Only the first parameter, `type`, was ever passed in. The others are
superfluous. The defaults have been set directly instead.
It's a bit silly to set the `rpcUrl` and `nickname` to an empty string,
but to make this more sensible would take much more effort. This at
least is simpler and guaranteed to be equivalent.
* Add stringified payload to metametrics controller trackEvent error message
* Don't include full payload when throwing error on missing event or category in trackEvent
The `assert` module has two modes: "Legacy" and "strict". When using
strict mode, the "strict" version of each assertion method is implied.
Whereas in legacy mode, by default it will use the deprecated, "loose"
version of each assertion.
We now use strict mode everywhere. A few tests required updates where
they were asserting the wrong thing, and it was passing beforehand due
to the loose matching.
* Add 'What's New' notification popup
* Move selectors from shared/notifications into ui/ directory
* Use keys for localized message in whats new notifications objects, to ensure notifications will be translated.
* Remove unused swaps intro popup locale messages
* Fix keys of whats new notification locales
* Remove notifications messages and descriptions from comment in shared/notifications
* Move notifcationActionFunctions to shared/notifications and make it stateless
* Get notification data from constants instead of state in whats-new-popup
* Code cleanup
* Fix build quote reference to swapsEthToken, broken during rebase
* Rename notificationFilters to notificationToExclude to clarify its purpose
* Documentation for getSortedNotificationsToShow
* Move notification action functions from shared/ to whats-new-popup.js
* Stop setting swapsWelcomeMessageHasBeenShown to state in app-state controller
* Update e2e tests for whats new popup changes
* Updating migration files
* Addressing feedback part 1
* Addressing feedback part 2
* Remove unnecessary div in whats-new-popup
* Change getNotificationsToExclude to getNotificationsToInclude for use in the getSortedNotificationsToShow selector
* Delete intro-popup directory and test files
* Lint fix
* Add notifiction state to address-entry fixture
* Use two separate functions for rendering first and subsequent notifications in the whats-new-popup
* Ensure that string literals are passed to t for whats new popup text
* Update import-ui fixtures to include notificaiton controller state
* Remove unnecessary, accidental change confirm-approve
* Remove swaps notification in favour of mobile swaps as first notifcation and TBD 3rd notification
* Update whats-new-popup to use intersection observer api to detect if notification has been seen
* Add notifications to send-edit and threebox e2e test fixtures
* Update ui/app/selectors/selectors.js
Co-authored-by: Mark Stacey <markjstacey@gmail.com>
* Update ui/app/selectors/selectors.js
Co-authored-by: Mark Stacey <markjstacey@gmail.com>
* Clean up locale code for whats-new-popup notifications
* Disconnect observers in whats-new-popup when their callback is first called
* Add test case for migration 58 for when the AppStateController does not exist
* Rename popover components containerRef to popoverWrapRef
* Fix messages.json
* Update notification messages and images
* Rename popoverWrapRef -> popoverRef in whats-new-popup and popover.component
* Only create one observer, and only after images have loaded, in whats-new-popup
* Set width and height on whats-new-popup image, instead of setting state on img load
* Update ui/app/components/app/whats-new-popup/whats-new-popup.js
Co-authored-by: Mark Stacey <markjstacey@gmail.com>
* Code clean up in whats new popup re: notification rendering and action functions
* Code cleanup in render notification functions of whats-new-popup
* Update ui/app/components/app/whats-new-popup/whats-new-popup.js
Co-authored-by: Mark Stacey <markjstacey@gmail.com>
* lint fix
* Update and localize notification dates
* Clean up date code in shred/notifications/index.js
Co-authored-by: ryanml <ryanlanese@gmail.com>
Co-authored-by: Mark Stacey <markjstacey@gmail.com>
* Remove periodic calls to the /featureFlag API
* Always show the Swap button on the main page
* Check if the Swaps feature is enabled, show loading animation while waiting
* Reuse an existing useEffect call
* Use ‘isFeatureFlagLoaded’ in React’s state, resolve lint issues
* Add a watch mode for Jest testing
* Add unit tests for Swaps: fetchSwapsLiveness, add /ducks/swaps into Jest testing
* Remove Swaps Jest tests from Mocha’s ESLint rules
* Ignore Swaps Jest tests while running Mocha, update paths
* Increase test coverage to the current max
* Fix ESLint issues for Swaps
* Enable the Swaps feature by default and after state reset, remove loading screen before seeing Swaps
* Update Jest config, fix tests
* Update Jest coverage threshold to the current maximum
* Update ESLint rule in jest.config.js
* Disable the “Review Swap” button if the feature flag hasn’t loaded yet
* Update jest threshold
* Swaps support for local testnet
* Create util method for comparison of token addresses/symbols to default swaps token
* Get chainId from txMeta in _trackSwapsMetrics of transaction controller
* Add comment to document purpose of getTransactionGroupRecipientAddressFilter
* Use isSwapsDefaultTokenSymbol in place of repeated defaultTokenSymbol comparisons in build-quote.js
Our automatic token detection was hard-coded to only work on our built-
in Infura Mainnet endpoint. It now works with custom Mainnet RPC
endpoints as well.
Relates to #6992
`eth_getProof` is an unpermissioned, read-only RPC method for getting account-related Merkle proofs, specified here: https://eips.ethereum.org/EIPS/eip-1186
It's been supported by major Ethereum clients, and Infura, for some time. By adding it to the safe methods list, we enable this method for our users.
This PR introduces the new approval controller to the extension codebase. We use it for the permissions controller's pending approval functionality.
The approval controller sets us up for a new pattern of requesting and managing user confirmations in RPC methods. Along with the generic RPC method middleware, the approval controller will allow us to eliminate our message managers, and decouple various method handlers from our provider stack, making the implementations more portable between the extension and mobile.
From a behavioral standpoint this PR fixes the issue with tracking, and persisting, tokens that the user hides. Whether we can/should optimize this to prevent duplicates of the accountHiddenTokens and hiddenToken is a point of contention, but it acts similiarly to how we track tokens and accountTokens.
Also to note, for tokens under a custom network there is no way to distinguish two different custom network sets of hidden tokens, they are all under the `rpc` property, same as accountTokens.
* @metamask/inpage-provider@^8.0.0
* Replace public config store with JSON-RPC notifications
* Encapsulate notification permissioning in permissions controller
* Update prefix of certain internal RPC methods and notifications
* Add accounts to getProviderState
* Send accounts with isUnlocked notification (#10007)
* Rename provider streams, notify provider of stream failures (#10006)
Attempts to send metrics would fail when no `options` were used. This
was because when the options parameter was not set, it was often sent
over our RPC connection as `undefined`, which gets serialized to `null`
when the message is converted to JSON. This `null` parameter didn't
trigger the default parameter set in the metametrics controller, as
default parameters are only used for `undefined`.
Instead the `options` parameter is now treated as fully optional, with
no default value set. The optional chaining operator is used to ensure
it won't blow up if it's not set. A fallback of `{}` was used for the
one destructure case as well.
If a `gasPrice` was specified in a transaction sent via a dapp, we
would include it in our `eth_estimateGas` call, causing it to fail if
the user had insufficient balance (for either the transaction amount or
the gas fee). This resulted in the fallback gas estimate being used;
the block gas limit. The block gas limit is quite a bit larger than
most transactions need, so this resulted in wildly inflated gas costs
being shown on our confirmation screen.
The `gasPrice` has been removed from the `txParams` object we pass to
`eth_estimateGas`, so now it won't perform any balance checks anymore.
This ensures that we'll get a valid gas estimate, as long as geth is
able to simulate the contract execution properly.
Fixes#9967
* Remove use of ethgassthat; use metaswap /gasPrices api for gas price estimates
* Remove references to ethgasstation
* Pass base to BigNumber constructor in fetchExternalBasicGasEstimates
* Update ui/app/hooks/useTokenTracker.js
Co-authored-by: Erik Marks <25517051+rekmarks@users.noreply.github.com>
* Delete gas price chart
* Remove price chart css import
* Delete additional fee chart code
* Lint fix
* Delete more code no longer used after ethgasstation removal
Co-authored-by: Erik Marks <25517051+rekmarks@users.noreply.github.com>
A few inconsistencies in JSDoc formatting have been fixed throughout
the project. Many issues remain; these were just the few things that
were easy to fix with a regular expression.
The changes include:
* Using lower-case for primitive types, but capitalizing non-primitive
types
* Separating the parameter identifier and the description with a dash
* Omitting a dash between the return type and the return description
* Ensuring the parameter type is first and the identifier is second (in
a few places it was backwards)
* Using square brackets to denote when a parameter is optional, rather
than putting "(optional)" in the parameter description
* Including a type and identifier with every parameter
* Fixing inconsistent spacing, except where it's used for alignment
* Remove incorrectly formatted `@deprecated` tags that reference non-
existent properties
* Remove lone comment block without accompanying function
Additionally, one parameter was renamed for clarity.
The `seedPhraseBackedUp` now tracks whether or not the seed phrase has
been backed up. Previously this defaulted to `true`, which left no way
to distinguish whether it had been backed up or not during onboarding.
The default is now `null`, and the UI logic has been updated to account
for this, so that "existing users" (i.e. users that have a backup that
is years old) aren't mistakenly considered to have not backed up their
seed phrase. This value is already set explicitly to `true` or `false`
during onboarding, in both the create and import flow.
This change was made primarily to make it easier to fix the onboarding
library integration, which will be done in a subsequent PR.
* Alternative savings fix
* Further required changes to savings fix
* Further fix to savings calculations that properly accounts for metamask fees
* metaMaskFeeInEth property on quotes to decimal string
* Fix swaps controller unit tests
* Improve documentation in swaps controller
* Prevent getMedianEthValueQuote from mutation passed quotes array with .sort() call
* Another fix and refactor to savings calculations in _findTopQuoteAndCalculateSavings
Cleaner structuring of conditionals for setting tokenValueOfQuoteForSorting, ethValueOfQuote and metaMaskFeeInEth in swaps controller
Stop subtracting medianMetaMaskFee from savings, but include it in savings data
Another fix and refactor to savings calculations in _findTopQuoteAndCalculateSavings
* Add and update unit tests for _findTopQuoteAndCalculateSavings
* Improve calculation of overallValueOfQuoteForSorting for case where ETH is the source token
* Clean up getMedianEthValueQuote code, test and comments
* Clean up _findTopQuoteAndCalculateSavings, create test input and expected results helper functions
* Update getMedianEthValueQuote to account for multiple quotes with overall values equal to the median
* Add jsdoc comment for meansOfQuotesFeesAndValue
* Fix jsdoc comment for getMedianEthValueQuote
* create custom addHexPrefix function
* switch to custom addHexPrefix
Co-authored-by: Mark Stacey <markjstacey@gmail.com>
Co-authored-by: Erik Marks <rekmarks@protonmail.com>
The incoming transactions controller now uses the `chainId` for the
current network instead of the `networkId`. This ensures that custom
RPC endpoints for the built-in supported networks do correctly receive
incoming transactions.
As part of this change, the incoming transactions controller will also
cease keeping track of the "last block fetched" for networks that are
not supported. This piece of state never really represented the last
block fetched, as _no_ blocks were fetched for any such networks. It
been removed.
If the swaps state is cleared in between the initial quote fetch and
the subsequent poll fetch, a `TypeError` will be thrown due to
`fetchParams` being set to `null`.
This is of no functional consequence, as `fetchParams` _should_ be
`null` in this case, and and no further action should be taken.
The optional chaining operator is now used to ensure the call no longer
throws.
This is a continuation of #9726, which did not fix the problem
described.
If the initial network when the extension is started is something other
than Mainnet, the swaps controller will never successfully retrieve
swap quotes. This is because `ethers` will continue to communicate
with whichever network the provider was initially on.
We tried fixing this by hard-coding the `chainId` to Mainnet's
`chainId` when constructing the Ethers provider, but this did not work.
I suspect this failed because the `provider` we pass to `ethers` is not
compliant with EIP 1193, as `ethers` doubtless expects it to be.
Instead the entire `ethers` provider is now reconstructed each time the
network changes. This mirrors the approach we take in some other
controllers.
If the initial network when the extension is started is something other
than Mainnet, the swaps controller will never successfully retrieve
swap quotes. This is because the `ethers` provider used by the swaps
controller doesn't allow network changes by default - it assumes that
the network remains the same as when the provider was initialized.
This was fixed by hard-coding Mainnet as the initial chain ID for this
`ethers` provider used by the swaps controller.
Some adjustments needed to be made to the `provider` stub to allow
setting `1` as the network ID and chain ID in unit tests.
Refs #9663
See [`node/no-deprecated-api`][1] for more information.
This change enables `node/no-deprecated-api` and fixes the issues raised by the rule.
[1]:https://github.com/mysticatea/eslint-plugin-node/blob/v11.1.0/docs/rules/no-deprecated-api.md
The change to the way that `punycode` is imported is to address the fact that
third-party module is hidden by the built-in. This is a silly hack but it works.
Consolidates the background and UI segment implementations into a shared solution.
This results in the introduction of our first shared module.
Co-authored-by: Erik Marks <25517051+rekmarks@users.noreply.github.com>
`@metamask/eslint-config` has been updated to v4.1.0. This update
requires that we update `eslint` to v7 as well, which in turn requires
updating most `eslint`-related packages.
Most notably, `babel-eslint` was replaced with `@babel/eslint-parser`,
and `babel-eslint-plugin` was replaced by `@babel/eslint-plugin`. This
required renaming all the `babel/*` rules to `@babel/*`.
Most new or updated rules that resulted in lint errors have been
temporarily disabled. They will be fixed and re-enabled in subsequent
PRs.
* Calculate savings per swap relative to median values
* Update test mock quotes, add getMedian tests
* Identify assets by sourceToken and destinationToken
* Add data point to 'Swaps Completed' segment event: estimated vs used gas
* Linted
* Correct property name for estimated gas on swapMetaData in _trackSwapsMetrics()
* Set estimated_gas property on swapMetaData to a hex string
* Correct base when dividing by estimated_gas
Co-authored-by: Dan Finlay <dan@danfinlay.com>
* Delete localhost provider type
* Use ganache-cli default chain ID for tests
* Delete unused test firstTimeState variable
* Migrate default ganache-cli network to frequentRpcListDetail
* Add default test provider state
* Add test functionality to createJsonRpcClient
* Lint locales
* Update test middleware creation
* fixup! Update test middleware creation
* Update txMeta after postTxBalance has been retrieved
* Use gas used from txReceipt to calculate eth received
* Return null from getSwapsTokensReceivedFromTxMeta in tokenSymbol is ETH and txReceipt is missing
* Get latest txMeta before updating it with postTxBalance in case of a swaps tx in confirmTransaction
* Lint fix
* Use initial transaction for settings swap transaction title params, and remove addition of swap properties to cancel transcations
* Update unit test data
* Use token symbol properties from initial transaction for filitering in transaction list
* call this.txStateManager.setTxStatusConfirmed before async call in confirmTransaction in the transactions controller
* Clone txMeta before setTxStatusConfirmed in confirmTransaction
* Correctly updateTx in confirmTransaction
* Track swaps event only if it is a swap transaction
* Create wrapper function for segment events
* Extract transaction controller metrics calls into own function
Co-authored-by: Mark Stacey <markjstacey@gmail.com>
* Stop passing a gas param to the estimateGas call initiated in the swaps controller timedoutGasReturn
* Stop passing gas params to timedoutGasReturn
* Lint fix
* Stop passing no longer used param to setInitialGasEstimate
When the `chainId` for a custom RPC endpoint is edited, we now migrate
the corresponding address book entries to ensure they are not orphaned.
The address book entries are grouped by the `metamask.network` state,
which unfortunately was sometimes the `chainId`, and sometimes the
`networkId`. It was always the `networkId` for built-in Infura
networks, but for custom RPC endpoints it would be set to the user-set
`chainId` field, with a fallback to the `networkId` of the network.
A recent change will force users to enter valid `chainId`s on all
custom networks, which will be normalized to be hex-prefixed. As a
result, address book contacts will now be keyed by a different string.
The contact entries are now migrated when this edit takes place.
There are some edge cases where two separate entries share the same set
of contacts. For example, if two entries have the same `chainId`, or if
they had the same `networkId` and had no `chainId` set. When the
`chainId` is edited in such cases, the contacts are duplicated on both
networks. This is the best we can do, as we don't have any way to know
which network the contacts _should_ be on.
The `typed-message-manager` unit tests have also been updated as part
of this commit because the addition of `sinon.restore()` to the
preferences controller tests ended up clearing a test object in-between
individual tests in that file. The test object is now re-constructed
before each individual test.
The `_fetchAndSetSwapsLiveness` was accidentally passed to
`setInterval` without being bound first, so the `this` reference was
not defined when it was called. It is now bound before being passed to
`setInterval`.
* Remove network config store
* Remove inline networks variable in network controller
* Re-key network controller 'rpcTarget' to 'rpcUrl'
* Require chainId in lookupNetwork, implement eth_chainId
* Require chain ID in network form
* Add alert, migrations, and tests
* Add chainId validation to addToFrequentRpcList
* Update public config state selector to match new network controller
state
* Use network enums in networks-tab.constants
* Ensure chainId in provider config is current
* Update tests
* Fix require-unicode-regexp issues
See [`require-unicode-regexp`](https://eslint.org/docs/rules/require-unicode-regexp) for more information.
This change enables `require-unicode-regexp` and fixes the issues raised by the rule.
* Remove case-insensitive flag from regexps
This method was accidentally broken with the introduction of the
permissions controller, as this was missing from the list of safe
methods.
It is now included in the list of safe methods.
Fixes#8993
It seems that this blocklist checker never worked correctly. Ever since
the initial commit, it was comparing the Number `1` to the `networkId`,
which is a string. Additionally, even if it did throw, the transaction
continued unhindered. The user could still approve it, and there was no
indication shown to the user that anything went wrong. Also some of the
blocklist entries were incorrectly mixed-case, and were never hit.
We can remove this for now, and re-add it later on after we rewrite the
transaction controller.
The `metamaskNetworkId` property in the `txMeta` for incoming
transactions was incorrectly set as a Number instead of a String. This
change was made accidentally as part of #8627.
As a result incoming transactions were being excluded from the
transaction list, as they didn't have a matching network ID.
`metamaskNetworkId` is now set to a string, and a migration has been
added to ensure `metamaskNetworkId` is converted to a string for any
incoming transactions in state.
* Use over the whole stringified error object which doesn't show the actual error message that is set as the
Co-authored-by: Mark Stacey <markjstacey@gmail.com>
* Feedback commit
The code for checking whether a transaction was dropped or not was
refactored in #8398, but in the process an off-by-one error was
introduced.
The old version of `_checkIfTxWasDropped` would query for an updated
transaction count from the network, and would consider the pending
transaction to be dropped if the count was above the nonce. However,
the version introduced in #8398 considers the transaction to be dropped
if the count is above *or equal to* the nonce.
The pending transaction nonce is expected to be equal to the
transaction count, because the nonce starts at zero. The transaction
count is equal to the expected next nonce.
The variable name has been updated to make this more clear
(`networkNextNonce` is how the `nonce-tracker` refers to this value).
`parseInt` is now called with an explicit radix of `16` as well, to
ensure both nonce strings are always parsed as hex. In all cases I am
aware of, these nonce strings were prefixed by `0x`, meaning that
`parseInt` would default to a radix of `16`, so this likely doesn't
constitute a functional change.
Fixes#8688
* update connected accounts appearance
* consolidate account alerts
* UnconnectedAccountAlert: use ConnectedAccountsList
* move switch account action out of menu in all views
Co-authored-by: Mark Stacey <markjstacey@gmail.com>
Imported accounts can be removed, but the permissions controller is not
informed when this happens. Permissions are now removed as part of the
account removal process.
Additionally, the `getPermittedIdentitiesForCurrentTab` selector now
filters out any non-existent accounts, in case a render occurs in the
middle of an account removal.
This was resulting in a render crash upon opening the popup on a site
that was connected to the removed account.
The `TokenRatesController` was accidentally broken in #8744, when the
logic for starting and stopping polling was moved from the `isActive`
property to start/stop functions.
A reference to the now-obsolete `isActive` property was accidentally
left behind, resulting in no exchange rate updates.
Defaults have been added for all three preferences. The default values
added are both falsey, so this shouldn't result in any functional
change. This was done to help make this preferences more easily
discoverable.
The max listener count of the preferences store has been increased to
12. Recently the 12th listener was added, which resulted in console
warnings during the unit tests - this prevents those warnings.
The default max listener value is 10; we didn't see this warning until
now because one of the twelve listeners is only setup when 3Box is
enabled, which doesn't occur during our unit tests.
Add alert suggesting that the user switch to a connected account. This
alert is displayed when the popup is opened over an active tab that is
connected to some account, but not the current selected account. The
user can choose to switch to a connected account, or dismiss the alert.
This alert is only shown once per account switch. So if the user
repeatedly opens the popup on a dapp without switching accounts, it'll
only be shown the first time. The alert also won't be shown if the user
has just dismissed an "Unconnected account" alert on this same dapp
and account, as that would be redundant.
The alert has a "Don't show me this again" checkbox that allows the
user to disable the alert. It can be re-enabled again on the Alerts
settings page.
All transaction status updates were moved into a `setTimeout` callback
and wrapped in a `try...catch` block in #4131, apparently in an attempt
to prevent failures in event subscribers from interrupting the
transaction logic. The `try...catch` block did accomplish that, but by
putting the status update in a `setTimeout` callback the operation was
made asynchronous.
Transaction status updates now happen unpredictably, in some future
event loop from when they're triggered. This creates a race condition,
where the transaction status update may occur before or after
subsequent state changes. This also introduces a risk of accidentally
undoing a change to the transaction state, as the update made to the
transaction inside the `setTimeout` callback uses a reference to
`txMeta` obtained synchronously before the `setTimeout` call. Any
replacement of the `txMeta` between the `setTxStatus` call and the
execution of the timeout would be erased. Luckily the `txMeta` object
is more often than not mutated rather than replaced, which may explain
why we haven't seen this happen yet.
Everything seems to work correctly with the `setTimeout` call removed,
and now the transaction logic is easier to understand.
The unconnected account alert can now be disabled. A "don't show this
again" checkbox has been added to the alert, which prevents that alert
from being shown in the future.
An alert settings page has been added to the settings as well. This
page allows the user to disable or enable any alert.
This controller was not used. It was used by the
`ComputedBalancesController`, which was removed in #7057 (as it was
also unused).
The pending balances calculator was only used by the balances
controller.
The inactive timer was being reset upon any change to the preferences
store. The intent was only to update the timer when the auto-lock
timeout had changed, so the subscription was updated to only update in
those cases.
There are no indications that this had any effect upon the user. It
looks like the preferences store never updates while the extension is
unattended, so in practice this may have been harmless. It was still
pointless however. This also protects against the possibility of the
preferences store being updated while unattended at some point in the
future.
A race condition exists where after adding an unapproved transaction,
it could be mutated and then replaced when the default gas parameters
are set. This happens because the transaction is added to state and
broadcast before the default gas parameters are set, because
calculating the default gas parameters to use takes some time.
Once they've been calculated, the false assumption was made that the
transaction hadn't changed.
The method responsible for setting the default gas now retrieves an
up-to-date copy of `txMeta`, and conditionally sets the defaults only
if they haven't yet been set.
This race condition was introduced in #2962, though that PR also added
a loading screen that avoided this issue by preventing the user from
interacting with the transaction until after the gas had been
estimated. Unfortunately this loading screen was not carried forward to
the new UI.
* Remove `estimatedGas` property from `txMeta`
The `estimatedGas` property was a cache of the gas value estimated for
a transaction when the default gas limit was set. This property wasn't
used anywhere. It may have been useful for debugging purposes, but the
same gas estimate is already stored on the `history` property so it
should be present in state logs regardless.
* Remove `gasLimitSpecified` txMeta property
The `gasLimitSpecified` property of `txMeta` wasn't used for anything.
It might have been useful for debugging purposes, but whether or not
the gas limit was specified can also be determined from looking at the
transaction history, so it's not a huge loss.
* Remove `gasPriceSpecified` txMeta property
The `gasPriceSpecified` property of `txMeta` wasn't used for anything.
It might have been useful for debugging purposes, but whether or not
the gas price was specified can also be determined from looking at the
transaction history, so it's not a huge loss.
* Remove `simpleSend` txMeta property
The `simpleSend` property of `txMeta` was used to ensure a buffer was
not added to the gas limit during gas estimation for simple send
transactions. It was made redundant by #8484, which accomplishes this
without the use of this property.
Previously a transaction would get assigned a default value during the
`addTxGasDefaults` function, after the transaction was added and sent
to the UI.
Instead the transaction is assigned a default value before it gets
added. This flow is simpler to follow, and it avoids the race condition
where the transaction is assigned a value from the UI before this
default is set. In that situation, the UI-assigned value would be
overridden, which is obviously not desired.
`analyzeGasUsage` now returns the results of the analysis rather than
setting them directly on `txMeta`. The caller is now responsible for
mutating `txMeta` instead. Functionally this should be identical to
before.
