Inline spell checking frequently underlines the "First Name, Last Name wrote: " line when replying to messages. We should automatically add the first and last name for each email receipient to the ignore list for inline spell check.

This gets us all the way there. It handles the following cases: 1) the initial identity being used 2) if you reply to a msg, all of the initial recipients get added to the ignore list. Ditto for edit as draft. 3) dragging addresses to the compose window causes the addresses to get added 3) If you change your identity, the new identity gets added. Still to do: 1) I can't figure out how to track individual names being added to the auto complete widget. You'd be surprised how tricky this is to detect. Would be nice to have: 1) Would be nice if the spell checker dialog could somehow leverage this code to get things added to its ignore list too. It works aready for the dialog if you have inline spell checking also turned on. But if you are a dialog user only then this won't do anything for you.

Comment on attachment 173320 [details] [diff] [review] most of the fix read the previous comment for more details on this patch. I can't figure out an easy way to detect the user using autocomplete to add recipients without spell checking the document easily so I'm punting on that part for now.

I added calls to add recipient to ignore list that get triggered whenever a user hits enter or tabs out of an autocomplete row. Note: calling IgnoreWords triggers a re-spell of the document. When bringing up a compose window with existing field data, we do only 1 spell check. But when adding autocompleted addresses, we will spell check again every time you add a recipient.

IMO, it's probably not a good idea to add names to the ignore list as some names could be misspelled words. If you ignore all names, it would be possible for the user to not see an obviously misspelled word if that misspelled word happened to be a first name of one of their recipients.. Example: tim cotten as the recipient name then user types 'the shirt was made of cotten' cotten should be cotton really... but would NOT be marked as misspelled Would it not be a better idea to ignore the quote header line itself, as is done with the quote itself ?

*** Bug 282678 has been marked as a duplicate of this bug. ***