642823 - Make nsISelection::Modify consistent on word granularity selection modifications

Status: RESOLVED FIXED

(Core :: DOM: Selection, defect)

Description

[(no longer active)]

Currently, the behavior of nsISelection::Modify depends on the layout.word_select.eat_space_to_next_word pref.  I don't think that makes a lot of sense, because the results of calling the API from the web page would depend on the OS.

I'm planning to change this behavior to make it more consistent.  I think assuming eat_space_to_next_word being false is the saner of the two approaches.

Comment 1

[(no longer active)]

Attachment: Patch (v1)

The webkit implementation also does this.

Comment 2

[Robert O'Callahan (:roc) (email my personal email if necessary)]

Comment on attachment 520250 [details] [diff] [review]
Patch (v1)

I guess this would be easy to test?

Comment 3

[(no longer active)]

Attachment: Part 2: adjust the behavior of the Windows-specific test

(In reply to comment #2)
> Comment on attachment 520250 [details] [diff] [review]
> Patch (v1)
> 
> I guess this would be easy to test?

Yes.  The richtext2 suite actually tests this, but I also found the test for bug 496275 which explicitly tested for the Windows specific behavior.  This patch removes that check, which makes us effectively test the new behavior of selection.modify across all platforms.

Comment 4

[(no longer active)]

Attachment: Part 2: adjust the behavior of the Windows-specific test

With the actual contents of the patch!

Comment 5

[Ojan Vafai]

I guess this is a question of use-cases. The original intent of this API was explicitly to expose the platform-specific selection behavior so that web apps could appropriately mimic it. Also, isn't this just choosing one platform's behavior other the other?

Maybe it's worth bringing this up on whatwg@? If we're going to deviate from platform-specific behavior, we should probably pick something consistent with the rest of the web platform as per http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2011-January/030120.html.

Comment 6

[(no longer active)]

http://hg.mozilla.org/mozilla-central/rev/d09549b1b664
http://hg.mozilla.org/mozilla-central/rev/57770167ce80

Comment 7

[(no longer active)]

(In reply to comment #5)
> I guess this is a question of use-cases. The original intent of this API was
> explicitly to expose the platform-specific selection behavior so that web apps
> could appropriately mimic it. Also, isn't this just choosing one platform's
> behavior other the other?
> 
> Maybe it's worth bringing this up on whatwg@? If we're going to deviate from
> platform-specific behavior, we should probably pick something consistent with
> the rest of the web platform as per
> http://lists.whatwg.org/htdig.cgi/whatwg-whatwg.org/2011-January/030120.html.

I'll post about this to whatwg.  Currently our behavior matches that of Webkit, and if it is decided that this needs to be platform specific, we can revert this change.

Comment 8

[Eric Shepherd [:sheppy]]

This change has been noted here:

https://developer.mozilla.org/en/DOM/Selection/modify

Also mentioned on Firefox 5 for developers.

Documented as shipping in Firefox 5. Please advise if this is not correct.