Upon the first render, the "original" signature request confirmation
page would trigger a PropType error. This was caused by unexpected
mutation of the state props. The container has been updated to avoid
mutating the props, and now the PropType warning is no longer present.
This fixes a bug where the network menu would remain present after a
second click on the network menu button. The bug was caused by the
click being handled _twice_, by two separate handlers. First it was
caught by the external click handler of the dropdown menu, which closed
the menu. Second, it was caught by the network button itself, which re-
opened the menu. This all happens quickly enough that to the user it
appears to stay open.
The external click handler of the menu now only fires if the menu is
open. Additionally, any click that is caught by the network menu is
stopped from propagating further, so that it can't trigger additional
click handlers.
The `menu-droppo` external click handler would fail to update when the
prop changed while the menu wasn't shown. This bug isn't exposed
anywhere yet as far as I know, but I ran into it when working on a
different bug fix.
The Chip component was emitting a PropType error because it was missing
the `labelProps.children` prop. It was never supposed to be given that
prop - it was a mistake in the PropType declaration. The PropTypes have
been fixed to prevent this warning.
Fixes#5611
The origin that suggests a transaction is now shown on the transaction
confirmation page. If the transaction was initiated from within
MetaMask (e.g. via the 'Send' flow or swaps), no origin is shown.
This was based upon designs that were linked in the PR #9377. This is a
temporary measure until our newer transaction confirmation designs can
be implemented.
Fixes#6071
The origin of the dapp that suggested signing has been added to the
signature request confirmation page. This only applies to `eth_sign`,
`personal_sign`, `eth_signTypedData`, and `eth_signTypedData_v1`. The
confirmation page for `eth_signTypedData_v3` and `eth_signTypedData_v4`
already featured the origin.
The decrupt message confirmation UI will crash if the origin metadata
is not present. This PR makes the UI tolerant of that metadata being
missing. It was always intended to be optional anyway.
Fixes#9244
When trying to connect a Trezor account on a fresh install of MetaMask,
the radio buttons on the account selection page would not respond to
being clicked.
When debugging this, it looks like the `onChange` event was never
triggered. A radio `<input>` element should trigger `onChange` whenever
the selection state change, but seemingly this wouldn't happen if the
change in selection state was undone during the same render cycle. If
I paused at a breakpoint during the render, I could see the checkbox
get selected then unselected again without triggering `onChange`.
The simplest fix was to use `onClick` instead of `onChange`. This seems
more appropriate anyway because we're treating the radio button as a
controlled component here, so the state of the underlying element isn't
really of any concern.
The "Next nonce" warning warns users when the custom nonce they set is
higher than our suggested nonce. This warning was mistakenly being
shown even when we didn't have a suggested nonce yet.
Fixes#9989
New year, new problems.
It's working as expected, but we had hard-coded some 2020 date values, and `formatDateWithYearContext` adds the year to its output formatted date if the date is not from the current year.
Ensures that `hideLoadingIndication` is always called in all actions that call `showLoadingIndication`. It's unclear how many of these actions were failing to hide the loading indication, because other actions superset `hideLoadingIndication`.
At the very least, `updateTransaction` was probably failing to hide the loading indication in the error case.
This PR also refactors a lot of actions to call `hideLoadingIndication` once in `finally` blocks as opposed to multiple times across `try` and `catch` blocks. We avoided making changes to functions using `Promise` methods, because `Promise.finally` is not supported by Waterfox, and it's not properly transpiled by Babel.
A data race was introduced in #9919 when the old synchronous storage
API was replaced with an async storage API. The problem arises when
`fetchWithCache` is called a second time while it's still processing
another call. In this case, the `cachedFetch` object can become
stale while blocked waiting for a fetch response, and result in a cache
being overwritten unintentionally.
See this example (options omitted for simplicity, and assuming an empty
initial cache):
```
await Promise.all([
fetchWithCache('https://metamask.io/foo'),
fetchWithCache('https://metamask.io/bar'),
]
```
The order of events could be as follows:
1. Empty cache retrieved for `/foo` route
2. Empty cache retrieved for `/bar` route
3. Call made to `/foo` route
4. Call made to `/bar` route
5. `/foo` response is added to the empty cache object retrieved in
step 1, then is saved in the cache.
6. `/bar` response is added to the empty cache object retrieved in
step 2, then is saved in the cache.
In step 6, the cache object saved would not contain the `/foo`
response set in step 5. As a result, `/foo` would never be cached.
This problem was resolved by embedding the URL being cached directly in
the cache key. This prevents simultaneous responses from overwriting
each others caches.
Technically a data race still exists when handing simultaneous
responses to the same route, but the result would be that the last call
to finish would overwrite the previous. This seems acceptable.