Closed Bug 930894 Opened 11 years ago Closed 9 years ago

Intermittent ico-bmp-corrupted/wrapper.html?invalid-compression-RLE4.ico | image comparison (==), max difference: 255, number of differing pixels: 4

Categories

(Core :: Graphics: ImageLib, defect)

Product:
Core
Component:
Graphics: ImageLib
Platform:
x86
Linux
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

()

Status:
RESOLVED WORKSFORME

People

(Reporter: cbook, Unassigned)

References

(
URL
)

Details

(Keywords: intermittent-failure)

Attachments

(2 files)

test screenshot
(deleted), image/png
Details
reference screenshot
(deleted), image/png
Details 
b2g_emulator b2g-inbound opt test reftest-1 on 2013-10-24 22:13:28 PDT for push c33d3181d2c5

slave: talos-r3-fed-032

https://tbpl.mozilla.org/php/getParsedLog.php?id=29640966&tree=B2g-Inbound


REFTEST TEST-UNEXPECTED-FAIL | http://10.0.2.2:8888/tests/image/test/reftest/ico/ico-bmp-corrupted/wrapper.html?invalid-compression-RLE4.ico | image comparison (==), max difference: 255, number of differing pixels: 4

10-25 05:38:02.030   660   660 I Gecko   : REFTEST TEST-UNEXPECTED-FAIL | http://10.0.2.2:8888/tests/image/test/reftest/ico/ico-bmp-corrupted/wrapper.html?invalid-compression-RLE4.ico | image comparison (==), max difference: 255, number of differing pixels: 4

R
Attached image test screenshot (deleted) — 

    
    

    
  

      
      
      
      

        Attached image
          
        reference screenshot (deleted)
        — 
      

    


  

    
    

    
  
So the test has some little dot on it and the ref is empty. Any idea what might be going on, Seth? :)
Flags: needinfo?(seth)

    
    

    
  
Possibly related to bug 921207?

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
  
David, we have a number of B2G reftest failures like these involving the various "corrupted" image reftests. Can we please find someone who can look into them?
Flags: needinfo?(dbaron)

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
  
Ryan, I'm looking into another variation of this in bug 921207. I'm going to leave 'needinfo' on until I manage to cook up a patch. I suspect all of the variations will be fixed in the same way. (Which hopefully does not involve converting them into mochitests yet again, but we can go there if needed.)

    
    

    
      
  

    
    

    
      
  

    
    

    
  
One theory is that bug 992324 might help.

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
  
comment 98 says that bug 992324 didn't fix this

    
    

    
  
I think it's interesting that the failure mode is always the same:

The test has:
 * a black (#000000) pixel at (8,21)
 * three gray (#919191) pixels at (9,21), (8,22), and (9,22)
 * a scrollbar
whereas the reference has only the scrollbar.

(Is the normal pass mode both having only the scrollbar?)

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
  
Looking into the (a) normal behavior (b) the invalidation sequence in both failure and normal cases might be informative.  I won't have time before mid-June, but others might.

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
  
Summary: TEST-UNEXPECTED-FAIL | tests/image/test/reftest/ico/ico-bmp-corrupted/wrapper.html?invalid-compression-RLE4.ico | image comparison (==), max difference: 255, number of differing pixels: 4 → Intermittent ico-bmp-corrupted/wrapper.html?invalid-compression-RLE4.ico | image comparison (==), max difference: 255, number of differing pixels: 4

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  
Depends on: 1051530

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
      
  

    
    

    
  
Inactive; closing (see bug 1180138).
Status: NEW → RESOLVED
Closed: 9 years ago
Resolution: --- → WORKSFORME
Flags: needinfo?(seth)
Flags: needinfo?(dbaron)

          You need to log in
          before you can comment on or make changes to this bug.
        






  

    
  







  

    

      
Attachment

      

      
      
      
      
    

    

      

        

          

            
General

            

              Creator: 



            

            
Created: 

            
Updated: 

            
Size: