Move sanbox_export.h to //sandbox from //sandbox/linux and split root OWNERS file.

Move sanbox_export.h to //sandbox from //sandbox/linux and split root OWNERS file.

BUG=none
R=jln@chromium.org, jschuh@chromium.org

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=269852

[jln (very slow on Chromium), 2014-05-12 18:44:22]

lgtm

Justin: is that fine with you too? The main goal is to let the presumbit script
suggest a suitable owners for each platform.
We can both remain owners of the root of sandbox/ for administrative purpose and
rare stuff like sandbox_export.h

[jschuh, 2014-05-12 19:49:46]

I think the owners change should be split out into a separate CL, which is a
good place to narrow down the size of the list if appropriate.

[Robert Sesek, 2014-05-12 19:56:18]

On 2014/05/12 19:49:46, Justin Schuh wrote:
> I think the owners change should be split out into a separate CL, which is a
> good place to narrow down the size of the list if appropriate.

Dropped OWNERS

[jschuh, 2014-05-12 19:57:31]

lgtm, and if you want i can make the CL to split the owners. :)

[Robert Sesek, 2014-05-12 20:03:46]

Committed patchset #3 manually as r269852 (presubmit successful).

[rvargas (doing something else), 2014-05-12 21:56:34]

https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h
File sandbox/sandbox_export.h (right):

https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h...
sandbox/sandbox_export.h:9: #if defined(WIN32)
I don't think we should have this part. At the very least it is misleading as it
suggest that the win sandbox is a shared library for the component build.

[Robert Sesek, 2014-05-12 22:09:47]

https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h
File sandbox/sandbox_export.h (right):

https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h...
sandbox/sandbox_export.h:9: #if defined(WIN32)
On 2014/05/12 21:56:35, rvargas wrote:
> I don't think we should have this part. At the very least it is misleading as
it
> suggest that the win sandbox is a shared library for the component build.

It seemed wrong to move it to the parent directory without adding support to
Windows. I figured doing so would in fact be misleading.

[rvargas (doing something else), 2014-05-12 22:49:57]

https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h
File sandbox/sandbox_export.h (right):

https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h...
sandbox/sandbox_export.h:9: #if defined(WIN32)
On 2014/05/12 22:09:48, rsesek wrote:
> On 2014/05/12 21:56:35, rvargas wrote:
> > I don't think we should have this part. At the very least it is misleading
as
> it
> > suggest that the win sandbox is a shared library for the component build.
> 
> It seemed wrong to move it to the parent directory without adding support to
> Windows. I figured doing so would in fact be misleading.

Now I'm confused. Are you saying that this file should not be at the parent
directory?

[Robert Sesek, 2014-05-12 23:29:42]

https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h
File sandbox/sandbox_export.h (right):

https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h...
sandbox/sandbox_export.h:9: #if defined(WIN32)
On 2014/05/12 22:49:57, rvargas wrote:
> On 2014/05/12 22:09:48, rsesek wrote:
> > On 2014/05/12 21:56:35, rvargas wrote:
> > > I don't think we should have this part. At the very least it is misleading
> as
> > it
> > > suggest that the win sandbox is a shared library for the component build.
> > 
> > It seemed wrong to move it to the parent directory without adding support to
> > Windows. I figured doing so would in fact be misleading.
> 
> Now I'm confused. Are you saying that this file should not be at the parent
> directory?

No, it's needed by Mac and Linux. It didn't seem smart to move it to a parent
directory that also included Windows files, without at least adding support for
Windows, even if it is unused, because it's such a minimal amount of
boilerplate.

[rvargas (doing something else), 2014-05-13 00:49:26]

On 2014/05/12 23:29:42, rsesek wrote:
> https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h
> File sandbox/sandbox_export.h (right):
> 
>
https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h...
> sandbox/sandbox_export.h:9: #if defined(WIN32)
> On 2014/05/12 22:49:57, rvargas wrote:
> > On 2014/05/12 22:09:48, rsesek wrote:
> > > On 2014/05/12 21:56:35, rvargas wrote:
> > > > I don't think we should have this part. At the very least it is
misleading
> > as
> > > it
> > > > suggest that the win sandbox is a shared library for the component
build.
> > > 
> > > It seemed wrong to move it to the parent directory without adding support
to
> > > Windows. I figured doing so would in fact be misleading.
> > 
> > Now I'm confused. Are you saying that this file should not be at the parent
> > directory?
> 
> No, it's needed by Mac and Linux. It didn't seem smart to move it to a parent
> directory that also included Windows files, without at least adding support
for
> Windows, even if it is unused, because it's such a minimal amount of
> boilerplate.

hmm... then I still prefer this file to not have code for windows. It's not like
we _may_ use that code at some point; doing that would be clearly wrong so it's
not just a matter of unused (but correct) code.

[Robert Sesek, 2014-05-13 00:52:17]

On 2014/05/13 00:49:26, rvargas wrote:
> On 2014/05/12 23:29:42, rsesek wrote:
> >
https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h
> > File sandbox/sandbox_export.h (right):
> > 
> >
>
https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h...
> > sandbox/sandbox_export.h:9: #if defined(WIN32)
> > On 2014/05/12 22:49:57, rvargas wrote:
> > > On 2014/05/12 22:09:48, rsesek wrote:
> > > > On 2014/05/12 21:56:35, rvargas wrote:
> > > > > I don't think we should have this part. At the very least it is
> misleading
> > > as
> > > > it
> > > > > suggest that the win sandbox is a shared library for the component
> build.
> > > > 
> > > > It seemed wrong to move it to the parent directory without adding
support
> to
> > > > Windows. I figured doing so would in fact be misleading.
> > > 
> > > Now I'm confused. Are you saying that this file should not be at the
parent
> > > directory?
> > 
> > No, it's needed by Mac and Linux. It didn't seem smart to move it to a
parent
> > directory that also included Windows files, without at least adding support
> for
> > Windows, even if it is unused, because it's such a minimal amount of
> > boilerplate.
> 
> hmm... then I still prefer this file to not have code for windows. It's not
like
> we _may_ use that code at some point; doing that would be clearly wrong so
it's
> not just a matter of unused (but correct) code.

If Windows never intends to build sandbox as a <(component) then yes we should
remove this.

[rvargas (doing something else), 2014-05-13 01:10:05]

On 2014/05/13 00:52:17, rsesek wrote:
> On 2014/05/13 00:49:26, rvargas wrote:
> > On 2014/05/12 23:29:42, rsesek wrote:
> > >
> https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h
> > > File sandbox/sandbox_export.h (right):
> > > 
> > >
> >
>
https://codereview.chromium.org/273423007/diff/40001/sandbox/sandbox_export.h...
> > > sandbox/sandbox_export.h:9: #if defined(WIN32)
> > > On 2014/05/12 22:49:57, rvargas wrote:
> > > > On 2014/05/12 22:09:48, rsesek wrote:
> > > > > On 2014/05/12 21:56:35, rvargas wrote:
> > > > > > I don't think we should have this part. At the very least it is
> > misleading
> > > > as
> > > > > it
> > > > > > suggest that the win sandbox is a shared library for the component
> > build.
> > > > > 
> > > > > It seemed wrong to move it to the parent directory without adding
> support
> > to
> > > > > Windows. I figured doing so would in fact be misleading.
> > > > 
> > > > Now I'm confused. Are you saying that this file should not be at the
> parent
> > > > directory?
> > > 
> > > No, it's needed by Mac and Linux. It didn't seem smart to move it to a
> parent
> > > directory that also included Windows files, without at least adding
support
> > for
> > > Windows, even if it is unused, because it's such a minimal amount of
> > > boilerplate.
> > 
> > hmm... then I still prefer this file to not have code for windows. It's not
> like
> > we _may_ use that code at some point; doing that would be clearly wrong so
> it's
> > not just a matter of unused (but correct) code.
> 
> If Windows never intends to build sandbox as a <(component) then yes we should
> remove this.

yes, that is the case.