[Webkit-unassigned] [Bug 48976] stepUp/stepDown for values in step-mismatching state for input elements

[bugzilla-daemon at webkit.org]

https://bugs.webkit.org/show_bug.cgi?id=48976





--- Comment #16 from Dai Mikurube <dmikurube at google.com>  2010-11-16 02:15:10 PST ---
(In reply to comment #15)
Thanks for comments and advices on the flag and style.


I found a worse result for the 0.00392156863.

1. <input type="number" min="0" value="0" step="0.00392156863" /> (max removed.)
2. stepUp() for 255 times. => 1.00000000065

This is an essential numeric error. 0.00392156863 * 255 is completely equal to 1.00000000065. I guess we can avoid it only with stepwise adjustment at stepUp(). The only reason to round 1.00000000065 is stepwise.

Another option is ignoring it. We can delegate such rounding-error handling to a page author, as you said "a page author have the responsibility of validity of the resultant values". It happens just on JavaScript stepUp()/stepDown() API. It doesn't happen on the renderer.

But it is less useful (stepUp()/stepDown() accumulates rounding errors) and non-compliant with the spec's example.

-- 
Configure bugmail: https://bugs.webkit.org/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug.