TAK Server (for bridging Meshtastic into TAK clients) #1547
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
* Remove extra want config call when adding a contact * App badge and unnecessary notification fixes (#1455) * - Fix for app badge not going to zero if a message arrives while you have that chat open - Fix for push notifications popping up when a message is received while that chat is open * Fix for cancelling notifications, works now. And scroll to bottom of conversation upon new message * Fix: Channels help grammer fix (#1452) * remove outdated TCP not available on Apple devices (#1450) * Update initial onboarding view * remove toggle gating for mac * Update crash reporting opt out in real time * Update onboarding text * Use mDNS text records for node name * TCP IP and port on the connection screen * Hide app icon chooser on mac * Infinite loop hang bugfixes and performance improvements for both `UserMessageList` and `ChannelMessageList` (#1465) * 2.7.5 Working Changes (#1460) * Remove extra want config call when adding a contact * App badge and unnecessary notification fixes (#1455) * - Fix for app badge not going to zero if a message arrives while you have that chat open - Fix for push notifications popping up when a message is received while that chat is open * Fix for cancelling notifications, works now. And scroll to bottom of conversation upon new message * Fix: Channels help grammer fix (#1452) * remove outdated TCP not available on Apple devices (#1450) * Update initial onboarding view * remove toggle gating for mac * Update crash reporting opt out in real time * Update onboarding text --------- Co-authored-by: Gnome Adrift <[email protected]> Co-authored-by: Zain Kergaye <[email protected]> Co-authored-by: NillRudd <[email protected]> * UserEntity: add mostRecentMessage and unreadMessages with early exit when lastMessage is nil, and fetch 1 row (not N) otherwise * UserList: replace 5 slow calls to user.messageList with new fast calls * NodeList: always put the connected node at the top of list (if it matches the node filters) * ChannelEntity: add faster mostRecentPrivateMessage and unreadMessages which fetch 1 row (not N) * ChannelList: replace 5 calls to channel.allPrivateMessage with new fast calls * Fix incorrect appState.unreadDirectMessages calculations * MyInfoEntity: also fix unreadMessages count here to be fast, and use it for appState.unreadChannelMessages * UserMessageList: use @fetchrequest to prevent the N^2 behavior that was happening in calls to allPrivateMessages * Refactor ChannelEntityExtension and MyInfoEntityExtension to be more similar to UserEntityExtension * Remove SwiftUI-infinite-loop-causing `.id(redrawTapbacksTrigger)` in ChannelMessageList and UserMessageList (duplicate row ids) * MyInfoEntityExtension: exclude emoji tapbacks (which never get marked as read anyway) from unread message count * Add SaveChannelLinkData so MessageText and MeshtasticApp can use .sheet(item: ...) and avoid infinite loop hang due to Binding rebuild * ChannelMessageList and UserMessageList: switch to stable messageId for ForEach SwiftUI row identity * ChannelMessageList and UserMessageList: debouncedScrollToBottom; keyboardWillShowNotification/keyboardDidShowNotification * ChannelMessageList and UserMessageList: scroll to bottom onFirstAppear * ChannelMessageList and UserMessageList: block spurious markMessagesAsRead when this View is not active --------- Co-authored-by: Garth Vander Houwen <[email protected]> Co-authored-by: Gnome Adrift <[email protected]> Co-authored-by: Zain Kergaye <[email protected]> Co-authored-by: NillRudd <[email protected]> * message-list-performance: revert scrolling changes (#1472) * Revert e0f0b4a (ChannelMessageList and UserMessageList: scroll to bottom onFirstAppear) * Revert "ChannelMessageList and UserMessageList: debouncedScrollToBottom; keyboardWillShowNotification/keyboardDidShowNotification" This reverts commit ee1a7c4. --------- Co-authored-by: Gnome Adrift <[email protected]> Co-authored-by: Zain Kergaye <[email protected]> Co-authored-by: NillRudd <[email protected]> Co-authored-by: Jake-B <[email protected]> Co-authored-by: Mike Robbins <[email protected]>
…d from #1424 (#1477) * Fix: "Retrieving nodes" significantly slower after reconnect (#1424) The node database retrieval was calling context.save() for every single NodeInfo packet received (250 saves for 250 nodes). This caused severe performance degradation on reconnect when CoreData had accumulated state. Root Cause: - nodeInfoPacket() called context.save() immediately for each node - With 250 nodes, this meant 250 individual CoreData save operations - On first connection, CoreData is fresh and fast - On reconnect, CoreData has accumulated change tracking, undo management, and memory pressure, making each save progressively slower - This resulted in 10+ second retrieval times vs 1-2 seconds initially Solution: - Added deferSave parameter to nodeInfoPacket() function - During database retrieval (.retrievingDatabase state), defer all saves - Perform a single batch save when database retrieval completes (when NONCE_ONLY_DB configCompleteID is received) - This reduces 250 saves to 1 save Performance Impact: - Eliminates N individual saves during node database sync - Reduces database retrieval time back to 1-2 seconds on reconnect - Matches first-connection performance consistently Fixes #1424 * Revert *MessageListUnified files --------- Co-authored-by: Martin Bogomolni <[email protected]> Co-authored-by: Jake-B <[email protected]>
…ckable at the highest zoom levels (#1478) Co-authored-by: Garth Vander Houwen <[email protected]>
Co-authored-by: Copilot <[email protected]>
…wift Co-authored-by: Copilot <[email protected]>
Co-authored-by: Copilot <[email protected]>
Co-authored-by: Copilot <[email protected]>
Co-authored-by: Copilot <[email protected]>
* Keep list of previous manual connections * More descriptive manual connection rows * Merge fixes and new way to show IP on Connect view --------- Co-authored-by: Jake-B <[email protected]>
* Add identification for node that relayed text messages and add count for ammount of relayers of your message * Ack Relays
Co-authored-by: Copilot <[email protected]>
Co-authored-by: Copilot <[email protected]>
* MeshMap: change onMapCameraChange frequency to .onEnd so that zooming doesn't cause continuous SwiftUI reevaluation on every frame * MeshMapContent: factor out reducedPrecisionMapCircles into a separate function * MeshMapContent: when multiple reducedPrecisionCircles have the same (lat,lon,radius), just draw one (big perf boost in dense areas)
* NodeMapContent: move Route Lines out of ForEach * NodeMapContent: move Convex Hull out of ForEach * NodeMapContent: Replace `position.nodePosition?` with `node` * NodeMapContent: drop unnecessary LazyVStack in showNodeHistory * NodeMapContent: hoist out nodeColorSwift * Move lineCoords, loraCoords calculations within showRouteLines, showConvexHull respectively * Hoist out repeated node.metadata?.positionFlags lookups / PositionFlags creation * NodeMapContent: remove unused @State * NodeMapSwiftUI: add NodeMapContentEquatableWrapper and NodeMapContentSignature to prevent frequent NodeMapContent recomputation and infinite render loops * NodeMapSwiftUI: disable animation during SwiftUI transactions * NodeMapContent: hoist nodeBorderColor and set allowsHitTesting(false) on history point views * NodeMapContent: prerenderHistoryPointCircle and prerenderHistoryPointArrow to avoid thousands of vector draw operations * NodeMapContent: Shared coordinate list for Route Lines and Convex Hull * NodeMapContent.prerenderHistoryPointArrow: add .frame(width: 16, height: 16)
* Update Muzi R1 Neo to actively supported * update & improve zh-Hans translation rt --------- Co-authored-by: Jonathan Bennett <[email protected]> Co-authored-by: Ben Meadors <[email protected]>
Update protobufs to 2.7.1
Add long-turbo preset support
…1512) * Update Muzi R1 Neo to actively supported * Disable Range Test module when primary channel is public/unsecured Updated RangeTestConfig.swift to determine whether the primary channel (index 0) is operating without encryption or with a 1-byte minimal PSK. Disabled Range Test UI controls when on a public/default channel to prevent user interaction. Added safety enforcement in the save operation: Range Test enabled flag is automatically forced to false before sending updates to the device. Introduced a computed property isPrimaryChannelPublic following existing code patterns and security indicators (e.g., hexDescription PSK length). Matches the behavior implemented in the Android client for consistent policy across platforms. --------- Co-authored-by: Jonathan Bennett <[email protected]> Co-authored-by: Ben Meadors <[email protected]>
* Bump version * Message list performance fixes into 2.7.6 (#1475) * Remove extra want config call when adding a contact * App badge and unnecessary notification fixes (#1455) * - Fix for app badge not going to zero if a message arrives while you have that chat open - Fix for push notifications popping up when a message is received while that chat is open * Fix for cancelling notifications, works now. And scroll to bottom of conversation upon new message * Fix: Channels help grammer fix (#1452) * remove outdated TCP not available on Apple devices (#1450) * Update initial onboarding view * remove toggle gating for mac * Update crash reporting opt out in real time * Update onboarding text * Use mDNS text records for node name * TCP IP and port on the connection screen * Hide app icon chooser on mac * Infinite loop hang bugfixes and performance improvements for both `UserMessageList` and `ChannelMessageList` (#1465) * 2.7.5 Working Changes (#1460) * Remove extra want config call when adding a contact * App badge and unnecessary notification fixes (#1455) * - Fix for app badge not going to zero if a message arrives while you have that chat open - Fix for push notifications popping up when a message is received while that chat is open * Fix for cancelling notifications, works now. And scroll to bottom of conversation upon new message * Fix: Channels help grammer fix (#1452) * remove outdated TCP not available on Apple devices (#1450) * Update initial onboarding view * remove toggle gating for mac * Update crash reporting opt out in real time * Update onboarding text --------- Co-authored-by: Gnome Adrift <[email protected]> Co-authored-by: Zain Kergaye <[email protected]> Co-authored-by: NillRudd <[email protected]> * UserEntity: add mostRecentMessage and unreadMessages with early exit when lastMessage is nil, and fetch 1 row (not N) otherwise * UserList: replace 5 slow calls to user.messageList with new fast calls * NodeList: always put the connected node at the top of list (if it matches the node filters) * ChannelEntity: add faster mostRecentPrivateMessage and unreadMessages which fetch 1 row (not N) * ChannelList: replace 5 calls to channel.allPrivateMessage with new fast calls * Fix incorrect appState.unreadDirectMessages calculations * MyInfoEntity: also fix unreadMessages count here to be fast, and use it for appState.unreadChannelMessages * UserMessageList: use @fetchrequest to prevent the N^2 behavior that was happening in calls to allPrivateMessages * Refactor ChannelEntityExtension and MyInfoEntityExtension to be more similar to UserEntityExtension * Remove SwiftUI-infinite-loop-causing `.id(redrawTapbacksTrigger)` in ChannelMessageList and UserMessageList (duplicate row ids) * MyInfoEntityExtension: exclude emoji tapbacks (which never get marked as read anyway) from unread message count * Add SaveChannelLinkData so MessageText and MeshtasticApp can use .sheet(item: ...) and avoid infinite loop hang due to Binding rebuild * ChannelMessageList and UserMessageList: switch to stable messageId for ForEach SwiftUI row identity * ChannelMessageList and UserMessageList: debouncedScrollToBottom; keyboardWillShowNotification/keyboardDidShowNotification * ChannelMessageList and UserMessageList: scroll to bottom onFirstAppear * ChannelMessageList and UserMessageList: block spurious markMessagesAsRead when this View is not active --------- Co-authored-by: Garth Vander Houwen <[email protected]> Co-authored-by: Gnome Adrift <[email protected]> Co-authored-by: Zain Kergaye <[email protected]> Co-authored-by: NillRudd <[email protected]> * message-list-performance: revert scrolling changes (#1472) * Revert e0f0b4a (ChannelMessageList and UserMessageList: scroll to bottom onFirstAppear) * Revert "ChannelMessageList and UserMessageList: debouncedScrollToBottom; keyboardWillShowNotification/keyboardDidShowNotification" This reverts commit ee1a7c4. --------- Co-authored-by: Gnome Adrift <[email protected]> Co-authored-by: Zain Kergaye <[email protected]> Co-authored-by: NillRudd <[email protected]> Co-authored-by: Jake-B <[email protected]> Co-authored-by: Mike Robbins <[email protected]> * Explicitly set unmessagable, seems unnessary * Add back missing mesh map features * Fix: "Retrieving nodes" significantly slower after reconnect extracted from #1424 (#1477) * Fix: "Retrieving nodes" significantly slower after reconnect (#1424) The node database retrieval was calling context.save() for every single NodeInfo packet received (250 saves for 250 nodes). This caused severe performance degradation on reconnect when CoreData had accumulated state. Root Cause: - nodeInfoPacket() called context.save() immediately for each node - With 250 nodes, this meant 250 individual CoreData save operations - On first connection, CoreData is fresh and fast - On reconnect, CoreData has accumulated change tracking, undo management, and memory pressure, making each save progressively slower - This resulted in 10+ second retrieval times vs 1-2 seconds initially Solution: - Added deferSave parameter to nodeInfoPacket() function - During database retrieval (.retrievingDatabase state), defer all saves - Perform a single batch save when database retrieval completes (when NONCE_ONLY_DB configCompleteID is received) - This reduces 250 saves to 1 save Performance Impact: - Eliminates N individual saves during node database sync - Reduces database retrieval time back to 1-2 seconds on reconnect - Matches first-connection performance consistently Fixes #1424 * Revert *MessageListUnified files --------- Co-authored-by: Martin Bogomolni <[email protected]> Co-authored-by: Jake-B <[email protected]> * Hide route lines filter from mesh map * Mesh Map: fuzz imprecise locations so they're distinguishable and clickable at the highest zoom levels (#1478) Co-authored-by: Garth Vander Houwen <[email protected]> * Fix bad merge * Update Meshtastic/Extensions/CoreData/UserEntityExtension.swift Co-authored-by: Copilot <[email protected]> * Update Meshtastic/Views/Nodes/Helpers/Map/MapContent/MeshMapContent.swift Co-authored-by: Copilot <[email protected]> * Update Meshtastic/Extensions/CoreData/ChannelEntityExtension.swift Co-authored-by: Copilot <[email protected]> * Update Meshtastic/Extensions/CoreData/MyInfoEntityExtension.swift Co-authored-by: Copilot <[email protected]> * Update Meshtastic/Extensions/CoreData/ChannelEntityExtension.swift Co-authored-by: Copilot <[email protected]> * Keep list of previous manual connections (#1484) * Keep list of previous manual connections * More descriptive manual connection rows * Merge fixes and new way to show IP on Connect view --------- Co-authored-by: Jake-B <[email protected]> * Show who relayed messages (#1486) * Add identification for node that relayed text messages and add count for ammount of relayers of your message * Ack Relays * upsertPositionPacket: don't use future timestamps to set node's lastHeard (#1488) * R1 NEO * Neo * Update Meshtastic/Views/Settings/AppSettings.swift Co-authored-by: Copilot <[email protected]> * Remove bad if * Git rid of extra environment variable * Update Meshtastic/Accessory/Transports/TCP/TCPTransport.swift Co-authored-by: Copilot <[email protected]> * MeshMap performance: quick wins (#1490) * MeshMap: change onMapCameraChange frequency to .onEnd so that zooming doesn't cause continuous SwiftUI reevaluation on every frame * MeshMapContent: factor out reducedPrecisionMapCircles into a separate function * MeshMapContent: when multiple reducedPrecisionCircles have the same (lat,lon,radius), just draw one (big perf boost in dense areas) * NodeMap performance improvements for high # positions history (#1480) * NodeMapContent: move Route Lines out of ForEach * NodeMapContent: move Convex Hull out of ForEach * NodeMapContent: Replace `position.nodePosition?` with `node` * NodeMapContent: drop unnecessary LazyVStack in showNodeHistory * NodeMapContent: hoist out nodeColorSwift * Move lineCoords, loraCoords calculations within showRouteLines, showConvexHull respectively * Hoist out repeated node.metadata?.positionFlags lookups / PositionFlags creation * NodeMapContent: remove unused @State * NodeMapSwiftUI: add NodeMapContentEquatableWrapper and NodeMapContentSignature to prevent frequent NodeMapContent recomputation and infinite render loops * NodeMapSwiftUI: disable animation during SwiftUI transactions * NodeMapContent: hoist nodeBorderColor and set allowsHitTesting(false) on history point views * NodeMapContent: prerenderHistoryPointCircle and prerenderHistoryPointArrow to avoid thousands of vector draw operations * NodeMapContent: Shared coordinate list for Route Lines and Convex Hull * NodeMapContent.prerenderHistoryPointArrow: add .frame(width: 16, height: 16) * Fix wantRangeTestPackets to correctly follow rangeTestConfig.enabled (#1489) * Fix interval drop down formatter * Clean up channel qr code functionality. * perferredPeripheralId fix * Set opt in * Retry once 5 second timer. dont throw the error * Queue for peripherals * Fix: hoplimit of dms would always fallback to hops away of the node even when configured hops was higher (#1495) * fix hops setting in dms * Fix hops for exchange position * Final fix * Don't favorite client base * Update device hardware * Prevent nil environment metrics * Bump datadog sdk * fix setting device telemetry enabled (#1515) * Update Muzi R1 Neo to actively supported * fix setting device telemetry enabled --------- Co-authored-by: Jonathan Bennett <[email protected]> Co-authored-by: Ben Meadors <[email protected]> * Don't subscribe to mqtt topic if downlink is not on (#1501) * Dont sub if no downlink * moved reload mqtt connect config * Preview enabled in connected devices (#1509) * Update Muzi R1 Neo to actively supported * Preview enabled in connected devices * Fixing indentation * Fixing indentation --------- Co-authored-by: Jonathan Bennett <[email protected]> Co-authored-by: Ben Meadors <[email protected]> * UpdateCoreData.updateAnyPacketFrom: mirror firmware's lastHeard/snr/rssi/hopsAway update logic from NodeDB::updateFrom (#1492) * `CLIENT_BASE` add-favorite/role-change confirmation dialog (#1493) * FavoriteNodeButton: refactor task out * AccessoryManager.connectedDeviceRole helper * FavoriteNodeButton: show confirmation dialog when a CLIENT_BASE is trying to add a favorite * addContactFromURL: add comment referencing upcoming change in meshtastic/firmware#8495 * DeviceConfig: role picker: show a warning when selecting CLIENT_BASE, similar to warning shown for ROUTER * Adjust device configuration Client Base warning text * Compass view (#1521) * Added compass view * Added Compass View * Node colors in compass * Update Muzi R1 Neo to actively supported * Update PositionPopover.swift Co-authored-by: Copilot <[email protected]> * Update CompassView.swift Co-authored-by: Copilot <[email protected]> * Update CompassView.swift Co-authored-by: Copilot <[email protected]> * Update Meshtastic/Views/Helpers/CompassView.swift Co-authored-by: Copilot <[email protected]> * Update Meshtastic/Views/Helpers/CompassView.swift Co-authored-by: Copilot <[email protected]> --------- Co-authored-by: Benjamin Faershtein <[email protected]> Co-authored-by: Jonathan Bennett <[email protected]> Co-authored-by: Ben Meadors <[email protected]> Co-authored-by: Copilot <[email protected]> * Remove discovery queue * revert problematic retry functionalliy * format file * Update & improve zh-Hans translation (#1523) * Update Muzi R1 Neo to actively supported * update & improve zh-Hans translation rt --------- Co-authored-by: Jonathan Bennett <[email protected]> Co-authored-by: Ben Meadors <[email protected]> * Update protobufs to 2.7.1 * Add long-turbo preset * Disable Range Test module when primary channel is public/unsecured (#1512) * Update Muzi R1 Neo to actively supported * Disable Range Test module when primary channel is public/unsecured Updated RangeTestConfig.swift to determine whether the primary channel (index 0) is operating without encryption or with a 1-byte minimal PSK. Disabled Range Test UI controls when on a public/default channel to prevent user interaction. Added safety enforcement in the save operation: Range Test enabled flag is automatically forced to false before sending updates to the device. Introduced a computed property isPrimaryChannelPublic following existing code patterns and security indicators (e.g., hexDescription PSK length). Matches the behavior implemented in the Android client for consistent policy across platforms. --------- Co-authored-by: Jonathan Bennett <[email protected]> Co-authored-by: Ben Meadors <[email protected]> * Add new device images * Remove print statement, get rid of cut and paste error --------- Co-authored-by: Gnome Adrift <[email protected]> Co-authored-by: Zain Kergaye <[email protected]> Co-authored-by: NillRudd <[email protected]> Co-authored-by: Jake-B <[email protected]> Co-authored-by: Mike Robbins <[email protected]> Co-authored-by: Martin Bogomolni <[email protected]> Co-authored-by: Copilot <[email protected]> Co-authored-by: jake-b <[email protected]> Co-authored-by: Benjamin Faershtein <[email protected]> Co-authored-by: Jonathan Bennett <[email protected]> Co-authored-by: Ben Meadors <[email protected]> Co-authored-by: Charles Pinesky <[email protected]> Co-authored-by: Radio <[email protected]> Co-authored-by: Jason Houk <[email protected]>
This is my initial implementation for a TAK Server running inside Meshtastic-Apple.
Contributor
There was a problem hiding this comment.
Pull request overview
This PR implements a TAK (Team Awareness Kit) server feature that enables Meshtastic devices to communicate with TAK clients (like ITAK/ATAK). The implementation includes bidirectional CoT (Cursor on Target) message conversion, TLS/mTLS support, and data package generation for client configuration.
Key Changes
- TAK server with TLS/mTLS support for secure TAK client connections
- Bidirectional protocol bridge converting between CoT messages and Meshtastic TAKPacket protobufs
- Certificate management for server identity and client CA validation
- Data package generation for TAK client configuration
Reviewed changes
Copilot reviewed 27 out of 32 changed files in this pull request and generated 1 comment.
Show a summary per file
| File | Description |
|---|---|
| TAKServerManager.swift | Core TAK server implementation with TLS connection handling |
| TAKMeshtasticBridge.swift | Bidirectional conversion between CoT and TAKPacket formats |
| TAKCertificateManager.swift | Certificate storage and management for TLS/mTLS |
| TAKServerConfig.swift | SwiftUI configuration interface for TAK server settings |
| CoTMessage.swift | CoT message data structures and XML generation |
| AccessoryManager+TAK.swift | Integration with Meshtastic mesh network |
| FountainCodec.swift | Fountain coding for reliable large message transfer |
| TAKConnection.swift | Individual TAK client connection handler |
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
Contributor
There was a problem hiding this comment.
Pull request overview
Copilot reviewed 27 out of 32 changed files in this pull request and generated no new comments.
💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.
thebentern
commented
Jan 6, 2026
thebentern
commented
Jan 6, 2026
Collaborator
Author
|
@niccellular I gave this a pretty thorough review this morning and I'm generally pretty happy with how it turned out. I left a couple of comments / questions. @garthvh my only concern is that we have a lot of logging in here and I was wondering if we should do anything to isolate the TAK based logs into it's own silo so that we don't polute the existing meshtastic logging in a significant way. |
thebentern
changed the title
Member
|
The logging is all filterable with "TAK" but yes there is an absurd amount of verbose logging including the output of the entire CoT XML -> Protobuf conversion including byte dumping. |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
@niccellular's original work. I'm just opening the PR for review