Move chrome::HostDesktopType to ui::HostDesktopType

Move chrome::HostDesktopType to ui::HostDesktopType

Core tab restore code that is due to be componentized uses the
chrome::HostDesktopType enum. To accomodate that usage, this CL splits that
enum out of //chrome/browser/host_desktop.h into //ui/gfx/host_desktop_type.h.
The //chrome-level utility functions stay in //chrome/browser/host_desktop.h.

This CL does the move of the enum and namespace change. A followup CL will
change includes of //chrome/browser/host_desktop.h to
//ui/gfx/host_desktop_type.h where appropriate; that change is large due to
existing IWYU violations.

BUG=528880

[blundell, 2015-09-11 13:06:35]

blundell@chromium.org changed reviewers:
+ sky@chromium.org

[blundell, 2015-09-11 13:06:36]

Hi Scott,

This CL is large, but the only meaningful changes are quite small:

- //chrome/browser/ui/host_desktop.h
- //ui

The rest of the changes in //chrome are just renamespacing of the HostDesktop
enum from ::chrome to ::ui and "git cl format".

Note that as the CL description states, I've left include path changes in users
of HostDesktopType for a followup CL (that I'll send out shortly).

[sadrul, 2015-09-11 14:03:33]

Drive-by comment.

https://codereview.chromium.org/1336823003/diff/60001/ui/gfx/host_desktop_type.h
File ui/gfx/host_desktop_type.h (right):

https://codereview.chromium.org/1336823003/diff/60001/ui/gfx/host_desktop_typ...
ui/gfx/host_desktop_type.h:28: #endif
gfx shouldn't normally refer to layers above it (ash in this case). And
desktop-type is probably something we can get rid of soon
(https://groups.google.com/a/chromium.org/forum/#!msg/chromium-dev/PkYHtUn8NI8...).

[blundell, 2015-09-11 14:09:47]

https://codereview.chromium.org/1336823003/diff/60001/ui/gfx/host_desktop_type.h
File ui/gfx/host_desktop_type.h (right):

https://codereview.chromium.org/1336823003/diff/60001/ui/gfx/host_desktop_typ...
ui/gfx/host_desktop_type.h:28: #endif
On 2015/09/11 14:03:32, sadrul wrote:
> gfx shouldn't normally refer to layers above it (ash in this case). And
> desktop-type is probably something we can get rid of soon
>
(https://groups.google.com/a/chromium.org/forum/#!msg/chromium-dev/PkYHtUn8NI8...).

I'm not very familiar with the dependency hierarchy at this level, so if there's
a more sensible location for this enum (that //components can see), I'm happy to
move it.

Re: the email thread, I'm looking to componentize tab restore code now (within
the next two weeks or so).

[blundell, 2015-09-11 14:21:05]

On 2015/09/11 14:09:47, blundell wrote:
>
https://codereview.chromium.org/1336823003/diff/60001/ui/gfx/host_desktop_type.h
> File ui/gfx/host_desktop_type.h (right):
> 
>
https://codereview.chromium.org/1336823003/diff/60001/ui/gfx/host_desktop_typ...
> ui/gfx/host_desktop_type.h:28: #endif
> On 2015/09/11 14:03:32, sadrul wrote:
> > gfx shouldn't normally refer to layers above it (ash in this case). And
> > desktop-type is probably something we can get rid of soon
> >
>
(https://groups.google.com/a/chromium.org/forum/#!msg/chromium-dev/PkYHtUn8NI8...).
> 
> I'm not very familiar with the dependency hierarchy at this level, so if
there's
> a more sensible location for this enum (that //components can see), I'm happy
to
> move it.
> 
> Re: the email thread, I'm looking to componentize tab restore code now (within
> the next two weeks or so).

If there's nowhere lower than //chrome that really makes sense for this enum to
live, I could also just change the core tab restore code to pass around an
|is_ash| boolean or some such, and convert when coming from/back to //chrome.

[sky, 2015-09-11 16:09:36]

I don't think we should promote this enum either. Is there a way the
tabrestoreclient could provide an opaque value (int) that is static_cast'd to
the host_desktop_type? This way code in components doesn't need to know about
host-desktop-type.

[blundell, 2015-09-11 16:20:37]

On 2015/09/11 16:09:36, sky wrote:
> I don't think we should promote this enum either. Is there a way the
> tabrestoreclient could provide an opaque value (int) that is static_cast'd to
> the host_desktop_type? This way code in components doesn't need to know about
> host-desktop-type.

How about just replacing the usage of HostDesktopType inside the tab restore
code with an is_native_desktop bool?

[sky, 2015-09-11 19:50:14]

That's still rather chrome specific, right? I prefer my earlier suggestion.

  -Scott

On Fri, Sep 11, 2015 at 9:20 AM,  <blundell@chromium.org> wrote:
> On 2015/09/11 16:09:36, sky wrote:
>>
>> I don't think we should promote this enum either. Is there a way the
>> tabrestoreclient could provide an opaque value (int) that is static_cast'd
>> to
>> the host_desktop_type? This way code in components doesn't need to know
>> about
>> host-desktop-type.
>
>
> How about just replacing the usage of HostDesktopType inside the tab restore
> code with an is_native_desktop bool?
>
> https://codereview.chromium.org/1336823003/

To unsubscribe from this group and stop receiving emails from it, send an email
to chromium-reviews+unsubscribe@chromium.org.

[blundell, 2015-09-14 07:58:22]

On 2015/09/11 19:50:14, sky wrote:
> That's still rather chrome specific, right? I prefer my earlier suggestion.

OK! Closing this patch. Should have reached out to consult before embarking on
the CL given my lack of experience in //ui.

> 
>   -Scott
> 
> On Fri, Sep 11, 2015 at 9:20 AM,  <mailto:blundell@chromium.org> wrote:
> > On 2015/09/11 16:09:36, sky wrote:
> >>
> >> I don't think we should promote this enum either. Is there a way the
> >> tabrestoreclient could provide an opaque value (int) that is static_cast'd
> >> to
> >> the host_desktop_type? This way code in components doesn't need to know
> >> about
> >> host-desktop-type.
> >
> >
> > How about just replacing the usage of HostDesktopType inside the tab restore
> > code with an is_native_desktop bool?
> >
> > https://codereview.chromium.org/1336823003/
> 
> To unsubscribe from this group and stop receiving emails from it, send an
email
> to mailto:chromium-reviews+unsubscribe@chromium.org.