[Bug] DCHECK fires in telemetry_perf_unittests |
||
Issue descriptionDCHECK fires when I run the telemetry tests: autoninja -C out/Android chrome_public_apk && CHROMIUM_OUTPUT_DIR=out/Android tools/perf/run_tests -v --browser=android-chromium --test-filter=benchmarks.system_health_smoke_test.SystemHealthBenchmarkSmokeTest.system_health.memory_mobile/load:tools:drive Stack trace: I 3.384s Main [FATAL:oauth2_token_service_delegate_android.cc(510)] Check failed: !account_name.empty() || account_id.empty(). Can't find account name, account_id=105787132709220979048' I 3.384s Main v------> base::debug::(anonymous namespace)::DebugBreak() ../../base/debug/debugger_posix.cc:229:5 I 3.384s Main 0014609d base::debug::BreakDebugger() ../../base/debug/debugger_posix.cc:263:0 I 3.384s Main 000e5093 logging::LogMessage::~LogMessage() ../../base/logging.cc:876:7 I 3.384s Main 00c23e07 OAuth2TokenServiceDelegateAndroid::MapAccountIdToAccountName(std::__ndk1::basic_string<char, std::__ndk1::char_traits<char>, std::__ndk1::allocator<char> > const&) const ../../chrome/browser/signin/oauth2_token_service_delegate_android.cc:510:3 I 3.385s Main 00c23ce1 OAuth2TokenServiceDelegateAndroid::RefreshTokenIsAvailable(std::__ndk1::basic_string<char, std::__ndk1::char_traits<char>, std::__ndk1::allocator<char> > const&) const ../../chrome/browser/signin/oauth2_token_service_delegate_android.cc:199:30 I 3.385s Main 00d237bf syncer::DetermineAccountToUse(identity::IdentityManager*, bool) ../../components/sync/driver/sync_auth_util.cc:36:27 I 3.385s Main 00ef162d autofill::PersonalDataManager::GetSyncSigninState() const ../../components/autofill/core/browser/personal_data_manager.cc:705:7 I 3.385s Main 00ecc545 autofill::AutofillManager::ShouldParseForms(std::__ndk1::vector<autofill::FormData, std::__ndk1::allocator<autofill::FormData> > const&, base::TimeTicks) ../../components/autofill/core/browser/autofill_manager.cc:325:50 I 3.385s Main 00ecb769 autofill::AutofillHandler::OnFormsSeen(std::__ndk1::vector<autofill::FormData, std::__ndk1::allocator<autofill::FormData> > const&, base::TimeTicks) ../../components/autofill/core/browser/autofill_handler.cc:70:8 I 3.385s Main 00858a11 autofill::mojom::AutofillDriverStubDispatch::Accept(autofill::mojom::AutofillDriver*, mojo::Message*) gen/components/autofill/content/common/autofill_driver.mojom.cc:783:13 I 3.385s Main 00016b47 mojo::InterfaceEndpointClient::HandleValidatedMessage(mojo::Message*) ../../mojo/public/cpp/bindings/lib/interface_endpoint_client.cc:423:32 I 3.385s Main 000167df mojo::FilterChain::Accept(mojo::Message*) ../../mojo/public/cpp/bindings/lib/filter_chain.cc:40:17 I 3.385s Main 000174d1 mojo::InterfaceEndpointClient::HandleIncomingMessage(mojo::Message*) ../../mojo/public/cpp/bindings/lib/interface_endpoint_client.cc:306:19 I 3.385s Main 000231a9 IPC::(anonymous namespace)::ChannelAssociatedGroupController::AcceptOnProxyThread(mojo::Message)
,
Dec 14
The following revision refers to this bug: https://chromium.googlesource.com/chromium/src.git/+/7887f7a41498776a05d84d88812d033ff43fb7a3 commit 7887f7a41498776a05d84d88812d033ff43fb7a3 Author: Mihai Sardarescu <msarda@chromium.org> Date: Fri Dec 14 17:43:08 2018 [android] Remove DCHECK from RefreshTokenIsAvailable. RefreshTokenIsAvailable is now called for accounts that are only present in the Gaia cookies and that were previously not seeded yet. This leads to the DCHECK in |OAuth2TokenServiceDelegateAndroid::MapAccountIdToAccountName()| being fired. This is unexpected as the code should simply return false in that case (not crash). This CL removes this DCHECK when MapAccountIdToAccountName is called from RefreshTokenIsAvailable. Bug: 915238 Change-Id: I863a822a5429f9aef54d78bcf39d573ebf647e84 Reviewed-on: https://chromium-review.googlesource.com/c/1378139 Commit-Queue: Mihai Sardarescu <msarda@chromium.org> Reviewed-by: Boris Sazonov <bsazonov@chromium.org> Cr-Commit-Position: refs/heads/master@{#616729} [modify] https://crrev.com/7887f7a41498776a05d84d88812d033ff43fb7a3/chrome/browser/signin/oauth2_token_service_delegate_android.cc
,
Jan 2
Mihai, can we close this now?
,
Jan 4
|
||
►
Sign in to add a comment |
||
Comment 1 by msarda@chromium.org
, Dec 14