Changed StartMouseInitiatedRemoveTabAnimation to use tab data indices rather than Model indices w...

Changed StartMouseInitiatedRemoveTabAnimation to use tab data indices rather than Model indices when adjusting mini-tab/non-mini-tab padding.

BUG=36504
TEST=See bug. Start Chrome and open a new tab. Pin the first one and then close it with a middle mouse click. Now look carefully at the left edge of the tab and open a new tab. The left most tab should not shift at all.
Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=91698

[SteveT, 2011-07-06 20:39:37]

Please have a look.

This change was necessary as the Tab Model was only seeing one tab in the
scenario where there is only one pinned tab and one normal tab. (It seems to no
longer see the tab that is closed) Because of that, model_index + 1 <
GetModelCount() is false and so the extra padding is not accounted for. Using
the tab index instead fixes this problem.

Let me know if I should note all of the above in a comment.

Thanks!


Steve

[sky, 2011-07-06 20:54:56]

LGTM

http://codereview.chromium.org/7307028/diff/1/chrome/browser/ui/views/tabs/ta...
File chrome/browser/ui/views/tabs/tab_strip.cc (right):

http://codereview.chromium.org/7307028/diff/1/chrome/browser/ui/views/tabs/ta...
chrome/browser/ui/views/tabs/tab_strip.cc:974: if (tab_closing->data().mini &&
next_tab_data_index < tab_count() &&
Insert:
DCHECK_NE(next_tab_index, tab_count())
between 973 and 974.

[sky, 2011-07-06 21:01:57]

One comment. The bug mentions all platforms. I'm not sure if that is correct,
but you should verify that.

  -Scott

[SteveT, 2011-07-06 21:15:31]

On 2011/07/06 21:01:57, sky wrote:
> One comment. The bug mentions all platforms. I'm not sure if that is correct,
> but you should verify that.
> 
>   -Scott

Yeah I did take a look at mac and linux_gtk and they do not appear to repro
there. It does repro on linux_chromeos.

Thanks for the look. Will make your change and commit once I get some green
trys.