* Support `is:bookmarked` search option to search my bookmarked items
* Update the worker search module to include `bookmarkedBy: Array<Number>` which
contains the list of user ids which have bookmarked a given item
* Add a trigger on the `Bookmark` DB table to re-index the corresponding item when
a bookmark is added/removed
* Update the Search resolver to check for a `is:bookmarked` query option. If provided,
include it as an option in the search request. This updates search to look for items
which are bookmarked by the current user. By default, this preserves stacker privacy
so you can only search your own bookmarks
* Update the search page UI to show how to invoke searching your own bookmarks
* undo `is:bookmarked` support, add `bookmarks` item in search select
* short circuit return empty payload for anon requesting bookmarks
* remove console.log for debugging
* fix indexing a new item that has yet to be bookmarked
* update db migration to re-index all existing bookmarked items one time
* fix the case where deleting a bookmark doesn't trigger a new index of items
explictly specify a `updatedAt` value when deleting a bookmark, to ensure that
deleting a bookmark results in a new indexed version of the bookmarked item
* update search indexer to use the latest of all three choices for the latest version
* give bookmark index jobs longer expiration
---------
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
Co-authored-by: keyan <keyan.kousha+huumn@gmail.com>
* Also log autowithdrawal routing errors
* Only log autowithdrawal success in worker
* Use WalletType for WalletLog.wallet
* Fix autowithdrawal success message
* Infer walletName from walletType in upsertWallet
* Fix first zap when modal closed (#771)
- Extract handlers
- Remove unnecessary async keyword from callback
- Assign a new key to force remounting of LongPressable component when modal is closed from long press
- Remove hover state when closing modal, otherwise it stays colored
* Replace LongPressable with custom component
* Remove yarn.lock
* first pass of a mute mgmt page, ported from subscription mgmt page pr
* adjust error message for mutes
* muted users -> muted stackers
* fix typo in component name
* env vars for polling intervals
add env vars for 4 different common polling intervals,
fast (1000), normal (30000), long (60000), extra long (300000)
use env vars in all `pollInterval` params to `useQuery`
* replace `setInterval`'s interval with `FAST_POLL_INTERVAL`
* use fixed position + div placeholder
* hide footer padding when not shown
* account for mobile inset
---------
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
Co-authored-by: keyan <keyan.kousha+huumn@gmail.com>
* Add .vscode/settings.json to .gitignore to allow local vscode settings without making the work tree dirty
* Swap (fix) Login & SignUp button ids + Make them both 112px wide
This is the onclose listener from getInfo.
This might also be a potential fix for undefined errors that I am getting.
With "undefined errors" I mean that the error itself is literally undefined.
* less confusing expression of when the autowithdraw will be initiated
* only change display value
* tentatively resolved linting errors: spacing around * and removed unused var import
* Remove useRef for NWC relay
* connect to relay for every payment for more reliable payments
* remove getInfoWithRelay method (no longer needed since we no longer use useRef)
* fix 'WebSocket is already in CLOSING or CLOSED state.' errors
* improve logging
* Log connection failures
* Fix no error thrown on validation error
* first pass of a subscription management page under settings
* add tabs to settings ui
* NymActionDropdown
* update Apollo InMemoryCache to merge paginated list of my subscribed users
* various updates
* switch from UsersNullable to Users
* bake the nym action dropdwon into the user component
* add back fields to the user query
* `meSubscriptionPosts`, `meSubscriptionComments`, `meMute`
* Refetch my subscribed users when a user subscription is changed
* update user list to hide stats in the subscribed list users
* update my sub'd users fragment to remove unnecessary user fields
* memoize subscribe user context provider value to avoid re-renders
* use inner join instead of left join
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* throw error when unauthenticated
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
---------
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
Co-authored-by: coderabbitai[bot] <136622811+coderabbitai[bot]@users.noreply.github.com>
* nwc wallet logs
* persist logs in IndexedDB
* Potential fix for empty error message
* load logs limited to 5m ago from IDB
* load logs from past via query param
* Add 5m, 1h, 6h links for earlier logs
* Show end of log
* Clamp to logStart
* Add log.module.css
* Remove TODO about persistence
* Use table for logs
* <table> fixes bad format with fixed width and message overflow into start of next row
* also using ---start of log--- instead of ---end of log--- now
* removed time string in header nav
* Rename .header to .logNav
* Simply load all logs and remove navigation
I realized the code for navigation was most likely premature optimization which even resulted in worse UX:
Using the buttons to load logs from 5m, 1h, 6h ago sometimes meant that nothing happened at all since there were no logs from 5m, 1h, 6h ago.
That's why I added a time string as "start of logs" so it's at least visible that it changed but that looked bad so I removed it.
But all of this was not necessary: I can simply load all logs at once and then the user can scroll around however they like.
I was worried that it would be bad for performance to load all logs at once since we might store a lot of logs but as mentioned, that's probably premature optimization.
WHEN a lot of logs are stored AND this becomes a problem (What problem even? Slow page load?), THEN we can think about this.
If page load ever becomes slow because of loading logs, we could probably simply not load the logs at page load but only when /wallet/logs is visited.
But for now, this works fine.
* Add follow checkbox
* Create WalletLogs component
* Embed wallet logs
* Remove test error
* Fix level padding
* Add LNbits logs
* Add logs for attaching LND and lnAddr
* Use err.message || err.toString?.() consistently
* Autowithdrawal logs
* Use details from LND error
* Don't log test invoice individually
* Also refetch logs on error
* Remove obsolete and annoying toasts
* Replace scrollIntoView with scroll
* Use constant embedded max-height
* Fix missing width: 100% for embedded logs
* Show full payment hash and preimage in logs
* Also parse details from LND errors on autowithdrawal failures
* Remove TODO
* Fix accidental removal of wss:// check
* Fix alignment of start marker and show empty if empty
* Fix sendPayment loop
* Split context in two
* Fix title from draft being dismissed
Add a way to bypass the title generation query when the url change is from local storage (draft) and not user interaction.
* Check draft title from storage
* Remove unused
If an underlying toast finished, an empty toast that automatically immediately hides was dispatched to end the flow ("end flow hack").
If this empty toast had the same tag, the code marked the top toast as hidden even though it was not hidden.
This meant that during render, the animation-delay for the top toast (which was already rendered) was added again, leading to a progress bar jump.
This is fixed by no longer using this "end flow hack" where a toast is dispatched but a dedicated function to end flows.
* Handle archived territories in territory form
* Use dedicated mutation
* Add sanity check for eternal territories
* Fix fields and cost ignored
* Remove no longer needed manual validation in upsertSub
* Remove founder check
* Always check if sub is archived
Using { abortEarly: false } now since previously, if no description was not given, we wouldn't detect if the sub was archived since validation would abort on empty descriptions.
Only on submission all fields would get validated but since we ignore archived errors during submission, the user would never see that the sub is archived before submission
+ the wrong mutation would run if archived is not already true before submission.
Hence, we need to validate all fields always.
There is currently still a bug where the validation does not immediately run but maybe this can be fixed by simply using validateImmediately on the Formik component.
* Fix archived warning not shown after first render
* Only create transfers if owner actually changes
* Reuse helper functions in lib/territory.js
* Rename var to editing
* Use onChange instead of validation override
* Run same validation on server for unarchiving
* Fix 'territory archived' shown during edits
* Use && instead of ternary operator for conditional query
---------
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
* Generate API key in settings
* Check x-api-key for GraphQL API requests
* Don't fallback to cookie if x-api-key header was provided
* Select all session fields
* Fix error if API key not found
* Fix style in settings via form-label className
---------
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
* Ignore if sub belongs to user during existence check
* Remove code no longer needed
* Fix territory edit
Territory edits were broken because validation failed for existing territories and if you edit an territory, it obviously already exists.
This commit fixes this by ignoring the territory that we're currently editing.
* Fix existence check using stale cache
---------
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
* Don't hide self in top even if hidden
* Also don't hide self in top cowboys
* only use anon icon for anon stuff
---------
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
Co-authored-by: keyan <keyan.kousha+huumn@gmail.com>
* Allow founders to transfer territories
* Log territory transfers in new AuditLog table
* Add territory transfer notifications
* Use polymorphic AuditEvent table
* Add setting for territory transfer notifications
* Add push notification
* Rename label from user to stacker
* More space between cancel and confirm button
* Remove AuditEvent table
The audit table is not necessary for territory transfers and only adds complexity and unrelated discussion to this PR.
Thinking about a future-proof schema for territory transfers and how/what to audit at the same time made my head spin.
Some thoughts I had:
1. Maybe using polymorphism for an audit log / audit events is not a good idea
Using polymorphism as is currently used in the code base (user wallets) means that every generic event must map to exactly one specialized event.
Is this a good requirement/assumption? It already didn't work well for naive auditing of territory transfers since we want events to be indexable by user (no array column) so every event needs to point to a single user but a territory transfer involves multiple users.
This made me wonder: Do we even need a table? Maybe the audit log for a user can be implemented using a view? This would also mean no data denormalization.
2. What to audit and how and why?
Most actions are already tracked in some way by necessity: zaps, items, mutes, payments, ...
In that case: what is the benefit of tracking these things individually in a separate table?
Denormalize simply for convenience or performance? Why no view (see previous point)? Use case needs to be more clearly defined before speccing out a schema.
* Fix territory transfer notification id conflict
* Use include instead of two separate queries
* Drop territory transfer setting
* Remove trigger usage
* Prevent transfers to yourself
* Adding tabindex to ImageUpload div wrapper in order to make it selectable
* Adding keyboard event handler to listen to Enter key stroke and trigger file upload selection
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
* show placeholder for hidden stackers in top
* top rewardability views
* make territory revenue idependent job
* monthly rewards and leaderboard on rewards pages
* fix earn reschedule
* add query for rewards leaderboard
* reduce likelihood of rewards racing with views
* fix earn and refine values views
Turbo zaps had different toast bodies so they weren't merged together. This gave stackers the option to undo these zaps out of order.
When zaps are undone out of order, the client cache can get in a bad state. Using the item id as a tag fixes that such that zaps for the same item will always get merged together.
This can be seen as a workaround for hacky zap undo code but I think it's also better UX so maybe we should do this anyway.
The progress bar indicates when the invoice will expire.
This works by passing in a timeout to the withToastFlow wrapper.
If timeout is set, progressBar option will be true for the toast and delay will be set to the timeout.
If progressBar is set, the progress bar will use the delay for its duration.
* Don't throw error if invoice attached
* Only show progress bar for undo toasts
* Update zap undo info in settings
* Skip zap undo toast flow for external payments
* Fix toast progress bar desync
If a toast gets rendered again with the same animation-delay, the animation-delay seems to get added.
This commit fixes that by ensuring that animation-delay is only set if the toast was not rendered before.
* Fix comment
* Territory notifications
* Migrate old setting to new table
* Auto subscribe founders to their territories on creation
* Fix (un)subscribe not shown to founder
* Rename to toggleSubSubscription
* Fix inconsistency between toggleSubSubscription and toggleMuteSub
* Add dedicated button in header for following territories
* Don't drop noteTerritoryPosts column
* Fix db dip in Sub.meSubscription resolver
* Move territory subscribe to new territory context menu
* Decrease space between share icon and mute button
* Fix eslint
* add nterritories field to User
* add userSubs query
* show territories tab on user profiles
hide the tab if user has 0 territories, except when the
viewer navigated directly to the user's territories page
* add USER_WITH_SUBS query for user territories page
* add user territories page
* crosspost-item
* crosspost old items, update with nEventId
* Updating noteId encoding, cleaning up a little
* Fixing item-info condition, cleaning up
* Linting
* Add createdAt variable back
* Change instances of eventId to noteId
* Adding upsertNoteId mutation
* Cleaning up updateItem, using toasts to communivate success/failure in crosspost-item
* Linting
* Move crosspost to share button, make sure only OP can crosspost
* Lint
* Simplify conditions
* user might have no nostr extension installed
Co-authored-by: ekzyis <27162016+ekzyis@users.noreply.github.com>
* change upsertNoteId to updateNoteID for resolver and mutations, change isOp to mine, remove unused noteId params
* Basic setup for crossposting poll / link items
* post rebase fixes and Bounty and job crossposts
* Job crossposting working
* adding back accidentally removed import
* Lint / rebase
* Outsource as much crossposting logic from discussion-form into use-crossposter as possible
* Fix incorrect property for user relays, fix itemId param in updateNoteId
* Fix toast messages / error cases in use-crossposter
* Update item forms to for updated use-crossposter hook
* CrosspostDropdownItem in share updated to accomodate use-crossposter update
* Encode paramaterized replacable event id's in naddress format with nostr-tools, bounty to follw nip-99 spec
* Increase timeout on relay connection / cleaning up
* No longer crossposting job
* Add blastr, fix crosspost button in item-info for polls/discussions, finish removing job crosspostr code
* Fix toaster error, create reusable crossposterror function to surface toaster
* Cleaning up / comments / linting
* Update copy
* Simplify CrosspostdropdownItem, keep replies from being crossposted
* Moved query for missing item fields when crossposting to use-crossposter hook
* Remove unneeded param in CrosspostDropdownItem, lint
* Small fixes post rebase
* Remove unused import
* fix nostr-tools version, fix package-lock.json
* Update components/item-info.js
Co-authored-by: ekzyis <ek@stacker.news>
* Remove unused param, determine poll item type from pollCost field, add mutiny strfry relay to defaults
* Update toaster implementations, use no-cache for item query, restructure crosspostItem to use await with try catch
* crosspost info modal that lives under adv-post-form now has dynamic crossposting info
* Move determineItemType into handleEventCreation, mover item/event handing outside of do ... while loop
* Lint
* Reconcile skip method with onCancel function in toaster
* Handle failedRelays being undefined
* determine item type from router.query.type if available otherwise use item fields
* Initiliaze failerRelays as undefined but handle error explicitly
* Lint
* Fix crosspost default value for link, poll, bounty forms
---------
Co-authored-by: ekzyis <27162016+ekzyis@users.noreply.github.com>
Co-authored-by: ekzyis <ek@stacker.news>
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
* add poll expires at column to Item table
* update upsertPoll mutation for pollExpiresAt param
* use pollExpiresAt to show time left for poll
* correctly pluralize days for timeLeft
* correctly update pollExpiresAt when item is updated to remove poll expiration
* add DateTimePicker and DateTimeInput components to select datetimes
* update pollExpiresAt to be nullable and more than 1 day in the future
* hide time left text if poll has no expiration
* initialize pollExpiresAt with current value or default of 25 hours in the future
we add a one hour time buffer so that the user doesn't get a validation error
for pollExpiresAt if they post their poll within an hour from creation. there's
still a chance they'll hit the validation error but they should see the error
message toast
* add DateTimeInput into the options part of the poll form
add right padding to make room for the "clear" button.
allow field to be cleared (i.e. null pollExpiresAt) to allow
non-ending polls.
---------
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
* Use toast flows
"Toast flows" are a group of toasts that should be shown after each other.
Before this commit, they were implemented by manually removing previous toasts in the same flow.
Now a flowId can be passed and ToastProvider will make sure that there always only exists one toast with the same flowId.
This is different to toast tags since tags don't replace toasts with the same tag, they only are shown "above" them.
* Create wrapper for toast flows
* Fix passing of bolt11 for QR payments
* Fix missing provider check
* Only cancel invoice if hash and hmac were given
* Fix duplicate toast on error
* Fix relay might not be set yet when sendPayment is called
* Validate pubkey, relay URL and secret of NWC URL
* Fix NWC secret regexp
* Use sequential validation in Yup schema
* Add note about possible mismatch between hostnames and pubkeys
* Remove unused param
* add subViewGroup function to create view to read sub stats from
* add topSubs resolver to graphql query
* add TOP_SUBS query fragment
* add SUB_SORTS for top territory sorting
* add custom cache policy for topSubs
* add territories to top header select
* add top territories page
* add db views for sub stats
* configure sub_stats views to refresh by worker
* filter rows with empty subName
* update msats_spent calculation to include all ItemAct in sub
---------
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
* add nsfw column to sub
* add nsfw boolean to territorySchema
* save nsfw value in upsertSub mutation
* return nsfw value from Sub query for correct value in edit territory form
* add nsfw checkbox to territory form
* add nsfw badge to territory header
* add nsfwMode to user
* show nsfw badge next to item territory
* exclude nsfw sub from items query
* show nsfw mode checkbox on settings page
* fix nsfw badge formatting
* separate user from current, signed in user
* update relationClause to join with sub table
* refactor to simplify hide nsfw sql
* filter nsfw items when viewing user items
* hide nsfw posts for logged out users
* filter nsfw subs based on user preference
* show nsfw sub name if logged out user is viewing the page
* show current sub at the top of the list instead of bottom
* always join item with sub to check nsfw
* check for sub presence before showing nsfw badge on item
* skip manually adding sub to select if sub is null
* fix relationClause to join with root item
* move moderation and nsfw into accordion
---------
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
* Refactor setting of default providers
* fixed warning about component update while rendering another component
* individual providers no longer need to know if they are the default or not
* default setting is now handled by WebLNContext -- the same context that returns the provider. this makes a lot more sense and is a lot easier to read
* default payment checkbox is now also disabled if there is only one enabled provider or if it is the default provider
* Fix order lost on page reload
On page reload, the providers were synced in the order they were loaded.
This means that the default payment provider setting was lost.
Fixed this by syncing order to local storage and on page reload, only syncing providers when they were initialized (else the order would have been lost again).
* Add LNbits card
* Save LNbits Provider in WebLN context
* Check LNbits connection on save
* refactor: put LNbitsProvider into own file
* Pay invoices using WebLN provider from context
* Remove deprecated FIXME
* Try WebLN provider first
* Fix unhandled promise rejection
* Fix this in sendPayment
* Be optimistic regarding WebLN zaps
This wraps the WebLN payment promise with Apollo cache updates.
We will be optimistics and assume that the payment will succeed and update the cache accordingly.
When we notice that the payment failed, we undo this update.
* Bold strike on WebLN zap
If lightning strike animation is disabled, toaster will be used.
* Rename undo variable to amount
* Fix zap undo
* Add NWC card
* Attempt to check NWC connection using info event
* Fix NaN on zap
Third argument of update is reserved for context
* Fix TypeError in catch of QR code
* Add basic NWC payments
* Wrap LNbits getInfo with try/catch
* EOSE is enough to check NWC connection
* refactor: Wrap WebLN providers into own context
I should have done this earlier
* Show red indicator on error
* Fix useEffect return value
* Fix wrong usage of pubkey
The event pubkey is derived from the secret. Doesn't make sense to manually set it. It's also the wrong pubkey: we're not the wallet service.
* Use p tag in NWC request
* Add comment about required filter field
* Aesthetic changes to NWC sendPayment
* Add TODO about receipt verification
* Fix WebLN attempted again after error
* Fix undefined name
* Add code to mock NWC relay
* Revert "Bold strike on WebLN zap"
This reverts commit a9eb27daec0cd2ef30b56294b05e0056fb5b4184.
* Fix update undo
* Fix lightning strike before payment
* WIP: Wrap WebLN payments with toasts
* add toasts for pending, error, success
* while pending, invoice can be canceled
* there are still some race conditions between payiny the invoice / error on payment and invoice cancellation
* Fix invoice poll using stale value from cache
* Remove unnecessary if
* Make sure that pay_invoice is declared as supported
* Check if WebLN provider is enabled before calling sendPayment
* Fix bad retry
If WebLN payments failed due to insufficient balances, the promise resolved and thus the action was retried but failed immediately since the invoice (still) wasn't paid.
* Fix cache undo update
* Fix no cache update after QR payment
* refactor: Use fragments to undo cache updates
* Remove console.log
* Small changes to NWC relay mocking
* Return SendPaymentResponse
See https://www.webln.guide/building-lightning-apps/webln-reference/webln.sendpayment
* Also undo cache update on retry failure
* Disable NWC mocking
* Fix initialValue not set
But following warning is now shown in console:
"""
Warning: A component is changing a controlled input to be uncontrolled.
This is likely caused by the value changing from a defined to undefined, which should not happen.
Decide between using a controlled or uncontrolled input element for the lifetime of the component. More info: https://reactjs.org/link/controlled-components
"""
* Remove comment since only relevant for blastr (mutiny relay)
* Remove TODO
* Fix duplicate cache update
* Fix QR modal not closed after payment
* Ignore lnbits variable unused
* Use single relay connection for all NWC events
* Fix missing timer and subscription cleanup
* Remove TODO
Confirmed that nostr-tools verifies events and filters for us.
See https://github.com/nbd-wtf/nostr-tools/blob/master/abstract-relay.ts#L161
* Fix switch from controlled to uncontrolled input
* Show 'configure' on error
* Use budgetable instead of async
* Remove EOSE listener
Only nostr.mutinywallet.com didn't respond with info events due to implementation-specific reasons. This is no longer the case.
* Use invoice expiry for NWC timeout
I don't think there was a specific reason why I used 60 seconds initially.
* Validate LNbits config on save
* Validate NWC config on save
* Also show unattach if configuration is invalid
If unattach is only shown if configuration is valid, resetting the configuration is not possible while it's invalid. So we're stuck with a red wallet indicator.
* Fix detection of WebLN payment
It depended on a Apollo cache update function being available. But that is not the case for every WebLN payment.
* Fix formik bag lost
* Use payment instead of zap in toast
* autoscale capture svc by response time
* docs and changes for testing lnbits locally
* Rename configJSON to config
Naming of config object was inconsistent with saveConfig function which was annoying.
Also fixed other inconsistencies between LNbits and NWC provider.
* Allow setting of default payment provider
* Update TODO comment about provider priority
The list 'paymentMethods' is not used yet but is already implemented for future iterations.
* Add wallet security disclaimer
* Update labels
---------
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
Co-authored-by: keyan <keyan.kousha+huumn@gmail.com>
Currently, all the billing types radios are being assigned the
same "billingType" id, so clicking on any of the labels
always selects the monthly one. If you inspect the HTML, all the
billing type labels have 'for="billingType"' which is how the HTML
knows which input to select.
We have to keep the "name" attribute the same because that's how
the input values are linked to the billingType form field.
To fix, we explicitly assign the "id" prop for each radio so
that the <label>'s "for" attribute is tied to the correct
radio input.
Currently, all the post types checkbox are being assigned the
same "postTypes" id, so clicking on any of the post type labels
always toggles the first one. If you inspect the HTML, all the
post type labels have 'for="postTypes"' which is how the HTML
knows which checkbox to toggle.
We have to keep the "name" attribute the same because that's how
the checkbox values are linked to the postTypes field.
To fix, we explicitly assign the id prop for each checkbox so
that the <label>'s "for" attribute is tied to the correct
checkbox input.
* refactor: replace recursion with promise sequence
This commit refactors `useInvoicable`. The hard-to-follow recursion was replaced by awaiting promises which resolve or reject when one step of our JIT invoice flow is done.
Therefore, `onSubmit` is now fully agnostic of JIT invoices. The handler only returns when payment + action was successful or canceled - just like when a custodial zap was successful.
* refactor more and fix bugs
* move invoice cancel logic into hook where invoice is also created
* fix missing invoice cancellation if user closes modal or goes back.
* refactor promise logic: it makes more sense to wrap the payment promise with the modal promise than the other way around.
* Fix unhandled rejection
* Fix unnecessary prop drilling
* Fix modal not closed after successful action
* Fix unnecessary async promise executor
* Use function to set state
* Show territory details in post form
* Style territory details in post form
* Keep details closed by default
* Use SUB_FULL
* Undo unused changes to specify accordian default
---------
Co-authored-by: ekzyis <ek@stacker.news>
* Only close notifications manually on iOS
* Use function instead of hardcoded string
---------
Co-authored-by: ekzyis <ek@stacker.news>
Co-authored-by: Keyan <34140557+huumn@users.noreply.github.com>
* refactor: Use log function in service worker
* Add verbose logging on push listener
* Fix TypeError: Cannot read properties of null (reading 'postMessage')
navigator.serviceWorker.controller is null on forced refreshes:
"""
This property returns null if the request is a force refresh (Shift + refresh) or if there is no active worker.
"""
-- https://developer.mozilla.org/en-US/docs/Web/API/ServiceWorkerContainer/controller
This means when we unregister a service worker manually (like I do for debugging purposes) and then reload the page, there is no service worker available when this code is run.
Adding a check with a more helpful error message should improve UX.
This error might also happen in other cases where a page refresh should also help.
---------
Co-authored-by: ekzyis <ek@stacker.news>
ekzyis
Abhi Shandilya
Ben Allen
keyan
Keyan
SatsAllDay
✨JP⚡
Felipe Bueno
nl
benalleng
austinkelsay
Satoshi Nakamoto
JP Melanson
mzivil
Austin Kelsay
benthecarman
Alex Lewin
ekzyis