Crash in [@ mozilla.appservices.relay.InternalException: at mozilla.appservices.relay.RelayKt.uniffiCheckCallStatus(relay.kt:46)]
Categories
(Firefox for Android :: Privacy, defect)
Tracking
()
People
(Reporter: clouseau-bot, Unassigned)
References
(Blocks 1 open bug)
Details
(Keywords: crash)
Crash Data
Crash report: https://crash-stats.mozilla.org/report/index/ff21a6ec-42fb-4f96-a115-0e5690261015
Top 10 frames:
0 mozilla.appservices.relay.RelayKt uniffiCheckCallStatus toolkit/components/uniffi-bindgen-gecko-js/android/components/relay/android/src/main/java/mozilla/appservices/relay/relay.kt
1 mozilla.appservices.relay.RelayKt access$uniffiCheckCallStatus toolkit/components/uniffi-bindgen-gecko-js/android/components/relay/android/src/main/java/mozilla/appservices/relay/relay.kt
2 mozilla.appservices.relay.RelayClient fetchProfile toolkit/components/uniffi-bindgen-gecko-js/android/components/relay/android/src/main/java/mozilla/appservices/relay/relay.kt
3 mozilla.components.service.fxrelay.FxRelayImpl$fetchProfile$2$2 invokeSuspend mobile/android/android-components/components/service/firefox-relay/src/main/java/mozilla/components/service/fxrelay/FxRelay.kt
4 mozilla.components.service.fxrelay.FxRelayImpl$fetchProfile$2$2 invoke mobile/android/android-components/components/service/firefox-relay/src/main/java/mozilla/components/service/fxrelay/FxRelay.kt
5 mozilla.components.service.fxrelay.FxRelayImpl handleRelayExceptions mobile/android/android-components/components/service/firefox-relay/src/main/java/mozilla/components/service/fxrelay/FxRelay.kt
6 mozilla.components.service.fxrelay.FxRelayImpl$fetchProfile$2 invokeSuspend mobile/android/android-components/components/service/firefox-relay/src/main/java/mozilla/components/service/fxrelay/FxRelay.kt
7 mozilla.components.service.fxrelay.FxRelayImpl$fetchProfile$2 invoke mobile/android/android-components/components/service/firefox-relay/src/main/java/mozilla/components/service/fxrelay/FxRelay.kt
8 kotlinx.coroutines.intrinsics.UndispatchedKt startUndispatched
9 kotlinx.coroutines.BuildersKt__Builders_commonKt withContext
Filed because this signature's crash volume spiked. 11 reports from 3 distinct installations on the nightly build of 2026-10-15 (buildid 20261015093624); none of the preceding 3 build-days had any, and no build in the 21 days before had any report of it. That is an appearance from zero over the loudest of the preceding build-days and of this signature's builds over the 21 days before (Clouseau's bar is 3x, and the Poisson excess of the day against that baseline is z = 5.52, bar 3.62). This signature is new on nightly: crash-stats has no report of it there before this build.
Clouseau analysis (automated -- nothing below was checked by a human; a claim it could not ground is deliberately absent):
Fenix nightly crashes in the parent process, here 3 s after launch, when the Rust relay client's fetchProfile() fails with something UniFFI turns into mozilla.appservices.relay.InternalException (normally a Rust panic). FxRelayImpl.handleRelayExceptions only catches RelayApiException, so the exception escapes the Dispatchers.IO coroutine and kills the app. This spike is almost certainly the crash already tracked as bug 2082335 under a new name. That spelling (relay.kt:40) has 29 reports, all on build 20261014215550. This one (relay.kt:46) has 11, all on the next build. In between, 07d2cc0a1254 bumped application-services, and relay.kt is generated code shipped in that package. Recommended: catch InternalException (or any Exception) in handleRelayExceptions and return the fallback, and ask application-services what makes fetch_profile panic in the 2026-10-14 nightly versions.
Possible path to the crash:
Mechanism (observed): a mozilla.appservices.relay.InternalException thrown by the generated uniffiCheckCallStatus from RelayClient.fetchProfile is not caught, and the process dies.
Path (observed from source and stack, pinned d546d4277bd0):
FxRelayImpl.fetchAccountDetailscallsfetchProfile()insidewithContext(Dispatchers.IO).fetchProfilecallsclient.fetchProfile()insidehandleRelayExceptions(FETCH_PROFILE, { null }).- The Rust call returns a failure status that
uniffiCheckCallStatusraises asInternalException, not as the declaredRelayApiExceptionerror type. handleRelayExceptionshas onlycatch (e: RelayApiException), so the exception escapes the coroutine and the parent process crashes.
Trigger (inferred, low): something in the application-services relay crate that Fenix nightly picked up around 2026-10-14 makes fetch_profile panic or fail internally, possibly on a server response it cannot handle. The panic message and the Rust change were not available to check.
Signature change (inferred, medium): the move from relay.kt:40 to relay.kt:46 comes from the regenerated relay.kt in application-services bump 07d2cc0a1254, which only changes the version pin.
Checked:
handleRelayExceptionscatches onlyRelayApiException;fetchProfilerunsclient.fetchProfile()inside it onDispatchers.IO. (mobile/android/android-components/components/service/firefox-relay/src/main/java/mozilla/components/service/fxrelay/FxRelay.kt@d546d4277bd0:117-143,192-200)- The stack goes
uniffiCheckCallStatus<-fetchProfile(relay.kt) <-FxRelay.ktinvokeSuspend/handleRelayExceptions<-withContext. Parent process, arm64, Android 33, uptime 3 s, 0% bit-flip or Raptor Lake share. (brief stack; crashstats report ff21a6ec-42fb-4f96-a115-0e5690261015) - The sibling spelling
...uniffiCheckCallStatus(relay.kt:40)has 29 nightly reports, all on build 20261014215550 (first seen there, searched over 364 days), all parent process and arm64. (socorro crash_stats and facets build_id, 30 days, signature relay.kt:40) - This spelling (
relay.kt:46) appears only on build 20261015093624, the build right after therelay.kt:40build. (brief: signature age and spike counts) - 07d2cc0a1254, in the spiking build's window, only changes the application-services version pin (160.20261014050201 -> 160.20261014160830, revision ec31dc5b -> 7afac3d5). (mcp__patch__diff 07d2cc0a1254 (ApplicationServices.kt, moz.yaml))
- The previous application-services bump, 9e71cc00bd25 (bug 2081627, to ec31dc5b), landed 2026-10-14 and is the closest change to the onset build 20261014215550. (history file_history ApplicationServices.kt)
- The two spellings are the same crash; the line number changed because the generated
relay.ktwas rebuilt in the 07d2cc0a1254 bump. [inferred, medium confidence] (same frames and exception class, back-to-back builds, 07d2cc0a1254 diff) - Bug 2082335 (Firefox for Android :: Privacy, NEW) tracks the
relay.kt:40spelling. (bugzilla 2082335)
Alternatives considered:
- Bug 1977879 / c0ac317234cf (the pipeline's candidate, which created the component) as the cause of this onset: disfavored -- the missing
InternalExceptionhandling is the defect that makes this a crash, but neither spelling has reports before build 20261014215550 in 30 days of data, so something newer set it off. - A new defect introduced in the 2026-10-15 pushlog: disfavored -- the same crash was already happening as
relay.kt:40on the previous build; the only relevant change in the window is the application-services version pin. - Failing hardware: not supported -- 0% bit-flip and 0% Raptor Lake; Android arm64 Java exception.
Worth checking first:
- What makes Rust
fetch_profileraise an internal error or panic? Read the application-services commits between f22774fb and ec31dc5b (and up to 7afac3d5) in the relay crate, and get the panic message from theJavaExceptionannotation or from logcat. - Is the 2026-10-14 bump 9e71cc00bd25 the first build with it? Check that build 20261014215550 is the first nightly containing 9e71cc00bd25, and look for related relay signatures on earlier builds.
- Confirm the line-shift explanation: compare the generated
relay.ktin application-services ec31dc5b and 7afac3d5, or confirm that later builds report only therelay.kt:46spelling. - Defensive fix to land in android-components: in
FxRelayImpl.handleRelayExceptions, catchmozilla.appservices.relay.InternalException(or a genericException), log it and returnfallback(). Add a unit test with aRelayClientInterfacefake that throwsInternalException.
Filed automatically by Clouseau, which analyses nightly crashes with an LLM. Nothing above was written or checked by a human. Please close it as INVALID if it is wrong — that is useful feedback, not a nuisance.
This is not an independent Mozilla discovery: the crash is known here only because somebody submitted the crash report linked at the top. If this duplicates an existing report, please resolve THIS bug as the duplicate and leave the credit with the earlier reporter.
Updated•1 month ago
|
Updated•1 month ago
|
Description
•