You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
v0.169.0-alpha-main (current production build, via standard launcher)
π Issue Description:
When launching Explorer pointed at a local scene dev server on a non-loopback LAN IP over http (e.g. http://192.168.68.127:8000), the client shows a "you are about to enter a custom catalyst" prompt, then the loading screen hangs indefinitely with no further feedback or error message to the user. This reproduces identically on both the PR #9765 build and the current production build (v0.169.0-alpha-main) β confirming it's a pre-existing gap, not something introduced by the bugsweep PR. The underlying blocking mechanism differs by build (see notes below), but in both cases the user is left staring at a frozen loading screen with zero explanation.
π£ STR:
Start a local SDK7 scene server (npm run start) so it's reachable on the machine's LAN IP (not just loopback)
Launch Explorer with: open ./Decentraland.app --args --realm http://<LAN_IP>:8000 --position 0,0 --local-scene true --debug --scene-console --login-bridge-only
Accept the "custom catalyst" prompt and wait on the loading screen
βοΈ Expected Result:
Either the app clearly communicates that the connection was blocked/rejected (e.g. an "insecure connection not allowed" message or similar), or it fails fast with an actionable error β not an indefinite silent hang.
β Actual Result with evidence:
Loading screen hangs with no error shown to the user. All test sessions were manually force-quit after ~19-21 seconds of no progress.
On PR #9765 build, Player.log shows during asset bundle fetch attempts:
Curl error 35: Cert verify failed. Certificate could not be verified (either omitted or unsupported). UnityTls error code: 7
Occurred 3x per session, across two separate test attempts, immediately preceding manual app quit. Consistent with the new app-level cleartext-scheme enforcement (wearable-mainfile-insecure-url lane, EnforceSecureScheme) upgrading the non-loopback http request to https, then failing TLS handshake since the local dev server has no cert. (Inferred from the TLS failure signature β no EnforceSecureScheme/insecure/cleartext string is directly logged.)
On production build (v0.169.0-alpha-main), Player.log shows instead:
Non-secure network connections disabled in Player Settings
Same silent hang, same lack of any in-app user messaging β this time it's Unity's engine-level insecureHttpOption (still NotAllowed on production) hard-blocking the connection outright.
In both cases, the connection is correctly prevented from silently succeeding over cleartext β not a security bug. The bug is purely the missing user-facing feedback: nothing tells the user why the loading screen is stuck or what to do about it.
β Reproduction:
π Always - (3/3 attempts: 2x on PR #9765 build, 1x on production build, same silent-hang result every time)
π₯οΈ Operative system and additional Notes:
macOS β Apple M3 Pro.
Confirmed pre-existing β reproduces on current production (v0.169.0-alpha-main) as well as the PR #9765 build, so this is not a regression introduced by the bugsweep PR and does not block that PR's merge. Filing as its own item since it's a real UX gap: likely a good candidate to fix as part of the local-scene-development / Creator Hub preview flow specifically, since that's the workflow most likely to hit a non-loopback address in practice.
π οΈ Build version:
Reproduces on both:
v0.169.0-alpha-bugsweep/aug16-consolidated-22a1a3d(PR fix: bugsweep aug16Β #9765, commit22a1a3d)v0.169.0-alpha-main(current production build, via standard launcher)π Issue Description:
When launching Explorer pointed at a local scene dev server on a non-loopback LAN IP over http (e.g.
http://192.168.68.127:8000), the client shows a "you are about to enter a custom catalyst" prompt, then the loading screen hangs indefinitely with no further feedback or error message to the user. This reproduces identically on both the PR #9765 build and the current production build (v0.169.0-alpha-main) β confirming it's a pre-existing gap, not something introduced by the bugsweep PR. The underlying blocking mechanism differs by build (see notes below), but in both cases the user is left staring at a frozen loading screen with zero explanation.π£ STR:
npm run start) so it's reachable on the machine's LAN IP (not just loopback)open ./Decentraland.app --args --realm http://<LAN_IP>:8000 --position 0,0 --local-scene true --debug --scene-console --login-bridge-onlyβοΈ Expected Result:
Either the app clearly communicates that the connection was blocked/rejected (e.g. an "insecure connection not allowed" message or similar), or it fails fast with an actionable error β not an indefinite silent hang.
β Actual Result with evidence:
Loading screen hangs with no error shown to the user. All test sessions were manually force-quit after ~19-21 seconds of no progress.
On PR #9765 build, Player.log shows during asset bundle fetch attempts:
Occurred 3x per session, across two separate test attempts, immediately preceding manual app quit. Consistent with the new app-level cleartext-scheme enforcement (
wearable-mainfile-insecure-urllane,EnforceSecureScheme) upgrading the non-loopback http request to https, then failing TLS handshake since the local dev server has no cert. (Inferred from the TLS failure signature β noEnforceSecureScheme/insecure/cleartextstring is directly logged.)On production build (
v0.169.0-alpha-main), Player.log shows instead:Same silent hang, same lack of any in-app user messaging β this time it's Unity's engine-level
insecureHttpOption(stillNotAllowedon production) hard-blocking the connection outright.In both cases, the connection is correctly prevented from silently succeeding over cleartext β not a security bug. The bug is purely the missing user-facing feedback: nothing tells the user why the loading screen is stuck or what to do about it.
β Reproduction:
π Always - (3/3 attempts: 2x on PR #9765 build, 1x on production build, same silent-hang result every time)
π₯οΈ Operative system and additional Notes:
macOS β Apple M3 Pro.
Confirmed pre-existing β reproduces on current production (
v0.169.0-alpha-main) as well as the PR #9765 build, so this is not a regression introduced by the bugsweep PR and does not block that PR's merge. Filing as its own item since it's a real UX gap: likely a good candidate to fix as part of the local-scene-development / Creator Hub preview flow specifically, since that's the workflow most likely to hit a non-loopback address in practice.Player test 1 - wearable-mainfle.log
Player.log