Pasted links opened with alt-enter were opened next to the current tab instea...

Pasted links opened with alt-enter were opened next to the current tab instead of at the end of the strip.  This was because the LINK transition type triggered the TabStripModel to apply heuristics about where to open the URL, even though all URLs opened from the address bar should bypass these heuristics.  There were already hooks on the low-level functions to bypass the heuristics, I just had to expose them one level higher.  This meant an expansion to one of the TabContentsDelegate function's argument list, hence the number of files touched.  (This seems like a good capability to expose anyway, though.)BUG=6797
TEST=Have multiple tabs in your tab strip.  Select the first tab, paste in a URL, and hit alt-enter.  The newly opened tab should appear at the far end of the strip.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=17432

[Ben Goodger (Google), 2009-05-31 00:57:52]

Adding brettw to the reviewers list since this modifies TCD to include knowledge
of tabstrip indices.

[brettw, 2009-05-31 04:27:43]

As far as I can tell, the only place you actually need this parameter is in
Browser::OpenCurrentlURL. I probably missed a place but with all the function
calls it is very difficult to follow what's happening.

If tabs and their friends aren't going to (and shouldn't) know about these
parameters, it shouldn't be on the TCDelegate. Can you make this change internal
to Browser?

[Peter Kasting, 2009-06-01 01:34:03]

On 2009/05/31 04:27:43, brettw wrote:
> As far as I can tell, the only place you actually need this parameter is in
> Browser::OpenCurrentURL.

Yes.

> Can you make this change internal
> to Browser?

Maybe... I guess what you're suggesting is to make the
TabContentsDelegate::OpenURLFromTab() implementation itself a shell around a
function with a signature like what I have now.  I can try that, sure.

[brettw, 2009-06-01 02:59:41]

On 2009/06/01 01:34:03, pkasting wrote:
> On 2009/05/31 04:27:43, brettw wrote:
> > As far as I can tell, the only place you actually need this parameter is in
> > Browser::OpenCurrentURL.
> 
> Yes.
> 
> > Can you make this change internal
> > to Browser?
> 
> Maybe... I guess what you're suggesting is to make the
> TabContentsDelegate::OpenURLFromTab() implementation itself a shell around a
> function with a signature like what I have now.  I can try that, sure.

Yes, exactly. You changed the TabContentsDelegate interface, which isn't really
"Browser." It's the "stuff TabContents needs to call to whoever embeds it."
Browser can implement that however it needs to internally that make it most
convenient.

[Peter Kasting, 2009-06-02 19:25:12]

OK, new much simpler snap uploaded.  I feel embarrassed that I didn't do it this
way in the first place.

[brettw, 2009-06-02 19:50:06]

LGTM. thanks!