If an page is using background in an element, the reverse selection system is faild when the background color is not similar to background image. If the element's foreground color and background color is similar and existing background image, we should reverse the current reverse selection result.
Perhaps, but I wonder if exposing (if not already exposed, forgive me for not knowing image related api) a method such as HasTransparant in the Image would be worth it, and if there are transparant bits for the image, we could then assume that the author does intend the specified background color to be 'correct'. [and as such may be worth exchanging selection colors as currently designed]
Justin: If webpage author using transparent image, I think that the foreground color and background color are _not_ similar. So, I want to change the behavior when the colors are similar.
Good enough for me, but just for thought, in some of my personal designs I have done slight gradients (transparant/alpha-scale) where the actual specified background color is very close to what the image shows, though I can easily conceive that I am a minority in that use.
*** Bug 292660 has been marked as a duplicate of this bug. ***
Created attachment 192919 [details] [diff] [review] Patch rv1.0
*** Bug 294485 has been marked as a duplicate of this bug. ***
My patch doesn't work fine on http://www.uoguelph.ca/~stuartr/ This is very difficult problem. The page can look the text if we don't load images. So, the page's text color and background color conbination has sufficient contrast. On this case, we cannot fix with my approach.
I have an idea. But I don't like this way. If the page uses background image and it's spcified by other element(not equal background color specified element), we use text color for selection background color. If so, the text color and actual background color combination may have sufficient contrast. But I think that the looks is not good. # If background image and color specified same element, we trust the background color.
I won't be able to review this any time in the near future (at least 2+ weeks). I'd really appreciate it if you could find another reviewer...
Umm... I wrote test patch by comment 10. But it is not better than previous patch. Please ignore comment 10.
Comment on attachment 192919 [details] [diff] [review] Patch rv1.0 I cancel the review. Because I try to fix bug 170951 and bug 113161 before this.
*** Bug 345210 has been marked as a duplicate of this bug. ***
*** Bug 332612 has been marked as a duplicate of this bug. ***
http://www.ed.ac.uk/studying/ is another example (text in the white column on the right will select with a white background due to the blue bg colour of the page itself). Masayuki, now that the two bugs you mentioned in comment #13 are fixed, is there any chance you'll try to fix this one next? :-)
(In reply to comment #16) > http://www.ed.ac.uk/studying/ > > is another example (text in the white column on the right will select with a > white background due to the blue bg colour of the page itself). I'm not seeing a white column on the right here. Did the page change?