feat: map sharing by cimigree · Pull Request #1693 · digidem/comapeo-mobile

cimigree · 2026-02-12T20:53:57Z

closes #1570
closes #1757
closes #1755
closes #1756
closes #1784

Adds the Map Sharing Feature.

awana-lockfile-bot · 2026-02-12T20:54:11Z

`src/backend/package-lock.json` changes

Click to toggle table visibility

Name	Previous	Current
`@comapeo/core`	6.0.2	7.1.0
`@comapeo/default-categories`	1.0.0	1.0.1
`@comapeo/ipc`	7.0.0	8.0.0
`@comapeo/map-server`	1.0.1	1.1.3
`@emnapi/core`	1.4.3	1.9.2
`@emnapi/runtime`	1.4.3	1.9.2
`@emnapi/wasi-threads`	1.0.2	1.2.1
`@gmaclennan/zip-reader`	-	1.0.0
`@mapbox/mapbox-gl-style-spec`	-	14.21.0
`@mapbox/point-geometry`	-	1.1.0
`@napi-rs/wasm-runtime`	0.2.9	0.2.12
`@sqlite.org/sqlite-wasm`	-	3.51.2-build8
`@tybys/wasm-util`	0.9.0	0.10.1
`cheap-ruler`	-	4.0.0
`csscolorparser`	-	1.0.3
`ready-resource`	1.1.1	1.2.0
`smp-noto-glyphs`	-	1.0.0-pre.0
`styled-map-package-api`	-	5.0.0-pre.4
`wsl-utils`	0.1.0	-
`zip-writer`	-	2.2.0

awana-lockfile-bot · 2026-02-12T20:54:12Z

`package-lock.json` changes

Summary

Status	Count
	22
	10
	1

Click to toggle table visibility

Name	Previous	Current
`@comapeo/cloud`	0.3.0	0.4.0
`@comapeo/core-react`	10.0.1	11.0.4
`@comapeo/core`	6.0.2	7.1.0
`@comapeo/ipc`	7.0.0	8.0.0
`@comapeo/map-server`	1.0.1	1.1.3
`@gmaclennan/zip-reader`	-	1.0.0
`@hyperswarm/secret-stream`	6.8.1	6.9.1
`@mapbox/mapbox-gl-style-spec`	-	14.21.0
`@mapbox/point-geometry`	-	1.1.0
`@sqlite.org/sqlite-wasm`	-	3.51.2-build8
`@turf/bbox-polygon`	-	7.3.4
`@turf/boolean-point-in-polygon`	6.5.0	7.3.4
`@turf/clone`	-	7.3.4
`@turf/point-to-line-distance`	-	7.3.4
`@turf/point-to-polygon-distance`	-	7.3.4
`@turf/polygon-to-line`	-	7.3.4
`@turf/projection`	-	7.3.4
`@turf/rhumb-bearing`	-	7.3.4
`@turf/rhumb-distance`	-	7.3.4
`bare-ansi-escapes`	-	2.2.3
`bare-assert`	-	1.2.0
`bare-inspect`	-	3.1.4
`bare-type`	-	1.1.0
`csscolorparser`	-	1.0.3
`noise-handshake`	4.1.0	4.2.0
`point-in-polygon-hao`	-	1.2.4
`ready-resource`	1.1.2	1.2.0
`robust-predicates`	-	3.0.3
`smp-noto-glyphs`	-	1.0.0-pre.0
`styled-map-package-api`	-	5.0.0-pre.4
`typebox`	1.1.10	1.1.16
`wsl-utils`	0.1.0	-
`zip-writer`	-	2.2.0

socket-security · 2026-02-12T20:54:58Z

Review the following changes in direct dependencies. Learn more about Socket for GitHub.

Package	Supply Chain Security	Quality	Maintenance
npm/@comapeo/default-categories@1.0.0 ⏵ 1.0.1			⁺³
npm/@comapeo/map-server@1.0.1 ⏵ 1.1.3		⁺¹	⁺¹
npm/@comapeo/ipc@7.0.0 ⏵ 8.0.0			⁺¹
npm/@comapeo/core@6.0.2 ⏵ 7.1.0	⁺¹	⁺¹

View full report

cimigree · 2026-02-12T21:17:03Z

  earlyAccessStore,
  appUsageStatsStore,
 }: AppProvidersProps) => {
+  const mapServerListenPromise = React.useMemo(


I had to memoize this because I was getting an error:
"Listen method has been called more than once without closing"

Happened on force close and restart
Because the appRpc.mapServer.listen() call was happening on every render of the AppProviders component.
It was also complaining about needing an argument (hence the empty object)

Good point, I'll have a think about improving the ergonomics of the api for this - it's too easy right now to use it in a way that causes errors.

socket-security · 2026-02-12T21:28:37Z

All alerts resolved. Learn more about Socket for GitHub.

This PR previously contained dependency changes with security issues that have been resolved, removed, or ignored.

View full report

…tendfor package json tests.

cimigree · 2026-02-12T21:37:48Z

    )) {
      const backendVersion = backendMapeoDependencies[dependency];
      if (!backendVersion) continue;
+      // For file: dependencies, compare just the filename (paths may differ)


Do you think it's ok to do this just for tests? That way more pressing, problematic test failures will be visible.

You could also just comment out the expect during dev until it's ready - we shouldn't be using file dependencies by the time the pr is finished

Or don't run these tests on draft prs

gmaclennan · 2026-02-13T18:51:00Z

I've update the core-react PR #143 with what I hope is a fix for this issue. Here is an updated tarball with the latest from commit digidem/comapeo-core-react@22f6d95

comapeo-core-react-7.3.0-pre.3-MAP-SHARING.tgz

gmaclennan · 2026-02-13T19:24:10Z

  earlyAccessStore,
  appUsageStatsStore,
 }: AppProvidersProps) => {
+  const mapServerListenPromise = React.useMemo(


Good point, I'll have a think about improving the ergonomics of the api for this - it's too easy right now to use it in a way that causes errors.

gmaclennan · 2026-02-13T19:28:40Z

    )) {
      const backendVersion = backendMapeoDependencies[dependency];
      if (!backendVersion) continue;
+      // For file: dependencies, compare just the filename (paths may differ)


You could also just comment out the expect during dev until it's ready - we shouldn't be using file dependencies by the time the pr is finished

gmaclennan · 2026-02-13T19:29:00Z

    )) {
      const backendVersion = backendMapeoDependencies[dependency];
      if (!backendVersion) continue;
+      // For file: dependencies, compare just the filename (paths may differ)


Or don't run these tests on draft prs

gmaclennan · 2026-02-13T19:31:57Z

+                {projectId, receiverDeviceId: deviceId, mapId: 'custom'},
+                {
+                  onSuccess: result => {
+                    Promise.resolve(result).then(mapShare => {


Something up with my code here - I'll take a look

cimigree · 2026-02-13T21:30:45Z

@gmaclennan
So sorry but getting more fetch/ http type errors.
For the useGetCustomMapInfo: [Error: fetch failed: [start] Cannot convert 'http://127.0.0.1:40461/maps/custom/info' to a Kotlin type.]

And for the sendMapShare: [Error: fetch failed: [start] Cannot convert 'http://127.0.0.1:45635/mapShares' to a Kotlin type.]

The bird thinks:
The fetch implementation in @comapeo/core-react is not compatible with React Native on Android.

But again, I can't figure anything else out so don't know if that is the actual case

cimigree · 2026-02-16T20:47:45Z

@gmaclennan
Mostly Good News!
I am using maps from here: https://drive.google.com/drive/folders/19iZUWrlTQIBpqr7MnP0u5E5dmFudJYKt (the QA testing assets)
The first map I uploaded and tried to share was this one:
maplibre-demotiles.smp
It uploaded fine but when I tried to share it I got this error:
[Error: Bounds at 3 must be within max of spherical mercator projection -180,-85.051129,180,85.051129]

But then I tried a couple other maps and they worked with no error to get to the next step. I haven't yet gotten further than that, yet. Sorry.
This works now for the two maps below but not for the first one (see above).

 const mapShare = await sendMapShare({
                  projectId,
                  receiverDeviceId: deviceId,
                  mapId: 'custom',
                });
                navigation.navigate('WaitingForMapAccept', {
                  shareId: mapShare.shareId,
                });

demotiles-z2 (1).smp

And this one:
osm-bright-z6 (1).smp

I am going to continue on to next steps after I clean this up a bit (UI in that background map screen isn't quite right), but I thought you wanted to know about where I was able to get so far.

cimigree · 2026-04-06T18:16:29Z

@ErikSin Sorry for the messy and repeated pushes but I somehow messed up the node_modules/ package lock and had to go back and start from scratch.

Added a keep alive for the map sharing screens so they don't go dark during map sharing, which would cancel the map sharing.
Also updated/ upgraded all the backend comapeo packages to the latest versions.
- This adds much faster map loading which closes Map Black after receiving map file #1755
- Doing so allows us to use the error handling which is a little less fragile than just checking for strings. See this PR to comapeo/core-react for details.
  - That closes Map Reciever Cannot properly cancel a map being shared #1757
  - And closes ReplaceBackgroundMap actions not working #1756
We are now also able to share maps between two users who are both IN the same project even if they are ON different projects, so I changed the select device for map sharing accordingly.
I also updated the map share cancelled sheet so that when it closes it just goes back instead of to the background maps screen as specified by design.
I did not take on feat: replace background map #1704 (comment) because I wasn't sure about it... I will wait to hear your thoughts.
Apk sent to you via Slack.

ErikSin · 2026-04-06T23:17:20Z

Just starting to review, and just noting that my package lock is slightly different when i run npm i. Its just different by one line but we should figure out where the change is coming from. Can you run npm i again and see what happens.

I am using node-v24.13.0 and npm-v11.6.2

ErikSin · 2026-04-07T00:14:48Z

We are now also able to share maps between two users who are both IN the same project even if they are ON different projects, so I changed the select device for map sharing accordingly.

This is not working as expected

Screen.Recording.2026-04-06.at.5.05.14.PM.mov

It seems like from one device it doesnt matter if your actively on the project. But for the other device it does matter?

I initially thought it was one of the devices was a particant and not a coordinator, but that doesn't effect the outcome.

I don't know if it matters that much? but weird behaviour none the less

ErikSin

Its better, now that we have upgraded core and core react, but its still not working fully as expected. See the videos below, but it seems like my samsung Galaxy A03s is having a lot of troubles sending and recieving the cancel in a timely manner. It also is having trouble finding the other device.

Im not sure what the best way to move forward is. This feature doesn't feel ready to release as is.

ErikSin · 2026-04-07T00:19:57Z

+        <View style={styles.buttonsContainer}>
+          {warningInfo.warning !== 'space' && (
+            <PrimaryButton fullSize text={t(m.accept)} onPress={handleAccept} />
+          )}
+          {declineStatus === 'pending' ? (
+            <Loading size={12} />
+          ) : (
+            <SecondaryButton
+              fullSize
+              text={t(m.decline)}
+              onPress={handleDecline}
+            />
+          )}
+        </View>
+      </View>


This UI is a little off.

for buttons we use a <UIActivtyIndicator/>, not <Loading/>

we need block both button if were going to show a loader, not just the decline button

ErikSin · 2026-04-07T00:42:58Z

+    ? Math.floor((currentTime.getTime() - mapShare.mapShareCreatedAt) / 1000)
+    : 0;
+
+  const cancelShare = React.useCallback(() => {


This is taking several minutes for the other device to register:

Screen.Recording.2026-04-06.at.5.39.06.PM.mov

cimigree · 2026-04-07T18:48:37Z

Just starting to review, and just noting that my package lock is slightly different when i run npm i. Its just different by one line but we should figure out where the change is coming from. Can you run npm i again and see what happens.

I am using node-v24.13.0 and npm-v11.6.2

yes actually there was a change! So sorry I have no idea why!
This is where it is:
"node_modules/bare-events": {
"version": "2.8.2",
"resolved": "https://registry.npmjs.org/bare-events/-/bare-events-2.8.2.tgz",
"integrity": "sha512-riJjyv1/mHLIPX4RwiK+oW9/4c3TEUeORHKefKAKnZ5kyslbN+HXowtbaVEqt4IMUB7OXlfixcs6gsFeo/jhiQ==",
"license": "Apache-2.0",
"peer": true,
"peerDependencies": {
"bare-abort-controller": "*"
},
"peerDependenciesMeta": {
"bare-abort-controller": {
"optional": true
}
}
},

Adds the bolded line.

ErikSin · 2026-04-07T19:37:18Z

Just starting to review, and just noting that my package lock is slightly different when i run npm i. Its just different by one line but we should figure out where the change is coming from. Can you run npm i again and see what happens.
I am using node-v24.13.0 and npm-v11.6.2

yes actually there was a change! So sorry I have no idea why! This is where it is: "node_modules/bare-events": { "version": "2.8.2", "resolved": "https://registry.npmjs.org/bare-events/-/bare-events-2.8.2.tgz", "integrity": "sha512-riJjyv1/mHLIPX4RwiK+oW9/4c3TEUeORHKefKAKnZ5kyslbN+HXowtbaVEqt4IMUB7OXlfixcs6gsFeo/jhiQ==", "license": "Apache-2.0", "peer": true, "peerDependencies": { "bare-abort-controller": "*" }, "peerDependenciesMeta": { "bare-abort-controller": { "optional": true } } },

Adds the bolded line.

Can you push these changes?

…ges in the same screens.

ErikSin

I tested it with an APK, and things do seem to be working better...Im still not convinced that its in a state to release. But i am going to send an APK to QA and let them decide whether the frontend should pause on development (and wait for the backend team to update), or if we can go ahead with merging and releasing as is

ErikSin · 2026-04-07T21:18:32Z

+  const {shareId} = route.params;
+
+  const mapShare = useSingleSentMapShare({shareId});
+  const {mutate: cancelMapShare} = useCancelSentMapShare();


I think we need catch the status==='pending' here, can show a loading state when that is true

…atus for showing a loading state. Makes sure to navigate back instead of to background maps.

Should fix issues with delays on some actions on certain phones.

gmaclennan · 2026-04-15T16:35:11Z

Just to mention again, the flashing of screens and the need to catch different states that @ErikSin is catching in the review are mostly related to the architectural decision to split this into screens, which I pointed out in #1704 (comment) and in several conversations before that. It is possible to keep finding these edge-cases and fixing them as you have been, but changing the architecture would simplify this a lot and avoid these bugs appearing in the first place.

cimigree · 2026-04-15T16:58:43Z

Just to mention again, the flashing of screens and the need to catch different states that @ErikSin is catching in the review are mostly related to the architectural decision to split this into screens, which I pointed out in #1704 (comment) and in several conversations before that. It is possible to keep finding these edge-cases and fixing them as you have been, but changing the architecture would simplify this a lot and avoid these bugs appearing in the first place.

I changed it to the way you wanted it which is what was actually causing that flash that Erik mentioned in
2f5341b

I left out the one screen that is arrived at through user interaction but combined all of the restl.

…r mapshare. Updates test for same.

ErikSin

WOWWW That was several months that we can never get back...

Thanks for having the patience to finish this!

gmaclennan and others added 9 commits February 12, 2026 14:57

Add map sharing integration to backend (WIP)

2d5c2ee

First pass attempt to incorporate map server api. Not working so far.

9dcffb0

patch and fix

1609596

roll back changes for testing

a12372f

remove debug statement

afd321d

Incorporates backend work and dependency fixes.

f27cfa2

Tweaks to code to get it to load a custom map.

bceae52

Deletes extra tarball

79930d4

Merge branch 'develop' into feat/waiting-map-accept

c33f08b