WordPress.org

Make WordPress Core

Changes between Initial Version and Version 2 of Ticket #34955


Ignore:
Timestamp:
12/10/2015 01:19:49 AM (4 years ago)
Author:
joemcgill
Comment:

Thanks for the report @smerriman.

This looks like it's similar to the issue brought up in #34931, and I agree we need a better way to test for ratios. If I recall correctly, we were using a method similar to the one you're suggesting early in the development of this feature and we ran into some issues. Let me read back through the history and see if I can remember what the issue was and then we can come up with a fix.

Legend:

Unmodified
Added
Removed
Modified
  • Ticket #34955

    • Property Status changed from new to assigned
    • Property Keywords needs-patch added
    • Property Milestone changed from Awaiting Review to 4.4.1
    • Property Owner set to joemcgill
    • Property Type changed from enhancement to defect (bug)
  • Ticket #34955 – Description

    initial v2  
    11In #34810:
    22
    3   As the difference between the widths being compared increases, the probability that rounding will introduce false positives also grows. We could make the aspect ratio variance we allow relative to the difference between the widths we're comparing, but the extra calculation is probably unnecessary unless the difference between image widths is in the tens of thousands.
     3> As the difference between the widths being compared increases, the probability that rounding will introduce false positives also grows. We could make the aspect ratio variance we allow relative to the difference between the widths we're comparing, but the extra calculation is probably unnecessary unless the difference between image widths is in the tens of thousands.
    44
    55This is not quite true. The issue does not arise when serving huge images with tens of thousands of pixels difference - but when serving small images.