Bug 81760

Summary: Unable to traverse from results.html into EFL test failures on master buildbot
Product: WebKit Reporter: Dominik Röttsches (drott) <d-r>
Component: Tools / TestsAssignee: Nobody <webkit-unassigned>
Status: RESOLVED FIXED    
Severity: Normal CC: cdumez, dbates, lforschler, rakuco, thorton, tmpsantos, wsiegrist
Priority: P2    
Version: 528+ (Nightly build)   
Hardware: Unspecified   
OS: Unspecified   

Description Dominik Röttsches (drott) 2012-03-21 04:22:00 PDT
When looking at the layout test results for the EFL buildbot it is impossible to browse down into the individual failures. 

For example looking at
http://build.webkit.org/results/EFL%20Linux%20Release/r111519%20(33826)/results.html
and then checking the first failure for "css1/box_properties/clear_float.html"

The webserver first reports a "301 Moved permanently" redirecting to "old-results", e.g. from
"http://build.webkit.org/results/EFL%20Linux%20Release/r111519%20(33826)/css1/box_properties/clear_float-pretty-diff.html" to
"http://build.webkit.org/old-results/EFL%20Linux%20Release/r111519%20(33826)/css1/box_properties/clear_float-pretty-diff.html"
and then fails with 404 Not Found.

http://build.webkit.org/results/EFL%20Linux%20Release/r111519%20(33826).zip
Contains the build results for that particular run.
When downloading this to a mac with the per-user http enabled and uncompressing it using the same "ditto" command that's used on the buildbot master, I don't see any permission problems. Browsing down into individual failure cases starting from results.html works fine when testing it on the local webserver.

Also, the stdio of the uncompressing run does not show anything unusual, files should be successfully placed into the filesystem.

Are there any rewrite-rules in place that would incorrectly redirect to old-results?
Comment 1 Lucas Forschler 2012-03-21 12:18:15 PDT
Tim, you were looking at this for the internal bots.  Do you have time to look at it for webkit?
Comment 2 Tim Horton 2012-03-21 13:29:02 PDT
(In reply to comment #1)
> Tim, you were looking at this for the internal bots.  Do you have time to look at it for webkit?

I don't believe this is related to what I was looking at (and also didn't solve), and I won't have a chance to return to that (or look at this) for a while, sorry!
Comment 3 Dominik Röttsches (drott) 2012-04-13 07:12:55 PDT
Same happens on the new EFL Debug Buildbot, connected in bug 82711. Could somebody take a look?
Comment 4 William Siegrist 2012-04-13 07:57:17 PDT
The "old-results" are currently not being decompressed due to IO limitations on the long-term storage. Its a known issue. Please use the zip files for the time being.  We're working on a fix but it'll take a while.
Comment 5 Thiago Marcos P. Santos 2012-05-01 22:43:25 PDT
*** Bug 85198 has been marked as a duplicate of this bug. ***
Comment 6 Thiago Marcos P. Santos 2012-06-13 01:10:58 PDT
(In reply to comment #4)
> The "old-results" are currently not being decompressed due to IO limitations on the long-term storage. Its a known issue. Please use the zip files for the time being.  We're working on a fix but it'll take a while.

William,

Any update on this issue? It makes the TestFailures [1] service pretty hard to use for EFL. If there is something we could help, let us know.

Cheers,

[1] http://build.webkit.org/TestFailures/
Comment 7 William Siegrist 2012-06-15 10:37:12 PDT
We are working on new hardware for all of webkit.org. We hope to move to the new hardware soon.
Comment 8 Thiago Marcos P. Santos 2012-06-28 23:38:49 PDT
Fixed after the recent hardware upgrade. Thanks guys for handling this.