off the view it stored, `make check`
fails, including when it takes two corrupted fields at once, because the verdict
is the combined boot and the per-field re-boot that names a culprit can only
decorate the message. So does a field that gains a floor while its row still
claims it has none, and so does a field added to `PERSISTED_FIELDS` with no row
at all. The per-field justification that used to live in the header of
forward navigation opens, anything behind a click or a timer, and, of what a
healthy profile reaches, anything beyond its boot onto each view the popup can
reopen onto. The fields the router does not read share a slot on each boot, so
one of them truthy while another is falsy is reached only where the falsy slot
pairs a field that cannot be falsy with one that is. Within that boundary the
verdict is unconditional — if one of those boots leaves the popup unhealthy,
`make check` fails, including when it takes two corrupted fields at once,
because the verdict is the combined boot and the per-field re-boot that names a
culprit can only decorate the message. The combined boot must also land on the
view it stored, and each value driven only onto the restore path must land on
its view, or fall back to Home, as its row declares; a field the router reads
can legitimately change which view renders, so its own sweep is held to health
alone. `make check` also fails on a field that gains a floor while its row still
claims it has none, and on a field added to `PERSISTED_FIELDS` with no row at
all. The per-field justification that used to live in the header of
`src/shared/stateSchema.js` shipped a false claim in three consecutive changes,
each caught only by a reviewer re-deriving thirty fields by hand.
@@ -1283,11 +1318,14 @@ behind a "···" menu.
Navigation uses a stack model (like iOS): each forward action pushes the current
screen onto `state.viewStack`, and "Back" pops it (`pushCurrentView()` and
`goBack()` in `src/popup/views/helpers.js`). The root screen is either Welcome
(no wallets) or Home (has wallets). Each screen below gives its view id in
parentheses; the registry of view ids is the `VIEWS` array in
`src/popup/views/helpers.js`, and the markup for a screen is the element with id
`view-` plus that view id in `src/popup/index.html`.
`goBack()` in `src/popup/views/helpers.js`). "Back" skips an entry for the
screen already showing: ShowRecoveryPhrase and the two delete screens take
themselves off the stack when left, so the Settings gear on one of them leaves
Settings under Settings, and "Back" from there goes to the screen before
Settings. The root screen is either Welcome (no wallets) or Home (has wallets).
Each screen below gives its view id in parentheses; the registry of view ids is
the `VIEWS` array in `src/popup/views/helpers.js`, and the markup for a screen
is the element with id `view-` plus that view id in `src/popup/index.html`.
Three elements sit outside the screens and are present on all of them: the title
bar ("AutistMask by @sneak" plus the Settings gear), the flash message line
@@ -1413,14 +1451,17 @@ view would leave a wallet one click from deletion.
without it, the lost-password route on DeleteWallet is the first they
would hear of it. The hint line reserves its height, so switching tabs
cannot move the password fields under the pointer.
- "Import" button
- "Import" button, with the error line beside it so that it adds no height
to a screen whose button already starts near the bottom of the popup
- **Transitions**:
- "Import" with a valid entry and a matching password of at least 12
characters → creates the wallet, clears the navigation stack, and →
**Home**. The phrase and xprv modes then scan for further used addresses
and report the count as a flash message.
- "Import" with an invalid entry, a duplicate wallet or address, or a short
or mismatched password → flash message, no screen change
- "Import" with a missing, short or mismatched password → full-sentence
error on the error line, no screen change
- "Import" with an invalid entry or a duplicate wallet or address → flash
message, no screen change
- "Back" → previous screen (Welcome, Home, or Settings)
#### AddressDetail (`address`)
@@ -1459,8 +1500,8 @@ view would leave a wallet one click from deletion.
copy)
- Warning that anyone holding the private key can transfer all funds from
the address
- Error line
- Password input and "Reveal" button, shown until the key is revealed
- Password input, error line and "Reveal" button, shown until the key is
revealed
- The private key on a highlighted background, tap to copy, shown only after
the password has been accepted
- **Transitions**:
@@ -1479,6 +1520,14 @@ view would leave a wallet one click from deletion.
route, including the Settings gear. A decrypt still running when the screen is
left is discarded rather than written. The screen is not restorable, so
reopening the popup lands on Home rather than back on the key.
- **Clipboard**: tapping the key copies it to the clipboard, and the wallet
never clears the clipboard afterwards; leaving the screen wipes the key from
the page only. The clipboard is the user's, not the wallet's. Clearing it
would go against what the user expects, and by then they may have copied
something else vital that the clear would destroy. The user knows the key is
secret from the warning above the password input, which says that anyone with
it can access and transfer all funds from the address, and knows it is on the
clipboard because they copied it. From then on it is theirs to manage.
#### AddressToken (`address-token`)
@@ -1790,8 +1839,8 @@ view would leave a wallet one click from deletion.
- Wallet name
- Warning box stating that anyone holding these words can take everything in
the wallet, from any device, without the password
- Error line
- Password input + "Reveal" button, shown until the password is accepted
- Password input, error line and "Reveal" button, shown until the password
is accepted
- The recovery phrase itself, in full and click-to-copy, shown only after a
correct password and in place of the password prompt
- **Transitions**:
@@ -1809,6 +1858,9 @@ view would leave a wallet one click from deletion.
gear. A decrypt still running when the screen is left is discarded rather than
written. The screen is not restorable, so reopening the popup lands on Home
rather than back on the phrase.
- **Clipboard**: tapping the phrase copies it, and the wallet never clears the
clipboard afterwards, for the reasons given under ExportPrivKey; here the
warning box above the password input is what tells the user it is secret.
#### DeleteWallet (`delete-wallet-confirm`)
@@ -1817,8 +1869,8 @@ view would leave a wallet one click from deletion.
- "Back" button, "Delete Wallet" heading
- Warning naming the wallet and stating that deletion is permanent and any
funds are unrecoverable without the recovery phrase
- Error line
- Password input
- Error line
- "Confirm Delete" button
- An underlined "I have lost my password" control
- **Transitions**:
@@ -2596,7 +2648,7 @@ Currently supported:
### Non-Goals for 1.0
- Multi-chain support (Ethereum mainnet only)
- Chains other than Ethereum mainnet and the Sepolia testnet
- Hardware wallet support
## TODO
@@ -2630,7 +2682,10 @@ Currently supported:
## Policies
- We don't mention "the other wallet" by name in code or documentation. We're
our own thing.
our own thing. Written exception: the injected provider in
`src/content/inpage.js` sets the flag named after the other wallet to `true`.
It is an interface-compatibility flag many dApps check, and without it the
wallet stops working on their sites.
- The README is the complete authoritative technical documentation. It's ok if
it gets big.
+4 -1
View File
@@ -118,4 +118,7 @@ contradicts either, the originals govern.
- [ ] "Address" not "account" or "derived key"
- [ ] "Password" not "encryption key" or "vault passphrase"
- [ ] Error messages are full sentences
- [ ] No competitor mentioned by name in code or documentation
- [ ] No competitor mentioned by name in code or documentation. Written
exception: the injected provider in `src/content/inpage.js` sets the flag
named after the other wallet to `true`, an interface-compatibility flag
many dApps check (README.md, Policies)
+132 -14
View File
@@ -23,28 +23,149 @@
pre-1.0, working towards the 1.0.0 milestone. Tagged v0.1.0 on 2026-02-27. The
milestone is in flight on `next`; its `next` -> `main` PR is
[#190](https://git.eeqj.de/sneak/AutistMask/pulls/190). `make check` verified
green on `next` at `e9fa8be` on 2026-08-10, and `make build` produces
`dist/chrome/` and `dist/firefox/`, verified against the build's own receipt to
hold exactly the regular files and symlinks that build emitted, with `DEBUG`
compiled off.
[#388](https://git.eeqj.de/sneak/AutistMask/pulls/388). `make build` produces
`dist/chrome/` and `dist/firefox/` with `DEBUG` compiled off, and checks them
against the build's own receipt to hold exactly the regular files and symlinks
that build emitted.
The backlog lives on the
[Gitea tracker](https://git.eeqj.de/sneak/AutistMask/issues), which is
authoritative; this file does not duplicate it. Full policy file set present.
Real-browser end-to-end suites (`make test-e2e` for Chrome,
`make test-e2e-firefox` for Firefox) sit alongside `make check`, which now does
static analysis as well as formatting, and `.gitea/workflows/e2e.yml` runs both
of them on every push.
`make test-e2e-firefox` for Firefox) sit alongside `make check`, which runs the
tests, static analysis and the formatting check, and `.gitea/workflows/e2e.yml`
runs both of them on every push.
# Next Step
Pre-1.0 security review of the extension (key handling, DEBUG mode policy, RPC
input validation) before any 1.0rc tag. Individual filed issues are parts of it,
but the review is broader than any of them.
Cut 1.0.0 once the
[1.0.0 milestone](https://git.eeqj.de/sneak/AutistMask/milestone/6) is empty,
then continue tagging as milestones land.
# Completed Steps
- 2026-10-07: Two contradictions between the documents and the code are resolved
as ruled on [#165](https://git.eeqj.de/sneak/AutistMask/issues/165). The
README's 1.0 non-goal now puts Ethereum mainnet and the Sepolia testnet in
scope and other chains out of it. The injected provider's flag named after the
other wallet stays, as an interface-compatibility flag many dApps check, and
is written down as an exception to the rule against naming competitors in the
README's Policies section, in `RULES.md` and in a comment beside it in
`src/content/inpage.js`. `script/check-censored` already allows that flag only
in that file and its built copies, so it is unchanged.
- 2026-10-07: Two holes in what `tests/persistedFieldContract.test.js` checks
are closed ([#379](https://git.eeqj.de/sneak/AutistMask/issues/379)). The
check that every swept field is driven both truthy and falsy counts only
`hostile` and `falsy` values, which the sweep drives onto every restorable
view, and no longer a `hostileRestore` value, which reaches only the views its
entry names; the stale index 5 moves into `hostile` for `selectedWallet` and
`selectedAddress`, as the one value of either still truthy after the floor.
Each `hostileRestore` entry declares whether its boot lands on its view or
falls back to Home, and is held to it. The limits that remain, fields that
share a slot on one boot and anything no stored record reaches by itself, are
restated as built in the file's header and the README, which now also say
which boots are held to where the popup lands and that a healthy profile is
booted onto every restorable view.
- 2026-10-07: Every password error in the popup is shown the same way
([#493](https://git.eeqj.de/sneak/AutistMask/issues/493)): with `showError()`
and `hideError()` in a fixed-height error line below the password field, as
the send confirmation and approval screens already did. The add wallet screen
showed a missing, short or mismatched password in the flash line, and the
private key export, recovery phrase and delete wallet screens each had a line
of their own above the field. On the add wallet screen the line sits beside
the Import button, so the button stays where it was at 360x600. Each line
clears when the screen is shown again and when the password is tried again.
The add wallet screen's other messages, such as an invalid recovery phrase, a
duplicate wallet and the address scan, stay in the flash line.
`tests/passwordErrorLines.test.js` drives all four screens in the popup.
- 2026-10-07: Pre-1.0 security review of the extension
([#383](https://git.eeqj.de/sneak/AutistMask/issues/383)), reading the tree at
`99292b9` for key handling, the DEBUG mode policy, and what the background
accepts from pages, the configured RPC endpoint and the explorer, with what
the approval screens show from it; the site permission model and storage were
read as well. Its summary on that issue lists ten findings, each filed as its
own issue, and all ten are fixed on `next`; one,
[#399](https://git.eeqj.de/sneak/AutistMask/issues/399), put funds at risk.
Three decisions it raised are still open with the owner: the Argon2id cost for
the vault key ([#401](https://git.eeqj.de/sneak/AutistMask/issues/401)), a
connected site switching the network with no prompt
([#408](https://git.eeqj.de/sneak/AutistMask/issues/408)), and `eth_sign`
signing as a personal message
([#409](https://git.eeqj.de/sneak/AutistMask/issues/409)). Not covered: the
end-to-end suites were not run, the bundled phishing blocklist and token list
were not checked entry by entry, `ethers` and `libsodium-wrappers-sumo` were
taken as audited, and nothing was tried against a real network with real funds
([#385](https://git.eeqj.de/sneak/AutistMask/issues/385)). The planned
independent second check of each finding did not run; the findings rest on the
reviewer's own reading of the code.
- 2026-10-07: Stale branches pruned from `origin`
([#167](https://git.eeqj.de/sneak/AutistMask/issues/167)). The issue
classifies each branch it lists, with the evidence. The eighteen still on
`origin` that it classifies as landed, or superseded by merged pull requests,
were deleted. `feat/message-signing` and `fix/59-transaction-view-ui-policies`
had been deleted on 2026-09-09; both landed and stay deleted. Four are kept
because their work is not in `next`: `fix/consistent-error-display`,
`fix/87-consistent-error-display` and `fix/87-consistent-error-display-v2`,
the change for [#87](https://git.eeqj.de/sneak/AutistMask/issues/87) that
never landed, as reference for
[#493](https://git.eeqj.de/sneak/AutistMask/issues/493); and
`chore/token-list-enrichment`, deleted on 2026-09-09 and re-created at
`f7a2437`, whose `scripts/` tooling waits on
[#495](https://git.eeqj.de/sneak/AutistMask/issues/495).
- 2026-10-07: The README says that the wallet never clears the clipboard after
the private key or the recovery phrase is copied, and why
([#492](https://git.eeqj.de/sneak/AutistMask/issues/492)): the clipboard is
the user's, clearing it would go against what they expect, and it could
destroy something else they copied since. It is under ExportPrivKey, with a
line under ShowRecoveryPhrase, and names the warning each password screen
actually shows, which says nothing about the clipboard.
- 2026-10-07: The Firefox end-to-end suite no longer tolerates any uncaught
extension error ([#487](https://git.eeqj.de/sneak/AutistMask/issues/487)). Its
one entry, Firefox reporting a popup promise that settled after the page
unloaded, had lost its cause with
[#275](https://git.eeqj.de/sneak/AutistMask/issues/275); the entry and the
code that printed tolerated errors are gone from `tests/e2e/firefox/run.js`,
and so is the README paragraph that described it.
- 2026-10-07: The toolbar icons in `icons/` can be redrawn in the tree
([#378](https://git.eeqj.de/sneak/AutistMask/issues/378)): `make icons` runs
`script/lib/icons.js`, which draws the mark and writes each PNG with node's
own `zlib`, no new dependency, leaving alone a file that already holds the
drawn image. `tests/icons.test.js` requires each committed file to hold
exactly that image, the same IHDR and every pixel. The compressed bytes are
not compared, because node's bundled zlib does not compress them as the stock
zlib that made the committed files did. `build.js` now copies each manifest
from the same path `copyIcons()` reads its icons from.
- 2026-10-07: The README's end-to-end limits now match what the two browser
suites do to the extension
([#293](https://git.eeqj.de/sneak/AutistMask/issues/293)). The `window.close`
override the issue named went with
[#275](https://git.eeqj.de/sneak/AutistMask/issues/275), so it needs no entry.
Added: both suites load the popup in a tab rather than from the toolbar;
Chrome waits 1.5 seconds before each site-connection request, records what
approval windows send and click, grants clipboard permission, writes text
straight into the page in two layout tests, and forces the leave during a
decrypt; Firefox forces the site-connection prompt into a window. Corrected:
Firefox's one tolerated error, whose cause that fix removed (the entry goes in
[#487](https://git.eeqj.de/sneak/AutistMask/issues/487)), the phishing
blocklist the extension no longer fetches, and two stale figures.
- 2026-10-07: Back from Settings no longer shows Settings again after the
settings gear was pressed on the recovery phrase or a delete wallet screen
opened from Settings, with or without a reopen in between
([#481](https://git.eeqj.de/sneak/AutistMask/issues/481)). Those screens take
themselves off the Back stack when left, which leaves Settings under Settings;
`goBack()` now skips an entry for the screen already showing.
`tests/showPhrase.test.js`, `tests/deleteWalletLostPassword.test.js` and
`tests/backNavigation.test.js` drive each path.
- 2026-10-07: `script/discard-dist-on-failure` returns the failed step's own
exit status even when it cannot write its message, to a closed stderr or to a
pipe nobody reads any more
@@ -1823,6 +1944,3 @@ but the review is broader than any of them.
Only work that has no issue of its own belongs here; everything else is on the
tracker.
- Cut 1.0.0 once the milestone is empty, then continue tagging as milestones
land.
+4 -9
View File
@@ -588,15 +588,10 @@ async function build() {
copyIcons(distDir);
}
// copy manifests
copyEmitted(
path.join(__dirname, "manifest", "chrome.json"),
path.join(DIST_CHROME, "manifest.json"),
);
copyEmitted(
path.join(__dirname, "manifest", "firefox.json"),
path.join(DIST_FIREFOX, "manifest.json"),
);
// copy manifests, the same files copyIcons() read
for (const [distDir, manifestPath] of MANIFEST_SOURCES) {
copyEmitted(manifestPath, path.join(distDir, "manifest.json"));
}
assertForbiddenTableCovered(forbiddenRecord);
+203
View File
@@ -0,0 +1,203 @@
// Draws the toolbar icon and writes it to every file manifest/chrome.json
// declares under "icons". Run by `make icons`; tests/icons.test.js checks that
// the committed files hold exactly the image this draws.
//
// The icon is a dark-navy rounded square carrying a teal triangle with a
// smaller triangle cut out of it. Every coordinate below is a fraction of the
// icon's side, so one drawing serves every size. A pixel's colour is the
// average of an 8x8 grid of samples, one at the centre of each cell, which
// smooths the edges.
//
// The PNG is written here rather than by a library: RGBA at 8 bits per
// channel, filter type 0 (none) on every row, one IDAT chunk compressed by
// node's zlib at level 9. The committed files were compressed by stock zlib,
// which node's bundled zlib does not reproduce byte for byte, so the image is
// compared rather than the file: the IHDR and every pixel.
"use strict";
const fs = require("fs");
const path = require("path");
const zlib = require("zlib");
const { icons } = require("../../manifest/chrome.json");
const NAVY = [0x10, 0x1a, 0x2e];
const TEAL = [0x35, 0xe0, 0xc2];
const CORNER_RADIUS = 0.22;
const OUTER_TRIANGLE = [
[0.5, 0.15],
[0.115, 0.855],
[0.885, 0.855],
];
const INNER_TRIANGLE = [
[0.5, 0.4],
[0.29, 0.7],
[0.71, 0.7],
];
const SAMPLES_PER_SIDE = 8;
const PNG_SIGNATURE = Buffer.from([
0x89, 0x50, 0x4e, 0x47, 0x0d, 0x0a, 0x1a, 0x0a,
]);
// Points are in pixels from the icon's top-left corner.
function insideRoundedSquare(x, y, size) {
const r = CORNER_RADIUS * size;
// How far the point is past the start of a corner's curve, on each axis.
const dx = Math.max(r - x, 0, x - (size - r));
const dy = Math.max(r - y, 0, y - (size - r));
return dx * dx + dy * dy <= r * r;
}
// Inside or on an edge: the point is not on opposite sides of two edges.
function insideTriangle(x, y, [a, b, c]) {
const side = (p, q) =>
(q[0] - p[0]) * (y - p[1]) - (q[1] - p[1]) * (x - p[0]);
const sides = [side(a, b), side(b, c), side(c, a)];
return !(sides.some((s) => s < 0) && sides.some((s) => s > 0));
}
// Rounds to the nearest integer and a half to the even neighbour, as the
// committed icons were made. Math.round takes a half up, which would change
// some pixels by one.
function roundHalfToEven(v) {
const down = Math.floor(v);
if (v - down !== 0.5) {
return Math.round(v);
}
return down % 2 === 0 ? down : down + 1;
}
// The RGBA bytes of the pixel whose top-left corner is (px, py).
function pixel(px, py, size, outer, inner) {
let inSquare = 0;
let inTeal = 0;
for (let j = 0; j < SAMPLES_PER_SIDE; j++) {
const y = py + (j + 0.5) / SAMPLES_PER_SIDE;
for (let i = 0; i < SAMPLES_PER_SIDE; i++) {
const x = px + (i + 0.5) / SAMPLES_PER_SIDE;
if (!insideRoundedSquare(x, y, size)) {
continue;
}
inSquare++;
if (insideTriangle(x, y, outer) && !insideTriangle(x, y, inner)) {
inTeal++;
}
}
}
if (inSquare === 0) {
return [0, 0, 0, 0];
}
const colour = NAVY.map((navy, k) =>
roundHalfToEven(navy + ((TEAL[k] - navy) * inTeal) / inSquare),
);
const alpha = roundHalfToEven((255 * inSquare) / SAMPLES_PER_SIDE ** 2);
return [...colour, alpha];
}
// A PNG chunk: the data's length, the type, the data, then the CRC-32 of the
// type and the data.
function chunk(type, data) {
const typeAndData = Buffer.concat([Buffer.from(type, "ascii"), data]);
const length = Buffer.alloc(4);
length.writeUInt32BE(data.length);
const crc = Buffer.alloc(4);
crc.writeUInt32BE(zlib.crc32(typeAndData));
return Buffer.concat([length, typeAndData, crc]);
}
// The complete PNG file for the icon at `size` pixels square.
function drawIcon(size) {
const toPixels = (corners) => corners.map(([x, y]) => [x * size, y * size]);
const outer = toPixels(OUTER_TRIANGLE);
const inner = toPixels(INNER_TRIANGLE);
const rows = [];
for (let py = 0; py < size; py++) {
const row = [0]; // filter type 0: the row's bytes are stored as they are
for (let px = 0; px < size; px++) {
row.push(...pixel(px, py, size, outer, inner));
}
rows.push(Buffer.from(row));
}
const header = Buffer.alloc(13); // compression, filter, interlace: all 0
header.writeUInt32BE(size, 0); // width
header.writeUInt32BE(size, 4); // height
header[8] = 8; // bits per channel
header[9] = 6; // colour type: RGBA
return Buffer.concat([
PNG_SIGNATURE,
chunk("IHDR", header),
chunk("IDAT", zlib.deflateSync(Buffer.concat(rows), { level: 9 })),
chunk("IEND", Buffer.alloc(0)),
]);
}
// The image in a PNG: its IHDR data, and its rows after decompression, each
// row's filter type byte first. With filter type 0 on every row, as drawIcon
// writes, the rows are the pixels themselves; a file using another filter does
// not match even where its pixels do. Refuses a file that is not a PNG, a chunk
// whose CRC-32 is wrong, and any chunk but IHDR, IDAT and IEND, rather than
// ignoring what it cannot compare.
function decodePng(png) {
if (!png.subarray(0, 8).equals(PNG_SIGNATURE)) {
throw new Error("not a PNG");
}
let header = null;
const idat = [];
for (let pos = 8; pos < png.length; ) {
const length = png.readUInt32BE(pos);
const typeAndData = png.subarray(pos + 4, pos + 8 + length);
const type = typeAndData.toString("ascii", 0, 4);
if (zlib.crc32(typeAndData) !== png.readUInt32BE(pos + 8 + length)) {
throw new Error(`the ${type} chunk fails its CRC-32`);
}
if (type === "IHDR") {
header = typeAndData.subarray(4);
} else if (type === "IDAT") {
idat.push(typeAndData.subarray(4));
} else if (type !== "IEND") {
throw new Error(`unexpected ${type} chunk`);
}
pos += 12 + length;
}
if (header === null) {
throw new Error("no IHDR chunk");
}
return { header, rows: zlib.inflateSync(Buffer.concat(idat)) };
}
// True when `file` already holds the image in `png`. A file that is missing or
// cannot be read as a PNG does not, and is written over.
function holdsSameImage(file, png) {
let existing;
try {
existing = decodePng(fs.readFileSync(file));
} catch {
return false;
}
const drawn = decodePng(png);
return (
existing.header.equals(drawn.header) && existing.rows.equals(drawn.rows)
);
}
// A file already holding the drawn image is left alone, so running this does
// not rewrite the committed icons with node's compression.
if (require.main === module) {
for (const [size, file] of Object.entries(icons)) {
const target = path.join(__dirname, "..", "..", file);
const png = drawIcon(Number(size));
if (holdsSameImage(target, png)) {
console.log(`icons: ${file} already holds this image`);
continue;
}
fs.writeFileSync(target, png);
console.log(`icons: wrote ${file}`);
}
}
module.exports = { drawIcon, decodePng };
+4 -1
View File
@@ -99,7 +99,10 @@
const provider = {
isAutistMask: true,
isMetaMask: true, // compatibility — many dApps check this
// An interface-compatibility flag many dApps check. Kept as a written
// exception to the rule that no competitor is named in code: see
// Policies in README.md, and RULES.md.
isMetaMask: true,
chainId: currentChainId,
networkVersion: currentNetworkVersion,
selectedAddress: null,
+24 -14
View File
@@ -207,12 +207,22 @@
class="border border-border p-1 w-full font-mono text-sm bg-bg text-fg"
/>
</div>
<!-- The error line sits beside Import, not above it: at
360x600 the button already starts near the bottom of
the popup, and a line of its own would push it below
the fold. The longest error fits on one line here. -->
<div class="flex items-center gap-2">
<button
id="btn-add-wallet-confirm"
class="border border-border px-2 py-1 hover:bg-fg hover:text-bg cursor-pointer"
>
Import
</button>
<div
id="add-wallet-password-error"
class="text-xs min-h-[1.25rem] invisible"
></div>
</div>
</div>
<!-- ============ MAIN VIEW: ALL WALLETS & ADDRESSES ============ -->
@@ -396,10 +406,6 @@
Warning: anyone with this private key can access and
transfer all funds from this address. Never share it.
</p>
<div
id="export-privkey-flash"
class="text-xs mb-2 min-h-[1.25rem] invisible"
></div>
<div id="export-privkey-password-section" class="mb-2">
<label class="block mb-1">Password</label>
<input
@@ -408,9 +414,13 @@
class="border border-border p-1 w-full font-mono text-sm bg-bg text-fg"
placeholder="Enter your password to continue"
/>
<div
id="export-privkey-password-error"
class="text-xs mt-2 mb-2 min-h-[1.25rem] invisible"
></div>
<button
id="btn-export-privkey-confirm"
class="border border-border px-2 py-1 hover:bg-fg hover:text-bg cursor-pointer mt-2"
class="border border-border px-2 py-1 hover:bg-fg hover:text-bg cursor-pointer"
>
Reveal
</button>
@@ -1116,10 +1126,6 @@
<strong id="delete-wallet-name"></strong> is permanent. Any
funds will be unrecoverable without your recovery phrase.
</p>
<div
id="delete-wallet-flash"
class="text-xs text-red-500 mb-2 min-h-[1.25rem] invisible"
></div>
<div class="mb-2">
<label class="block mb-1">Password</label>
<input
@@ -1129,6 +1135,10 @@
placeholder="Enter your password to confirm"
/>
</div>
<div
id="delete-wallet-password-error"
class="text-xs mb-2 min-h-[1.25rem] invisible"
></div>
<button
id="btn-delete-wallet-confirm"
class="border border-border text-red-500 px-2 py-1 hover:bg-fg hover:text-bg cursor-pointer"
@@ -1273,10 +1283,6 @@
this wallet, from any device, without your password. Never
type them into a website and never show them to anyone.
</div>
<div
id="show-phrase-flash"
class="text-xs text-red-500 mb-2 min-h-[1.25rem] invisible"
></div>
<div id="show-phrase-password-section" class="mb-2">
<label class="block mb-1">Password</label>
<input
@@ -1285,9 +1291,13 @@
class="border border-border p-1 w-full font-mono text-sm bg-bg text-fg"
placeholder="Enter your password to continue"
/>
<div
id="show-phrase-password-error"
class="text-xs mt-2 mb-2 min-h-[1.25rem] invisible"
></div>
<button
id="btn-show-phrase-reveal"
class="border border-border px-2 py-1 hover:bg-fg hover:text-bg cursor-pointer mt-2"
class="border border-border px-2 py-1 hover:bg-fg hover:text-bg cursor-pointer"
>
Reveal
</button>
+12 -4
View File
@@ -2,6 +2,8 @@ const {
$,
showView,
showFlash,
showError,
hideError,
goBack,
clearViewStack,
onViewLeave,
@@ -98,6 +100,7 @@ function clear() {
$("add-wallet-password").value = "";
$("add-wallet-password-confirm").value = "";
$("add-wallet-phrase-warning").style.visibility = "hidden";
hideError("add-wallet-password-error");
}
// Each wallet has its own password (its own encryptedSecret), so adding a
@@ -125,15 +128,18 @@ function validatePassword() {
const pw = $("add-wallet-password").value;
const pw2 = $("add-wallet-password-confirm").value;
if (!pw) {
showFlash("Please choose a password.");
showError("add-wallet-password-error", "Please choose a password.");
return null;
}
if (pw.length < 12) {
showFlash("Password must be at least 12 characters.");
showError(
"add-wallet-password-error",
"Password must be at least 12 characters.",
);
return null;
}
if (pw !== pw2) {
showFlash("Passwords do not match.");
showError("add-wallet-password-error", "Passwords do not match.");
return null;
}
return pw;
@@ -342,8 +348,10 @@ function init(ctx) {
$("add-wallet-phrase-warning").style.visibility = "visible";
});
// Import / confirm
// Import / confirm. Each press starts with no password error on screen:
// validatePassword() puts it back if the password is still wrong.
$("btn-add-wallet-confirm").addEventListener("click", async () => {
hideError("add-wallet-password-error");
if (currentMode === "mnemonic") {
await importMnemonic(ctx);
} else if (currentMode === "privkey") {
+16 -11
View File
@@ -2,6 +2,8 @@ const {
$,
showView,
showFlash,
showError,
hideError,
goBack,
clearViewStack,
onViewLeave,
@@ -55,8 +57,7 @@ function confirmKey(name) {
function clear() {
deleteWalletIndex = null;
$("delete-wallet-password").value = "";
$("delete-wallet-flash").textContent = "";
$("delete-wallet-flash").style.visibility = "hidden";
hideError("delete-wallet-password-error");
}
// The lost-password screen holds no secret — a wallet name is not one —
@@ -232,19 +233,22 @@ function init(_ctx) {
$("btn-delete-wallet-confirm").addEventListener("click", async () => {
const pw = $("delete-wallet-password").value;
if (!pw) {
$("delete-wallet-flash").textContent =
"Please enter your password.";
$("delete-wallet-flash").style.visibility = "visible";
showError(
"delete-wallet-password-error",
"Please enter your password.",
);
return;
}
if (deleteWalletIndex === null) {
$("delete-wallet-flash").textContent =
"No wallet selected for deletion.";
$("delete-wallet-flash").style.visibility = "visible";
showError(
"delete-wallet-password-error",
"No wallet selected for deletion.",
);
return;
}
hideError("delete-wallet-password-error");
const btn = $("btn-delete-wallet-confirm");
btn.disabled = true;
btn.classList.add("text-muted");
@@ -256,9 +260,10 @@ function init(_ctx) {
try {
await decryptWithPassword(wallet.encryptedSecret, pw);
} catch {
$("delete-wallet-flash").textContent =
"That password is incorrect. Please try again.";
$("delete-wallet-flash").style.visibility = "visible";
showError(
"delete-wallet-password-error",
"That password is incorrect. Please try again.",
);
btn.disabled = false;
btn.classList.remove("text-muted");
return;
+13 -12
View File
@@ -20,6 +20,8 @@ const {
$,
showView,
showFlash,
showError,
hideError,
flashCopyFeedback,
goBack,
onViewLeave,
@@ -56,11 +58,6 @@ function isCurrentReveal(generation) {
);
}
function fail(message) {
$("export-privkey-flash").textContent = message;
$("export-privkey-flash").style.visibility = "visible";
}
// Wipe every trace of the key and drop the address selection. Safe to call
// when nothing was ever revealed, and safe to call twice.
function clear() {
@@ -71,8 +68,7 @@ function clear() {
$("export-privkey-password").value = "";
$("export-privkey-result").classList.add("hidden");
$("export-privkey-password-section").classList.remove("hidden");
$("export-privkey-flash").textContent = "";
$("export-privkey-flash").style.visibility = "hidden";
hideError("export-privkey-password-error");
}
function show(walletIdx, addrIdx) {
@@ -112,15 +108,19 @@ function show(walletIdx, addrIdx) {
async function reveal() {
const password = $("export-privkey-password").value;
if (!password) {
fail("Please enter your password.");
showError(
"export-privkey-password-error",
"Please enter your password.",
);
return;
}
if (walletIndex === null) {
fail("No address is selected.");
showError("export-privkey-password-error", "No address is selected.");
return;
}
const wallet = state.wallets[walletIndex];
hideError("export-privkey-password-error");
const btn = $("btn-export-privkey-confirm");
btn.disabled = true;
btn.classList.add("text-muted");
@@ -140,11 +140,12 @@ async function reveal() {
$("export-privkey-password-section").classList.add("hidden");
$("export-privkey-value").textContent = signer.privateKey;
$("export-privkey-result").classList.remove("hidden");
$("export-privkey-flash").textContent = "";
$("export-privkey-flash").style.visibility = "hidden";
} catch {
if (!isCurrentReveal(generation)) return;
fail("That password is incorrect. Please try again.");
showError(
"export-privkey-password-error",
"That password is incorrect. Please try again.",
);
} finally {
btn.disabled = false;
btn.classList.remove("text-muted");
+13 -2
View File
@@ -216,10 +216,21 @@ function pushCurrentView() {
// Pop the navigation stack and show the previous view. If the stack
// is empty, fall back to the main (home) view.
//
// An entry for the view already showing is skipped: landing on it would
// make Back seem to do nothing. Settings, the recovery phrase or delete
// wallet screen, then the gear leaves Settings under Settings, because that
// screen takes itself off the stack when left, and a reopened popup cuts it
// off the restored stack the same way
// (https://git.eeqj.de/sneak/AutistMask/issues/481).
function goBack() {
const stack = state.viewStack;
while (stack.length > 0 && stack[stack.length - 1] === state.currentView) {
stack.pop();
}
let target;
if (state.viewStack.length > 0) {
target = state.viewStack.pop();
if (stack.length > 0) {
target = stack.pop();
} else {
target = "main";
}
+14 -13
View File
@@ -21,6 +21,8 @@ const {
$,
showView,
showFlash,
showError,
hideError,
flashCopyFeedback,
goBack,
onViewLeave,
@@ -52,11 +54,6 @@ function isCurrentReveal(generation) {
);
}
function fail(message) {
$("show-phrase-flash").textContent = message;
$("show-phrase-flash").style.visibility = "visible";
}
// Wipe every trace of the phrase and drop the wallet selection. Safe to
// call when nothing was ever revealed, and safe to call twice.
function clear() {
@@ -66,8 +63,7 @@ function clear() {
$("show-phrase-password").value = "";
$("show-phrase-result").classList.add("hidden");
$("show-phrase-password-section").classList.remove("hidden");
$("show-phrase-flash").textContent = "";
$("show-phrase-flash").style.visibility = "hidden";
hideError("show-phrase-password-error");
}
function show(walletIdx) {
@@ -90,19 +86,23 @@ function show(walletIdx) {
async function reveal() {
const password = $("show-phrase-password").value;
if (!password) {
fail("Please enter your password.");
showError("show-phrase-password-error", "Please enter your password.");
return;
}
if (walletIndex === null) {
fail("No wallet is selected.");
showError("show-phrase-password-error", "No wallet is selected.");
return;
}
const wallet = state.wallets[walletIndex];
if (!walletHasRecoveryPhrase(wallet)) {
fail("This wallet does not have a recovery phrase.");
showError(
"show-phrase-password-error",
"This wallet does not have a recovery phrase.",
);
return;
}
hideError("show-phrase-password-error");
const btn = $("btn-show-phrase-reveal");
btn.disabled = true;
btn.classList.add("text-muted");
@@ -120,13 +120,14 @@ async function reveal() {
$("show-phrase-password-section").classList.add("hidden");
$("show-phrase-value").textContent = phrase;
$("show-phrase-result").classList.remove("hidden");
$("show-phrase-flash").textContent = "";
$("show-phrase-flash").style.visibility = "hidden";
} catch {
if (!isCurrentReveal(generation)) return;
// Deliberately not the caught error: the message is fixed so that
// nothing derived from the ciphertext or the attempt can surface.
fail("That password is incorrect. Please try again.");
showError(
"show-phrase-password-error",
"That password is incorrect. Please try again.",
);
} finally {
btn.disabled = false;
btn.classList.remove("text-muted");
+2
View File
@@ -39,6 +39,8 @@ jest.doMock("../src/popup/views/helpers", () => ({
$: element,
showView: () => {},
showFlash: () => {},
showError: () => {},
hideError: () => {},
goBack: () => {},
clearViewStack: () => {},
onViewLeave: () => {},
+18
View File
@@ -52,6 +52,7 @@ const {
resetRenderedViews,
} = require("../src/popup/viewRouter");
const { state } = require("../src/shared/state");
const { restorableStack } = require("../src/shared/persistedState");
const ADDRESS = "0x1111111111111111111111111111111111111111";
const TOKEN = "0xa0b86991c6218b36c1d19d4a2e9eb0ce3606eb48";
@@ -209,6 +210,23 @@ describe("Back onto a view the reopened popup never rendered", () => {
});
});
// https://git.eeqj.de/sneak/AutistMask/issues/481. Settings, the recovery
// phrase or delete wallet screen, the gear, then a reopen: the restored stack
// is cut at the screen the gear left, which leaves Settings under the Settings
// the popup reopens onto.
describe("Back from Settings reopened over its own entry", () => {
test.each(["show-phrase", "delete-wallet-confirm"])(
"goes to the screen under it after leaving %s",
(left) => {
const stored = ["main", "settings", left];
reopenedOn("settings", restorableStack(stored, "settings"));
goBack();
expect(calls).toEqual(["main"]);
expect(state.currentView).toBe("main");
},
);
});
// The guards are restoreView()'s, so a popped view whose backing data is
// gone lands on Home rather than on an empty template.
describe("Back onto a view whose backing data is gone", () => {
+8 -6
View File
@@ -253,7 +253,7 @@ describe("reaching the screen", () => {
const { decryptWithPassword } = require("../src/shared/vault");
decryptWithPassword.mockRejectedValue(new Error("nope"));
await click("btn-delete-wallet-confirm");
expect(node("delete-wallet-flash").textContent).toBe(
expect(node("delete-wallet-password-error").textContent).toBe(
"That password is incorrect. Please try again.",
);
});
@@ -618,9 +618,11 @@ describe("the password route's confirm button", () => {
// https://git.eeqj.de/sneak/AutistMask/issues/480: leaving either delete
// screen drops its wallet selection, so Back onto one showed a screen whose
// button could only answer "No wallet selected for deletion."
// button could only answer "No wallet selected for deletion." Taking the
// screen off the stack leaves Settings under Settings, and Back must not land
// there either (https://git.eeqj.de/sneak/AutistMask/issues/481).
describe("Back from Settings after leaving by the settings gear", () => {
test("does not land on the delete screen", () => {
test("goes past the delete screen to the screen under Settings", () => {
const { helpers, deleteWallet, state } = load();
deleteWallet.show(1);
// The settings gear: push the current view, then show Settings.
@@ -629,10 +631,10 @@ describe("Back from Settings after leaving by the settings gear", () => {
expect(state.viewStack).toEqual(["main", "settings"]);
helpers.goBack();
expect(state.currentView).not.toBe("delete-wallet-confirm");
expect(state.currentView).toBe("main");
});
test("does not land on the lost-password screen", async () => {
test("goes past the lost-password screen to the screen under Settings", async () => {
const { helpers, deleteWallet, state } = load();
await openLostPassword(deleteWallet, 1);
// The settings gear: push the current view, then show Settings.
@@ -641,7 +643,7 @@ describe("Back from Settings after leaving by the settings gear", () => {
expect(state.viewStack).toEqual(["main", "settings"]);
helpers.goBack();
expect(state.currentView).not.toBe(VIEW);
expect(state.currentView).toBe("main");
});
// The lost-password screen's own Back is "Back returns to the delete
+1 -52
View File
@@ -854,37 +854,6 @@ step(
// ------------------------------------------------------------- runner
// Uncaught extension errors that are known, tracked and deliberately
// tolerated, in the same spirit as ALLOWED_ERRORS in tests/e2e/harness.js:
// every entry names the issue that will delete it, and every occurrence is
// still printed, so tolerating one is visible in the log rather than silent.
// This is the only concession in an otherwise zero-tolerance policy.
const ALLOWED_ERRORS = [
{
// The site-connection buttons in src/popup/views/approval.js send
// their decision and call window.close() on the next line. Firefox's
// BaseContext.wrapPromise reports, through Cu.reportError, any
// extension-API promise that settles after its context unloaded —
// whether or not the caller attached a handler, so notify()'s catch
// cannot suppress it.
//
// Pre-existing, and not introduced by the promise shim: the send was
// already unawaited, and this suite is merely the first thing to
// drive that window on Firefox. It is the same teardown ordering as
// the issue below, whose fix — making the outcome independent of when
// the popup closes — removes this entry with it.
pattern: /Promise (?:resolved|rejected) after context unloaded/,
source: /\/src\/popup\/index\.js$/,
issue: "https://git.eeqj.de/sneak/AutistMask/issues/275",
},
];
function allowedFor(e) {
return ALLOWED_ERRORS.find(
(a) => a.pattern.test(e.msg) && a.source.test(e.src),
);
}
function formatError(e) {
return (
e.msg + " (" + e.src + ":" + e.line + (e.cat ? ", " + e.cat : "") + ")"
@@ -1001,20 +970,6 @@ async function main() {
installFailure = null;
}
// Tolerated errors are set aside, never dropped: each one is
// printed with the issue that keeps it on the list, so the
// concession stays in the run output.
const tolerated = found.filter((e) => allowedFor(e));
found = found.filter((e) => !allowedFor(e));
for (const e of tolerated) {
console.log(
"# tolerated (" +
allowedFor(e).issue +
"): " +
formatError(e),
);
}
// Any uncaught error from an extension source fails the step
// that provoked it, whether or not its assertions passed.
if (!failure && found.length > 0) {
@@ -1039,13 +994,7 @@ async function main() {
// blamed on any one step, but they are still reported and they
// still fail the run.
await sleep(1000);
const trailingAll = await errors.take();
for (const e of trailingAll.filter((x) => allowedFor(x))) {
console.log(
"# tolerated (" + allowedFor(e).issue + "): " + formatError(e),
);
}
const trailing = trailingAll.filter((e) => !allowedFor(e));
const trailing = await errors.take();
console.log(
"# " +
(steps.length - failed) +
+109 -6
View File
@@ -809,7 +809,7 @@ async function secretScreenState(page, view) {
return page.evaluate(
(v) => ({
value: document.getElementById(v + "-value").textContent,
error: document.getElementById(v + "-flash").textContent,
error: document.getElementById(v + "-password-error").textContent,
html: document.getElementById("view-" + v).innerHTML,
resultHidden: document
.getElementById(v + "-result")
@@ -906,7 +906,8 @@ test("a wrong password reveals nothing (#161)", async (env) => {
await env.page.click("#btn-show-phrase-reveal");
await env.page.waitForFunction(
() =>
document.getElementById("show-phrase-flash").textContent.length > 0,
document.getElementById("show-phrase-password-error").textContent
.length > 0,
null,
{ timeout: 60000 },
);
@@ -1993,13 +1994,14 @@ test("an over-long flash message keeps to one line (#252)", async (env) => {
// Every screen that asks for a password reserves room for one line of error.
// The two on the dApp approval screens also have a border and padding, which
// that reserved height has to cover too.
// that reserved height has to cover too. The add wallet screen's line sits
// beside its button rather than above it, and has its own test below.
const PASSWORD_ERROR_CONTAINERS = [
"approve-tx-error",
"approve-sign-error",
"export-privkey-flash",
"show-phrase-flash",
"delete-wallet-flash",
"export-privkey-password-error",
"show-phrase-password-error",
"delete-wallet-password-error",
"confirm-tx-password-error",
];
@@ -2064,6 +2066,107 @@ test("a password error moves nothing on any screen (#297)", async (env) => {
}
});
// ------------------------------------------ add wallet Import button (#493)
// At 360x600 the add wallet screen's Import button already starts near the
// bottom of the popup, so its password error line sits beside the button
// rather than above it. Measures the button and the line empty and again
// filled with the longest error addWallet.js puts there. Runs in the page.
function measureImportButton() {
const button = document.getElementById("btn-add-wallet-confirm");
const line = document.getElementById("add-wallet-password-error");
const measure = () => {
const b = button.getBoundingClientRect();
const l = line.getBoundingClientRect();
return {
buttonTop: b.top + window.scrollY,
buttonBottom: b.bottom + window.scrollY,
lineTop: l.top + window.scrollY,
lineBottom: l.bottom + window.scrollY,
};
};
const empty = measure();
line.textContent = "Password must be at least 12 characters.";
line.style.visibility = "visible";
const filled = measure();
line.textContent = "";
line.style.visibility = "hidden";
return { empty, filled };
}
test("the add wallet password error leaves Import where it was (#493)", async (env) => {
const page = await openPopup(env.ctx, env.popupUrl);
try {
await page.setViewportSize(POPUP_VIEWPORT);
// Brought up by toggling classes, as in the test above. The note
// addWallet.js shows once a wallet exists is the only part of the
// screen that differs between the first wallet and a later one.
await page.evaluate(() => {
const screen = document.getElementById("view-add-wallet");
for (const view of document.querySelectorAll(".view")) {
view.classList.toggle("hidden", view !== screen);
}
});
for (const walletExists of [false, true]) {
await page.evaluate(
(shown) =>
document
.getElementById("add-wallet-separate-password-note")
.classList.toggle("hidden", !shown),
walletExists,
);
for (const tab of ["tab-mnemonic", "tab-privkey", "tab-xprv"]) {
await page.click("#" + tab);
const { empty, filled } =
await page.evaluate(measureImportButton);
const where =
"#" +
tab +
(walletExists
? " with a wallet already added"
: " for the first wallet");
// Printed pass or fail, as the dust threshold test does.
console.log(
"# add wallet Import top, " +
where +
": " +
empty.buttonTop +
"px",
);
for (const m of [empty, filled]) {
assert(
m.lineTop >= m.buttonTop &&
m.lineBottom <= m.buttonBottom,
"the error line on " +
where +
" does not fit beside Import, so it adds height: " +
JSON.stringify(m),
);
}
assert(
filled.buttonTop === empty.buttonTop,
"Import moved " +
(filled.buttonTop - empty.buttonTop) +
"px when the error appeared on " +
where,
);
if (!walletExists) {
assert(
empty.buttonTop < POPUP_VIEWPORT.height,
"Import starts at " +
empty.buttonTop +
"px on " +
where +
", below the fold",
);
}
}
}
} finally {
await page.close();
}
});
// --------------------------------------------- confirmation screen (#238)
//
// The screen that decides what gets signed. The arithmetic underneath it
+6 -4
View File
@@ -207,7 +207,7 @@ describe("a decrypt still running when the screen is left", () => {
});
// Same hole on the failure path: a wrong-password error written after
// the wipe would restore the flash line on a screen the user has left.
// the wipe would restore the error line on a screen the user has left.
test("never writes the failure message either", async () => {
const { helpers, vault, exportPrivkey } = load();
exportPrivkey.show(0, 0);
@@ -217,8 +217,10 @@ describe("a decrypt still running when the screen is left", () => {
reveal.reject(new Error("decryption failed"));
await reveal.pending;
expect(node("export-privkey-flash").textContent).toBe("");
expect(node("export-privkey-flash").style.visibility).toBe("hidden");
expect(node("export-privkey-password-error").textContent).toBe("");
expect(node("export-privkey-password-error").style.visibility).toBe(
"hidden",
);
});
});
@@ -260,7 +262,7 @@ describe("a reveal that is not interrupted", () => {
await reveal.pending;
expect(node("export-privkey-value").textContent).toBe("");
expect(node("export-privkey-flash").textContent).toBe(
expect(node("export-privkey-password-error").textContent).toBe(
"That password is incorrect. Please try again.",
);
});
+33
View File
@@ -0,0 +1,33 @@
// The toolbar icons in icons/ are drawn by script/lib/icons.js (`make icons`).
// Each committed file must hold exactly the image it draws, the same IHDR and
// every pixel, so the drawing and the files cannot drift apart. The compressed
// bytes are not compared: the committed files were compressed by stock zlib,
// and node's bundled zlib compresses the same pixels differently.
const fs = require("fs");
const path = require("path");
const { icons } = require("../manifest/chrome.json");
const { drawIcon, decodePng } = require("../script/lib/icons");
describe("toolbar icons", () => {
test.each(Object.entries(icons))(
"the %spx icon %s holds the image script/lib/icons.js draws",
(size, file) => {
const side = Number(size);
const committed = decodePng(
fs.readFileSync(path.join(__dirname, "..", file)),
);
const drawn = decodePng(drawIcon(side));
expect(committed.header).toEqual(drawn.header);
// Each row is its filter type byte, then 4 bytes per pixel.
expect(drawn.rows.length).toBe(side * (side * 4 + 1));
expect(committed.rows.length).toBe(drawn.rows.length);
// The offset of the first byte that differs, not a diff of all.
expect(
committed.rows.findIndex((byte, i) => byte !== drawn.rows[i]),
).toBe(-1);
},
);
});
+159
View File
@@ -0,0 +1,159 @@
// Every screen that asks for a password shows a password error the same way:
// through showError() and hideError() in src/popup/views/helpers.js, in a
// fixed-height error line below the password field, and never in the flash
// line at the top of the popup
// (https://git.eeqj.de/sneak/AutistMask/issues/493). These boot the real popup
// over src/popup/index.html, so each error line has to exist in the markup,
// and check that the error appears in it and clears again.
jest.mock("../src/shared/vault", () => ({
decryptWithPassword: jest.fn(),
encryptWithPassword: jest.fn(),
}));
const {
bootPopup,
cleanupPopup,
unversionedValidProfile,
} = require("./support/popupBoot");
const PASSWORD = "correct horse battery staple";
const WRONG_PASSWORD = "That password is incorrect. Please try again.";
afterEach(() => {
cleanupPopup();
});
// The error line as the user sees it.
function errorLine(page, id) {
return {
inMarkup: page.document.authoredIds.has(id),
text: page.text(id),
visibility: page.node(id).style.visibility,
};
}
function shown(text) {
return { inMarkup: true, text, visibility: "visible" };
}
const cleared = { inMarkup: true, text: "", visibility: "hidden" };
describe("the add wallet screen", () => {
const ERROR = "add-wallet-password-error";
// First run: Welcome, "Add wallet", then the die for a valid phrase.
async function openAddWallet() {
const page = await bootPopup(undefined);
await page.click("btn-welcome-add");
await page.click("btn-generate-phrase");
return page;
}
function setPasswords(page, password, confirm) {
page.node("add-wallet-password").value = password;
page.node("add-wallet-password-confirm").value = confirm;
}
test("shows a password problem below the password fields, and clears it on the next press", async () => {
const page = await openAddWallet();
setPasswords(page, "short", "short");
await page.click("btn-add-wallet-confirm");
expect(errorLine(page, ERROR)).toEqual(
shown("Password must be at least 12 characters."),
);
expect(page.text("flash-msg")).toBe("");
// The password is fixed and the phrase emptied: the password error
// goes, and the phrase problem is still reported in the flash line.
setPasswords(page, PASSWORD, PASSWORD);
page.node("wallet-mnemonic").value = "";
await page.click("btn-add-wallet-confirm");
expect(errorLine(page, ERROR)).toEqual(cleared);
expect(page.text("flash-msg")).toBe(
"Enter a recovery phrase, or press the die.",
);
// Leaving clears the flash line and stops its timer, which would
// otherwise fire after this page is gone.
await page.click("btn-add-wallet-back");
});
test("clears the error when the screen is shown again", async () => {
const page = await openAddWallet();
setPasswords(page, PASSWORD, PASSWORD + " typo");
await page.click("btn-add-wallet-confirm");
expect(errorLine(page, ERROR)).toEqual(
shown("Passwords do not match."),
);
await page.click("btn-add-wallet-back");
await page.click("btn-welcome-add");
expect(errorLine(page, ERROR)).toEqual(cleared);
});
});
describe.each([
{
screen: "the private key export screen",
open: () => require("../src/popup/views/exportPrivkey").show(0, 0),
field: "export-privkey-password",
button: "btn-export-privkey-confirm",
error: "export-privkey-password-error",
},
{
screen: "the recovery phrase screen",
open: () => require("../src/popup/views/showPhrase").show(0),
field: "show-phrase-password",
button: "btn-show-phrase-reveal",
error: "show-phrase-password-error",
},
{
screen: "the delete wallet screen",
open: () => require("../src/popup/views/deleteWallet").show(0),
field: "delete-wallet-password",
button: "btn-delete-wallet-confirm",
error: "delete-wallet-password-error",
},
])("$screen", ({ open, field, button, error }) => {
// Opens the screen and enters a password the vault rejects.
async function failedAttempt() {
const page = await bootPopup(unversionedValidProfile());
open();
const { decryptWithPassword } = require("../src/shared/vault");
decryptWithPassword.mockRejectedValue(new Error("wrong password"));
page.node(field).value = "not the password";
await page.click(button);
return { page, decryptWithPassword };
}
test("shows a wrong password below the password field", async () => {
const { page } = await failedAttempt();
expect(errorLine(page, error)).toEqual(shown(WRONG_PASSWORD));
});
test("clears the error while the next password is checked", async () => {
const { page, decryptWithPassword } = await failedAttempt();
let rejectDecrypt;
decryptWithPassword.mockReturnValue(
new Promise((resolve, reject) => {
rejectDecrypt = reject;
}),
);
page.node(field).value = "another guess";
const pressed = page.click(button);
await page.settle();
expect(errorLine(page, error)).toEqual(cleared);
rejectDecrypt(new Error("wrong password"));
await pressed;
expect(errorLine(page, error)).toEqual(shown(WRONG_PASSWORD));
});
test("clears the error when the screen is shown again", async () => {
const { page } = await failedAttempt();
open();
expect(errorLine(page, error)).toEqual(cleared);
});
});
+142 -46
View File
@@ -49,22 +49,37 @@
// every path a stored record takes, and the difference is the whole of what
// this file does not cover:
//
// - Only the values in the table, in the SLOT arrangement below: four value
// combinations per view, not the product of twelve fields. A dereference
// reached only under a pairing no slot produces is not driven at all.
// - Only what a stored record reaches by ITSELF. A view only forward
// navigation opens, and anything behind a click, is not driven.
// - Nothing about the paths a HEALTHY profile takes, which is most of the
// popup. This file is a floor under one defect class, not a proof about
// the renderers.
// - Only the values in the table, in the SLOT arrangement below. On the
// restore path the twelve fields the router does not read are corrupted
// together, every field on the same slot, so a view gets four value
// combinations of them, not their product. A branch entered only when one
// of them is truthy and another falsy is reached only where the falsy
// slot happens to pair a field that cannot be falsy with one that is.
// Each field the router reads is corrupted alone, over an otherwise
// well-formed record.
// - Only what a stored record reaches by ITSELF, as the boot renders it. A
// view only forward navigation opens, anything behind a click, and
// anything behind a timer (bootPopup() records every interval, and this
// file never runs one) is not driven.
// - Of the paths a HEALTHY profile takes, only its boot onto each
// restorable view ("the base profile the sweep corrupts" below). The rest
// of the popup is not covered: this file is a floor under one defect
// class, not a proof about the renderers.
//
// Within that boundary it is unconditional: if one of these boots leaves the
// popup unhealthy or off the view it stored, this file goes red — including
// when it takes two corrupted fields at once, because the verdict is the
// combined boot itself and the per-field re-boot below can only decorate the
// message. That last part is the one thing an earlier version got wrong: it
// asserted on the per-field list, so an observed dead popup that no single
// field reproduced was reported green.
// Within that boundary it is unconditional: if a boot that corrupts a field
// leaves the popup unhealthy, this file goes red — including when it takes
// two corrupted fields at once, because the verdict is the combined boot
// itself and the per-field re-boot below can only decorate the message. That
// last part is the one thing an earlier version got wrong: it asserted on the
// per-field list, so an observed dead popup that no single field reproduced
// was reported green.
//
// Where the popup lands is held for some of those boots and not others. The
// combined boot must land on the view it stored, and each `hostileRestore`
// value must land on its view, or fall back to Home, as its entry declares.
// The boots in "a hostile routing value restoring onto" are held to health
// alone, because a value in a field the router reads legitimately changes
// which view renders; so are the boots onto Home, which store no view.
//
// Booting every field separately at every value would be several hundred boots
// and most of the suite's budget; this is forty-four. Widening it further is
@@ -128,7 +143,11 @@ const sweptValues = (row) => [...row.hostile, ...(row.falsy || [])];
// list short and pointed. `floorOnly` is extra values checked against the
// floor alone, which is pure and free. `hostileRestore` is extra values driven
// through the restore path only, for a value that means nothing until a
// particular branch's gate has let it past.
// particular branch's gate has let it past. Each of its entries names the
// `views` it is driven onto and declares whether the boot lands on them
// (`restored: true`) or falls back to Home (`restored: false`), so a value
// written for one renderer cannot stop reaching it unnoticed. A value driven
// onto every restorable view is written with everyRestorableView() below.
//
// `falsy` is the other POLARITY of a swept field, driven for the same reason.
// It is not a value src/ never writes — for three of these fields it is the
@@ -139,6 +158,23 @@ const sweptValues = (row) => [...row.hostile, ...(row.falsy || [])];
// falsy value stored under that field comes back TRUTHY from the floor, so no
// `!state.x` branch is reachable from a stored record at all.
// The `hostileRestore` entries for a value driven onto every restorable view:
// it falls back to Home on the views listed in `fallsBackOn` and lands on every
// other one, so a view added to RESTORABLE_VIEWS is driven, and expected to
// land, without editing the row.
function everyRestorableView(value, fallsBackOn) {
return [
{ value, views: fallsBackOn, restored: false },
{
value,
views: [...RESTORABLE_VIEWS].filter(
(view) => !fallsBackOn.includes(view),
),
restored: true,
},
];
}
const CONTRACT = [
{
field: "wallets",
@@ -280,28 +316,38 @@ const CONTRACT = [
kind: KIND.SCALAR,
// The prototype members are the whole point: `wallets["map"]` is
// TRUTHY, so hasValidAddress()'s `&&` does not short-circuit and
// `.addresses[…]` throws. A stale INTEGER is the safe case.
hostile: ["map", "__proto__", { a: 1 }],
// `.addresses[…]` throws. A stale INTEGER is the safe case and has to
// stay so. 5 is one, out of range for the one wallet in the profile,
// and it is also this field's truthy polarity: every other value here
// comes back from the floor as null.
hostile: ["map", "__proto__", { a: 1 }, 5],
floorOnly: ["length", "constructor", "toString", "0", -1, 1.5, true],
holds: isIndexOrNull,
// SCALAR, and swept anyway: the restore path is precisely why this
// field gained a floor, so the sweep is the regression guard on it.
alsoSweep: true,
routes: true,
// A stale INTEGER index, which reaches the restore path by a different
// route from the prototype members above — falsy or out of range
// rather than truthy — and has to keep being the safe case.
hostileRestore: [{ value: "length" }, { value: 5 }],
// Comes back from the floor as null, which hasValidAddress() reads as
// nothing selected: the popup falls back to Home on the five views
// that need an address, and lands on every other one.
hostileRestore: everyRestorableView("length", [
"address",
"address-token",
"receive",
"transaction",
"confirm-tx",
]),
},
{
field: "selectedAddress",
kind: KIND.SCALAR,
hostile: ["map", "__proto__", { a: 1 }],
// 5 for the same reason as in selectedWallet: a stale index, and the
// one value here still truthy after the floor.
hostile: ["map", "__proto__", { a: 1 }, 5],
floorOnly: ["length", "constructor", "toString", "0", -1, 1.5, true],
holds: isIndexOrNull,
alsoSweep: true,
routes: true,
hostileRestore: [{ value: 5 }],
},
{
field: "currentView",
@@ -325,7 +371,9 @@ const CONTRACT = [
// container shapes below onto every restorable view; hostileRestore
// adds the records that PASS a branch's gate and then hand its
// renderer something it dereferences, which is where the entries are
// actually decided.
// actually decided. The guard refuses each single-view record below,
// so each is declared to fall back to Home: one that started landing
// would be reaching the renderer it was written against.
hostile: [42, "notarecord", { a: 1 }, [1, 2]],
// `structuredClone(saved.viewData || {})`: the container is never falsy
// in state whatever was stored, so no `!state.viewData` branch exists to
@@ -334,11 +382,20 @@ const CONTRACT = [
hostileRestore: [
// success-tx passes on `data.hash`, and renderSuccess() then calls
// toAddressHtml(d.to) -> addressTitle() -> address.toLowerCase().
{ value: { hash: "0x1" }, views: ["success-tx"] },
{ value: { hash: "0x1", to: 42 }, views: ["success-tx"] },
{
value: { hash: "0x1" },
views: ["success-tx"],
restored: false,
},
{
value: { hash: "0x1", to: 42 },
views: ["success-tx"],
restored: false,
},
{
value: { hash: "0x1", to: ADDRESS, decoded: { details: 7 } },
views: ["success-tx"],
restored: false,
},
{
value: {
@@ -347,12 +404,25 @@ const CONTRACT = [
decoded: { details: [{ address: 42 }] },
},
views: ["success-tx"],
restored: false,
},
// error-tx passes on `data.message`, same dereference.
{ value: { message: "boom" }, views: ["error-tx"] },
{ value: { message: "boom", to: 42 }, views: ["error-tx"] },
{
value: { message: "boom" },
views: ["error-tx"],
restored: false,
},
{
value: { message: "boom", to: 42 },
views: ["error-tx"],
restored: false,
},
// transaction passes on `data.tx`.
{ value: { tx: { hash: "0x1" } }, views: ["transaction"] },
{
value: { tx: { hash: "0x1" } },
views: ["transaction"],
restored: false,
},
{
value: {
tx: {
@@ -363,9 +433,14 @@ const CONTRACT = [
},
},
views: ["transaction"],
restored: false,
},
// confirm-tx passes on `data.pendingTx`.
{ value: { pendingTx: { amount: "1" } }, views: ["confirm-tx"] },
{
value: { pendingTx: { amount: "1" } },
views: ["confirm-tx"],
restored: false,
},
{
value: {
pendingTx: {
@@ -376,6 +451,7 @@ const CONTRACT = [
},
},
views: ["confirm-tx"],
restored: false,
},
// wait-tx passes on `pendingWait.hash`; restoreWait() has checked
// the fields below it since it was written, and this is the
@@ -388,20 +464,29 @@ const CONTRACT = [
},
},
views: ["wait-tx"],
restored: false,
},
// A record that passes EVERY branch's gate at once, driven onto
// every restorable view: a branch a view does not read must stay
// one it does not read, and each renderer must survive the fields
// another branch left behind.
// one it does not read. The five views with a viewData branch
// refuse it; every other one renders it, and must survive the
// fields every branch left behind.
...everyRestorableView(
{
value: {
hash: "0x1",
message: "boom",
tx: { hash: "0x1" },
pendingTx: { amount: "1" },
pendingWait: { hash: "0x1" },
},
},
[
"confirm-tx",
"transaction",
"wait-tx",
"success-tx",
"error-tx",
],
),
],
},
{
@@ -583,6 +668,11 @@ const HEALTHY = { errors: [], blank: false };
// set is all-truthy by construction, so without a falsy slot a dereference
// behind `if (!state.x)` is never reached on the boot that corrupts x — the
// same falsy-collapse blind spot the fields below were floored for.
//
// Only `hostile` and `falsy` values count. Those go through the sweep below,
// which covers every restorable view; a `hostileRestore` entry is driven onto
// only the views it names, so a polarity it alone supplied might reach a single
// renderer.
describe("both polarities of every swept field are driven", () => {
const FALSY_STORED = [0, "", false, null];
const floored = (field, value) =>
@@ -606,10 +696,9 @@ describe("both polarities of every swept field are driven", () => {
test(`${row.field}: truthy and falsy`, () => {
// What the boots below actually drive, floored the way a renderer
// sees it — not what the row says it drives.
const driven = [
...sweptValues(row),
...(row.hostileRestore || []).map((entry) => entry.value),
].map((value) => floored(row.field, value));
const driven = sweptValues(row).map((value) =>
floored(row.field, value),
);
expect({
truthy: driven.some((value) => Boolean(value)),
@@ -865,20 +954,27 @@ describe("every field the router does not read, corrupted at once, onto", () =>
// The values that only mean something on the restore path: a viewData that
// PASSES a branch's gate and then hands its renderer something dereferenced,
// and the index values whose route through hasValidAddress() differs from the
// row's own hostile set.
// and a selectedWallet the floor turns into nothing selected. Each boot must
// throw nothing and show exactly the view its entry says it lands on: its own,
// or Home, so a value written for one renderer cannot stop reaching it and
// still pass.
describe("a restore-only hostile value onto", () => {
for (const row of CONTRACT) {
for (const entry of row.hostileRestore || []) {
for (const view of entry.views || RESTORABLE_VIEWS) {
for (const view of entry.views) {
test(`${view}: ${row.field} = ${JSON.stringify(
entry.value,
)}`, async () => {
await expect(
bootHealth(
const env = await bootPopup(
restoringOnto(view, { [row.field]: entry.value }),
),
).resolves.toEqual(HEALTHY);
);
expect({
errors: env.pageErrors,
visible: env.visibleViews(),
}).toEqual({
errors: [],
visible: [entry.restored ? view : "main"],
});
});
}
}
+19 -3
View File
@@ -131,8 +131,10 @@ describe("Back from Settings after leaving by the settings gear", () => {
// https://git.eeqj.de/sneak/AutistMask/issues/461: leaving drops the
// wallet selection, so Back onto this screen showed a password prompt
// that could only answer "No wallet is selected."
test("does not land on the recovery phrase screen", () => {
// that could only answer "No wallet is selected." Taking the screen off
// the stack leaves Settings under Settings, and Back must not land there
// either (https://git.eeqj.de/sneak/AutistMask/issues/481).
test("goes to the screen under Settings", () => {
const { helpers, state, showPhrase } = load();
// Opened from the wallet list in Settings, then left by the gear:
@@ -143,7 +145,21 @@ describe("Back from Settings after leaving by the settings gear", () => {
expect(state.viewStack).toEqual(["main", "settings"]);
helpers.goBack();
expect(state.currentView).not.toBe(SHOW_PHRASE_VIEW);
expect(state.currentView).toBe("main");
});
// Each round trip leaves one more Settings under Settings.
test("goes to the screen under Settings after two round trips", () => {
const { helpers, state, showPhrase } = load();
for (let i = 0; i < 2; i++) {
showPhrase.show(0);
helpers.pushCurrentView();
helpers.showView("settings");
}
expect(state.viewStack).toEqual(["main", "settings", "settings"]);
helpers.goBack();
expect(state.currentView).toBe("main");
});
// This Back takes Settings off the stack before the screen is left, so