Scenario: This is a test -------------- a| Actual result after backspace: This is a test| -------------- Expected behaviour: This is a test --------------|

Comment on attachment 61095 [details] [diff] [review] Patch v1 In editor some of us tend to stay away from using the combined pattern: - if (NS_FAILED(res)) return res; - if (!bCollapsed) return res; + if (NS_FAILED(res) || !bCollapsed) return res; since it has on occassion caused some very subtle bugs in the past dealing with an unexpected nsresult passed to the caller. I know that as it stands your changes are functionaly equivalent. I'm wondering if we should be returning NS_OK instead of res in the !bCollapsed and !nearNode cases. Having said that, the rest of the diff looks ok to me. Since I'm going on vacation, I'll give you a conditional sr=kin if you can get r=jfrancis.

Comment on attachment 61095 [details] [diff] [review] Patch v1 looks fine/r=jfrancis

Fixed.

Verified on build 2001121803.