Closed Bug 1330550 Opened 8 years ago Closed 8 years ago

stylo: Talos fail with "Could not find report in browser output"

Categories

(Testing :: Talos, defect, P3)

x86_64
Linux
defect

Tracking

(Not tracked)

RESOLVED WORKSFORME

People

(Reporter: shinglyu, Unassigned)

References

Details

Talos running on Stylo https://treeherder.mozilla.org/#/jobs?repo=try&revision=cc113d4a2d9acd3eb7cd30542bf1f079e7ae5705&selectedJob=66082395 g1, g2 and tp test failed with: Could not find report in browser output: [('tsformat', ('__start_report', '__end_report')), ('tpformat', ('__start_tp_report', '__end_tp_report'))]
Component: CSS Parsing and Computation → Talos
Product: Core → Testing
Will, since you are involved in the mail thread, could you take a look at this or redirect to the right person? Thank you.
Flags: needinfo?(wlachance)
Blocks: stylo-perf-test
No longer blocks: stylo
it looks like talos isn't running successfully due to errors in the browser. On g2 (damp test) I see this in the log: 02:51:49 INFO - PROCESS | 9209 | ERROR:geckoservo::glue: Unnecessary call to traverse_subtree 02:51:50 INFO - PROCESS | 9209 | stylo: skipping declaration without ParserContextExtraData 02:51:50 INFO - PROCESS | 9209 | ERROR:geckoservo::glue: Unnecessary call to traverse_subtree 02:51:50 INFO - PROCESS | 9209 | ERROR:geckoservo::glue: Unnecessary call to traverse_subtree 02:51:50 INFO - PROCESS | 9209 | console.error: 02:51:50 ERROR - PROCESS | 9209 | Message: TypeError: doc.documentElement is null 02:51:50 INFO - PROCESS | 9209 | Stack: 02:51:50 INFO - PROCESS | 9209 | supportsHighlighters@resource://gre/modules/commonjs/toolkit/loader.js -> resource://devtools/server/actors/inspector.js:2857:9 02:51:50 INFO - PROCESS | 9209 | handler@resource://gre/modules/commonjs/toolkit/loader.js -> resource://devtools/shared/protocol.js:1082:19 02:51:50 INFO - PROCESS | 9209 | onPacket@resource://gre/modules/commonjs/toolkit/loader.js -> resource://devtools/server/main.js:1752:15 02:51:50 INFO - PROCESS | 9209 | receiveMessage@resource://gre/modules/commonjs/toolkit/loader.js -> resource://devtools/shared/transport/transport.js:761:7 02:51:50 INFO - PROCESS | 9209 | 02:51:50 INFO - PROCESS | 9209 | supportsHighlighters@resource://gre/modules/commonjs/toolkit/loader.js -> resource://devtools/server/actors/inspector.js:2857:9 02:51:50 INFO - PROCESS | 9209 | handler@resource://gre/modules/commonjs/toolkit/loader.js -> resource://devtools/shared/protocol.js:1082:19 02:51:50 INFO - PROCESS | 9209 | onPacket@resource://gre/modules/commonjs/toolkit/loader.js -> resource://devtools/server/main.js:1752:15 02:51:50 INFO - PROCESS | 9209 | receiveMessage@resource://gre/modules/commonjs/toolkit/loader.js -> resource://devtools/shared/transport/transport.js:761:7 02:51:50 INFO - PROCESS | 9209 | ERROR:geckoservo::glue: Unnecessary call to traverse_subtree 02:51:50 INFO - PROCESS | 9209 | ERROR:geckoservo::glue: Unnecessary call to traverse_subtree 02:51:50 INFO - PROCESS | 9209 | Assertion failure: mSource.IsGeckoRuleNode(), at /home/worker/workspace/build/src/layout/style/nsStyleContext.h:315 02:51:50 INFO - PROCESS | 9209 | #01: ???[/builds/slave/test/build/application/firefox/libxul.so +0xa258ce] 02:51:50 INFO - PROCESS | 9209 | ExceptionHandler::GenerateDump cloned child 9330 02:51:50 INFO - PROCESS | 9209 | ExceptionHandler::SendContinueSignalToChild sent continue signal to child 02:51:50 INFO - PROCESS | 9209 | ExceptionHandler::WaitForContinueSignal waiting for continue signal... 02:51:50 INFO - Terminating psutil.Process(pid=9209, name='firefox') 02:51:51 INFO - TEST-INFO | 9209: exit 11 the assertion is probably the cause of the browser terminating and not producing the results we want. In tp and g1, I see a hundreds of these message |ERROR:geckoservo::glue: Unnecessary call to traverse_subtree|. Talos expect output and progress at reasonable intervals (usually within 5 seconds) or it will quit. I would run this locally and compare firefox vs stylo- that seems like the best bet for watching what is going on.
Flags: needinfo?(wlachance)
I agree with Joel's advice of running locally, though I would personally start with a conceptually simpler test, tp5o (find it under the "tp" symbol in Treeherder) is probably my favorite (https://wiki.mozilla.org/Buildbot/Talos/Tests#tp5). Looking there, I see some pretty weird output before it fails: 02:50:49 INFO - TEST-INFO | started process 29436 (/builds/slave/test/build/application/firefox/firefox -profile /tmp/tmp83QyoG/profile http://localhost:52623/getInfo.html) 1146 02:50:50 INFO - PROCESS | 29436 | 1483527050288 addons.xpi WARN Exception running bootstrap method install on shield-recipe-client@mozilla.org: [Exception... "Component returned failure code: 0x8000ffff (NS_ERROR_UNEXPECTED) [nsIPrefBranch.getBoolPref]" nsresult: "0x8000ffff (NS_ERROR_UNEXPECTED)" location: "JS frame :: resource://gre/modules/addons/XPIProvider.jsm -> jar:file:///builds/slave/test/build/application/firefox/browser/features/shield-recipe-client@mozilla.org.xpi!/bootstrap.js :: this.install :: line 38" data: no] Stack trace: this.install()@resource://gre/modules/addons/XPIProvider.jsm -> jar:file:///builds/slave/test/build/application/firefox/browser/features/shield-recipe-client@mozilla.org.xpi!/bootstrap.js:38 < callBootstrapMethod()@resource://gre/modules/addons/XPIProvider.jsm:4986 < processFileChanges()@resource://gre/modules/addons/XPIProvider.jsm -> resource://gre/modules/addons/XPIProviderUtils.js:2097 < checkForChanges()@resource://gre/modules/addons/XPIProvider.jsm:3827 < startup()@resource://gre/modules/addons/XPIProvider.jsm:2831 < callProvider()@resource://gre/modules/AddonManager.jsm:264 < _startProvider()@resource://gre/modules/AddonManager.jsm:771 < startup()@resource://gre/modules/AddonManager.jsm:957 < startup()@resource://gre/modules/AddonManager.jsm:2923 < observe()@resource://gre/components/addonManager.js:65 1147 02:50:55 INFO - PROCESS | 29436 | [Parent 29436] WARNING: pipe error (66): Connection reset by peer: file /home/worker/workspace/build/src/ipc/chromium/src/chrome/common/ipc_channel_posix.cc, line 346 1148 02:50:55 INFO - PROCESS | 29436 | 1149 02:50:55 INFO - PROCESS | 29436 | ###!!! [Parent][MessageChannel] Error: (msgtype=0x46001A,name=PContent::Msg_PreferenceUpdate) Channel error: cannot send/recv 1150 02:50:55 INFO - PROCESS | 29436 | 1151 02:50:55 INFO - PROCESS | 29436 | 1152 02:50:55 INFO - PROCESS | 29436 | ###!!! [Parent][MessageChannel] Error: (msgtype=0x2C0084,name=PBrowser::Msg_Destroy) Channel error: cannot send/recv 1153 02:50:55 INFO - PROCESS | 29436 | 1154 02:50:55 INFO - PROCESS | 29436 | ERROR:geckoservo::glue: Unnecessary call to traverse_subtree 1155 02:50:55 INFO - PROCESS | 29436 | ERROR:geckoservo::glue: Unnecessary call to traverse_subtree 1156 02:50:55 INFO - PROCESS | 29436 | ERROR:geckoservo::glue: Unnecessary call to traverse_subtree 1157 02:51:53 INFO - PROCESS | 29436 | 1483527113649 addons.productaddons ERROR Request failed certificate checks: [Exception... "SSL is required and URI scheme is not https." nsresult: "0x8000ffff (NS_ERROR_UNEXPECTED)" location: "JS frame :: resource://gre/modules/CertUtils.jsm :: checkCert :: line 145" data: no] 1158 02:55:50 INFO - PROCESS | 29436 | ************************* 1159 02:55:50 INFO - PROCESS | 29436 | A coding exception was thrown and uncaught in a Task. 1160 02:55:50 INFO - PROCESS | 29436 | 1161 02:55:50 INFO - PROCESS | 29436 | Full message: ReferenceError: fetch is not defined 1162 02:55:50 INFO - PROCESS | 29436 | Full stack: apiCall@resource://shield-recipe-client/lib/NormandyApi.jsm:37:5 1163 02:55:50 INFO - PROCESS | 29436 | get@resource://shield-recipe-client/lib/NormandyApi.jsm:44:12 1164 02:55:50 INFO - PROCESS | 29436 | this.NormandyApi.fetchRecipes<@resource://shield-recipe-client/lib/NormandyApi.jsm:52:34 1165 02:55:50 INFO - PROCESS | 29436 | TaskImpl_run@resource://gre/modules/Task.jsm:319:42 1166 02:55:50 INFO - PROCESS | 29436 | TaskImpl@resource://gre/modules/Task.jsm:277:3 1167 02:55:50 INFO - PROCESS | 29436 | asyncFunction@resource://gre/modules/Task.jsm:252:14 1168 02:55:50 INFO - PROCESS | 29436 | this.RecipeRunner.start<@resource://shield-recipe-client/lib/RecipeRunner.jsm:64:23 1169 02:55:50 INFO - PROCESS | 29436 | TaskImpl_run@resource://gre/modules/Task.jsm:319:42 1170 02:55:50 INFO - PROCESS | 29436 | TaskImpl@resource://gre/modules/Task.jsm:277:3 1171 02:55:50 INFO - PROCESS | 29436 | asyncFunction@resource://gre/modules/Task.jsm:252:14 1172 02:55:50 INFO - PROCESS | 29436 | setTimeout_timer@resource://gre/modules/Timer.jsm:30:5 1173 02:55:50 INFO - PROCESS | 29436 | 1174 02:55:50 INFO - PROCESS | 29436 | ************************* 1175 1176 https://treeherder.mozilla.org/logviewer.html#?job_id=66082404&repo=try&lineNumber=1176
is this still a problem?
I run the tp5o test locally, the first few runs OK until chinaz.com, the one failed to report its number. I believe the report is generated from an Addon? Where can I find it and debug it?
Also
Depends on: 1329919
Found it: testing/talos/talos/pageloader/ but not sure where to start debugging. hmm
Is there any way I can manually trigger the pageloader addon and maybe debug the JS?
Flags: needinfo?(jmaher)
I found that the gDisableE10S parameter is not correctly set in the pageloader addon, even if I use a non-e10s test suite but the addon still use the e10s handlers. But fixing that doesn't fix this bug.
OK, after some debugging, the error is clearly about cnn.com crashing the stylo build. Block on Bug 1329919.
Flags: needinfo?(jmaher)
Depends on: 1337305
Priority: -- → P3
Summary: [Stylo] Talos fail with "Could not find report in browser output" → stylo: Talos fail with "Could not find report in browser output"
I am not sure this is a problem anymore, :shinglyu- can you confirm this is a problem or resolve this bug?
Flags: needinfo?(shing.lyu)
https://treeherder.mozilla.org/#/jobs?repo=mozilla-central&filter-searchStr=stylo&selectedJob=104379964 All the Talos test on stylo platform seems to be running successfully now. Resolving this bug. Thanks for the reminder, Joel.
Flags: needinfo?(shing.lyu)
Status: NEW → RESOLVED
Closed: 8 years ago
Resolution: --- → WORKSFORME
You need to log in before you can comment on or make changes to this bug.