This PR ensures that the `getLastConnectedInfo` selector handles missing `eth_accounts` permission history. Historically, `eth_accounts` was the only permission in existence, and the only way that a permission subject ended up with a permission history in the first place. This will no longer the case as of #11837, and we were perhaps never right to bake in this assumption to begin with.
* Update the copy for the Flask welcome page (#13223)
* Update the copy for the Flask welcome page
The copy for the Flask Welcome page has been updated to better dissuade
users who are not the target audience, and to better explain the risks
of using Flask.
* Fix typo
* Suggested edits (#13225)
* Suggested edits
* fixup! Suggested edits
Co-authored-by: Erik Marks <25517051+rekmarks@users.noreply.github.com>
* Update app/_locales/en/messages.json
Co-authored-by: David Walsh <davidwalsh83@gmail.com>
Co-authored-by: Erik Marks <25517051+rekmarks@users.noreply.github.com>
Co-authored-by: David Walsh <davidwalsh83@gmail.com>
The first page of the Flask onboarding was causing a propType warning
to appear in the console. It was caused by the array of React Fragments
used to construct the ASCII fox; they were missing the `key` prop.
These fragments are static content, so React doesn't really need to
worry about what to do in the event they are re-ordered. The array
index has been used as the key to silence the warning.
When a lot of transactions are occurring on the network, such as during
an NFT drop, it drives gas fees up. When this happens, we want to not
only inform the user about this, but also dissuade them from using a
higher gas fee (as we have proved in testing that high gas fees can
cause bidding wars and exacerbate the situation).
The method for determining whether the network is "busy" is already
handled by GasFeeController, which exposes a `networkCongestion`
property within the gas fee estimate data. If this number exceeds 0.66 —
meaning that the current base fee is above the 66th percentile among the
base fees over the last several days — then we determine that the
network is "busy".
ESLint rules have been added to enforce our JSDoc conventions. These
rules were introduced by updating `@metamask/eslint-config` to v9.
Some of the rules have been disabled because the effort to fix all lint
errors was too high. It might be easiest to enable these rules one
directory at a time, or one rule at a time.
Most of the changes in this PR were a result of running
`yarn lint:fix`. There were a handful of manual changes that seemed
obvious and simple to make. Anything beyond that and the rule was left
disabled.
The ESLint config for the extension explicitly includes support for
Prettier. However, this is already being provided by our global ESLint
config (`@metamask/eslint-config`). Therefore there is no need to
include it here. In fact, this is causing weird issues where the `curly`
option is getting overridden somehow. After this change, these syntaxes
are invalid:
``` javascript
if (foo) return;
```
``` javascript
if (foo) return 'bar';
```
* Update support links for Flask
* Disable 'prefer-const' in code fence linting
* Add bespoke home footer for Flask and update logic
* fixup! Add bespoke home footer for Flask and update logic
* Fix code fence lint failure
* Fix support request link in account menu
* Fix unit test failure
There was a propType error shown on the snap install screen about the
`name` property of `targetSubjectMetadata` being missing despite it
being marked as required. This property should not have been required,
it does not always exist.
* Prevent automatic rejection of confirmations
Confirmations are now only automatically rejected if a user explicitly
closes the notification window. If we close the window programmatically
because there are no notifications left to show, nothing gets rejected.
This partially avoids a race condition where a confirmation gets
rejected automatically without the user having seen the confirmation
first. This could happen if the confirmation was processed just as the
notification window was being closed.
It's still possible for a confirmation that the user has never seen to
get rejected as a result of the user closing the window. But at least
now it's no longer possible for a confirmation to get rejected in this
manner after the user resolves the last confirmation in the queue.
* Fix bug that prevented automatic closure detection
All windows were being detected as explicit window closures,
essentially just as they were previously, because this variable was
cleared too soon.
* Re-open popup when necessary
After the window is automatically closed, a confirmation may have been
queued up while the window was closing. If so, the popup is now re-
opened.
This error was introduced with #13100, which was merged without CI
checks because CircleCI was not running on that branch for some reason.
This error was fixed with `yarn lint:styles --fix`.
The `mounted` state was used to derive state from props before the
first render of the Home component. Instead this state is now derived
in the constructor, which is also run before the first render. This
should behave exactly the same, except now we don't need the `mounted`
state or the `deriveStateFromProps` function anymore.
The call to `closeCurrentWindow` that was made in `componentDidUpdate`
has been moved to the constructor as well. There is no need to delay
that call, and this saves us from having to compare current with
previous state in that lifecycle function.
The confirmation page template system allows templates to have alerts,
but it throws an uncaught Promise rejection if a template has no
alerts. This is the unintended side-effect of input validation.
The `getTemplateAlerts` function was updated to always return an array.
The one callsite was updated to expect an empty array if there were no
alerts to render, rather than expecting `undefined`.
This selector is a duplicate of the `getSubjectMetadata` selector,
which does the same thing except that there is no fallback for the case
where the `subjectMetadata` is falsy. This is OK because that state can
never be falsy.
This change was extracted from the `snaps` branch.
* added wrapper around account list to prevent storybook from collapsing the list
* updated translation files
* added snap-connect page
* refactored account list out of the choose account component
* fixed width
* removed unnecessary scss from choose-account component, fixed props in choose account story
* removed snaps-connect page, added comments to ChooseAccount
* updated choose account subtitle text, updated styling for title & subtitle, removed redundant account list story
* updated component name, updated paths
* fixed linter errors
* added comments
* removed unused message
* removed selectAccounts key from all locales
* updated class name for account list header, updated allAreSelected function to use length checks
* Revert "removed unused message"
This reverts commit 32771bc83c08f120825ef75f0741f3034e7dbecb.
* Revert "removed selectAccounts key from all locales"
This reverts commit ccfa4a860f9a75693d893d7c404384e719de297e.
* updated locale messages to use selectAccounts key
* removed stray import
* updated scss
* updated translation key
* removed chooseAccounts key from en locale
* removed optional chaining
* changes
* updated subjectMetadata
* updated subject types
* update useOriginMetadata function to include unknown subject type
* updated permission connect header props, removed host and added subjectType to targetSubjectMetadata
* added subjectType to targetSubjectMetadata
* removed console.log
* changed prop name to iconUrl
* Make this a form
Similar to "import-with-seed-phrase" I would like to be able to restore my vault by pressing `<enter>` on my keyboard.
* actually test enter
* preventDefault()
The Home notification component now has a storybook file. This work was
extracted from the `snaps` branch, and was originally implemented in
PR #12860.
The storybook file was enhanced to include more stories demonstrating
each likely usage scenario.
The ESLint config has been updated to v8. The breaking changes are:
* The Prettier rule `quoteProps` has been changed from `consistent` to
`as-needed`, meaning that if one key requires quoting, only that key is
quoted rather than all keys.
* The ESLint rule `no-shadow` has been made more strict. It now
prevents globals from being shadowed as well.
Most of these changes were applied with `yarn lint:fix`. Only the
shadowing changes required manual fixing (shadowing variable names were
either replaced with destructuring or renamed).
The dependency `globalThis` was added to the list of dynamic
dependencies in the build system, where it should have been already.
This was causing `depcheck` to fail because the new lint rules required
removing the one place where `globalThis` had been erroneously imported
previously.
A rule requiring a newline between multiline blocks and expressions has
been disabled temporarily to make this PR smaller and to avoid
introducing conflicts with other PRs.
