Bug 172371 - Speedometer 2.0: stops at test 32
Summary: Speedometer 2.0: stops at test 32
Status: RESOLVED CONFIGURATION CHANGED
Alias: None
Product: WebKit
Classification: Unclassified
Component: Tools / Tests (show other bugs)
Version: WebKit Nightly Build
Hardware: Unspecified Unspecified
: P2 Normal
Assignee: Nobody
URL:
Keywords: InRadar
Depends on:
Blocks: 172339
  Show dependency treegraph
 
Reported: 2017-05-19 11:51 PDT by Benjamin Bouvier
Modified: 2017-05-23 12:48 PDT (History)
5 users (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description Benjamin Bouvier 2017-05-19 11:51:17 PDT
STR:
- clone latest webkit repository
- go to PerformancesTests/Speedometer
- serve (python -m SimpleHTTPServer) and visit in a browser

I'd expect the benchmark to finish, but it doesn't at the moment because test #32 tries to load /resources/todomvc/architecture-examples/preact/build/index.html, when the actual URI should be /resources/todomvc/architecture-examples/preact/dist/index.html (note the build => dist). It causes a 404 and the benchmark runner doesn't terminate.
Comment 1 Radar WebKit Bug Importer 2017-05-21 16:59:59 PDT
<rdar://problem/32319159>
Comment 2 Ryosuke Niwa 2017-05-21 17:56:49 PDT
Hm... I'm not seeing this.

https://trac.webkit.org/browser/webkit/trunk/PerformanceTests/Speedometer/resources/tests.js?rev=217198#L355
says:
todomvc/architecture-examples/preact/dist/index.html

which revision of WebKit did you check out?
Comment 3 Benjamin Bouvier 2017-05-22 02:13:16 PDT
I was using 4e56588e6e0b54becfc891fc5f111f61ed57f87c, but it seems the issue has been fixed in the meanwhile: running on trunk terminates now. Thanks!
Comment 4 Ryosuke Niwa 2017-05-22 14:53:11 PDT
Maybe you can shed a light on https://bugs.webkit.org/show_bug.cgi?id=172439. Firefox's score goes up by 40% on Speedometer 2 compared to 1, and we're afraid something might be broken in Firefox.
Comment 5 Benjamin Bouvier 2017-05-23 02:49:48 PDT
Sorry, I have no idea at this time. I've been running the benchmark and it doesn't seem anything goes wrong, since the benchmark properly terminates. Is there any error checking in this benchmark, i.e. checks against expected results?
Comment 6 Ryosuke Niwa 2017-05-23 12:48:10 PDT
No, there is no error checking. We'll investigate it further on https://bugs.webkit.org/show_bug.cgi?id=172439