Today's hypothesis is that inspector tests may be slow on low-spec bots because they have to synchronously load all of Test.html's resources individually from slow spinning hard drives. Let's try concatenating them (not minifying) to see if there's any significant difference.
<rdar://problem/22548705>
Created attachment 260463 [details] Proposed Fix Does not apply, depends on jsmin update and bugfix.
Comment on attachment 260463 [details] Proposed Fix Attachment 260463 [details] did not pass mac-ews (mac): Output: http://webkit-queues.webkit.org/results/134765 New failing tests: inspector/dom/pseudo-element-dynamic.html inspector/css/stylesheet-events-inspector-stylesheet.html inspector/dom/setOuterHTML.html inspector/model/remote-object-fake-object.html inspector/dom/template-content.html inspector/css/stylesheet-events-basic.html inspector/dom/getOuterHTML.html inspector/css/stylesheet-events-imports.html inspector/protocol/inspector-backend-invocation-return-value.html inspector/console/command-line-api.html http/tests/inspector/dom/disconnect-dom-tree-after-main-frame-navigation.html inspector/dom/pseudo-element-static.html inspector/css/stylesheet-events-multiple-documents.html
Created attachment 260468 [details] Archive of layout-test-results from ews101 for mac-mavericks The attached test failures were seen while running run-webkit-tests on the mac-ews. Bot: ews101 Port: mac-mavericks Platform: Mac OS X 10.9.5
Comment on attachment 260463 [details] Proposed Fix Attachment 260463 [details] did not pass mac-wk2-ews (mac-wk2): Output: http://webkit-queues.webkit.org/results/134777 New failing tests: inspector/dom/pseudo-element-dynamic.html inspector/dom/template-content.html inspector/model/remote-object-fake-object.html inspector/indexeddb/requestDatabaseNames.html inspector/dom/setOuterHTML.html inspector/css/stylesheet-events-basic.html inspector/dom/getOuterHTML.html inspector/css/stylesheet-events-imports.html inspector/css/stylesheet-events-inspector-stylesheet.html inspector/protocol/inspector-backend-invocation-return-value.html inspector/console/command-line-api.html http/tests/inspector/dom/disconnect-dom-tree-after-main-frame-navigation.html inspector/dom/pseudo-element-static.html inspector/css/stylesheet-events-multiple-documents.html
Created attachment 260469 [details] Archive of layout-test-results from ews106 for mac-mavericks-wk2 The attached test failures were seen while running run-webkit-tests on the mac-wk2-ews. Bot: ews106 Port: mac-mavericks-wk2 Platform: Mac OS X 10.9.5
Comment on attachment 260463 [details] Proposed Fix Why the Mac failures?
(In reply to comment #7) > Comment on attachment 260463 [details] > Proposed Fix > > Why the Mac failures? There are diffs caused by template strings getting minified wrong. Should not be a problem when dependencies land.
Created attachment 260657 [details] Proposed fix, no minifying This patch version just concats all test files together. Speed improvement was negligible (1-3s faster) on a newish imac, but might make a bigger difference on older hardware. Will test out on my fossilized macbook pro before landing.
On my old machine (2009 laptop with core 2 duo and 4GB RAM, but still faster than our build bots by a lot), this reduced inspector testing time from 1m50s to 1m30s (cold start) and from 1m15s to 1m on a warmed up file cache. Let's see what happens and roll out if it makes things worse.
SSD or hard drive?
Comment on attachment 260657 [details] Proposed fix, no minifying Clearing flags on attachment: 260657 Committed r189420: <http://trac.webkit.org/changeset/189420>
All reviewed patches have been landed. Closing bug.