Move BrokerDuplicateHandle() to be declared in content/public

Move BrokerDuplicateHandle() to be declared in content/public

This is necessary so that BrokerDuplicateHandle() can be used from
chrome/browser while passing the check_deps rules.

BUG=http://code.google.com/p/nativeclient/issues/detail?id=2719
TEST=build

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=132405

[Mark Seaborn, 2012-04-13 20:30:48]

This depends on https://chromiumcodereview.appspot.com/9958034 being re-applied
(since it was recently reverted in
https://chromiumcodereview.appspot.com/10081018).

[jschuh, 2012-04-14 19:59:53]

lgtm if @jam is cool with the header filename.

https://chromiumcodereview.appspot.com/10082018/diff/1/content/common/sandbox...
File content/common/sandbox_policy.cc (right):

https://chromiumcodereview.appspot.com/10082018/diff/1/content/common/sandbox...
content/common/sandbox_policy.cc:658: namespace content {
You're going to have to fix this up manually now that I've landed r132303
(sorry).

https://chromiumcodereview.appspot.com/10082018/diff/1/content/public/common/...
File content/public/common/sandbox_init.h (right):

https://chromiumcodereview.appspot.com/10082018/diff/1/content/public/common/...
content/public/common/sandbox_init.h:35: CONTENT_EXPORT bool
BrokerDuplicateHandle(HANDLE source_handle,
I hate to nit, but does sandbox_init.h really make sense for this? it seems like
it should be in it's own header. Maybe sandbox_win.h. Although, if @jam is cool
with it then I defer to him.

[Mark Seaborn, 2012-04-14 23:34:26]

On 14 April 2012 12:59, <jschuh@chromium.org> wrote:

> lgtm if @jam is cool with the header filename.
>
>
> https://chromiumcodereview.**appspot.com/10082018/diff/1/**
>
content/common/sandbox_policy.**cc<https://chromiumcodereview.appspot.com/10082018/diff/1/content/common/sandbox_policy.cc>
> File content/common/sandbox_policy.**cc (right):
>
> https://chromiumcodereview.**appspot.com/10082018/diff/1/**
>
content/common/sandbox_policy.**cc#newcode658<https://chromiumcodereview.appspot.com/10082018/diff/1/content/common/sandbox_policy.cc#newcode658>
> content/common/sandbox_policy.**cc:658: namespace content {
> You're going to have to fix this up manually now that I've landed
> r132303 (sorry).
>

I didn't get any conflicts when I merged -- the change was already based on
the earlier version of r132303.

https://chromiumcodereview.**appspot.com/10082018/diff/1/**
>
content/public/common/sandbox_**init.h<https://chromiumcodereview.appspot.com/10082018/diff/1/content/public/common/sandbox_init.h>
> File content/public/common/sandbox_**init.h (right):
>
> https://chromiumcodereview.**appspot.com/10082018/diff/1/**
>
content/public/common/sandbox_**init.h#newcode35<https://chromiumcodereview.appspot.com/10082018/diff/1/content/public/common/sandbox_init.h#newcode35>
> content/public/common/sandbox_**init.h:35: CONTENT_EXPORT bool
> BrokerDuplicateHandle(HANDLE source_handle,
> I hate to nit, but does sandbox_init.h really make sense for this? it
> seems like it should be in it's own header. Maybe sandbox_win.h.
> Although, if @jam is cool with it then I defer to him.
>

That's what I said, but jam preferred reusing sandbox_init.h to avoid
creating more headers:

jam: you'll need to declare this function in a header in content/public if
you want to use it from chrome
jam: i suggest content/public/common/sandbox_init.h
me: do i just move the function prototype to content/public, and leave the
definition in content/common/sandbox_policy.cc?
me: creating a new header file, e.g. sandbox_support_win.h, seems more
appropriate than using sandbox_init.h, because this is windows-specific and
it's not related to initialisation. does that sound OK?
jam: it's still used for startup
jam: we try to keep the headers in content/public to a minimum where it
makes sense
jam: since this can be considered as (nacl) process startup, i think it's
fine to put it in that file
me: hmm, it's not just used for startup of nacl, though
jam: ok if you want we can maybe rename that file to sandbox_util.h so
it';s more generic
jam: i'd just prefer not to add another file in content/public just for one
function
me: ok, will reuse the header

Mark

[jam, 2012-04-15 21:40:55]

lgtm with nit

http://codereview.chromium.org/10082018/diff/6003/content/public/common/sandb...
File content/public/common/sandbox_init.h (right):

http://codereview.chromium.org/10082018/diff/6003/content/public/common/sandb...
content/public/common/sandbox_init.h:35: CONTENT_EXPORT bool
BrokerDuplicateHandle(HANDLE source_handle,
nit: please add a comment here

[Mark Seaborn, 2012-04-16 15:32:44]

http://codereview.chromium.org/10082018/diff/6003/content/public/common/sandb...
File content/public/common/sandbox_init.h (right):

http://codereview.chromium.org/10082018/diff/6003/content/public/common/sandb...
content/public/common/sandbox_init.h:35: CONTENT_EXPORT bool
BrokerDuplicateHandle(HANDLE source_handle,
On 2012/04/15 21:40:55, John Abd-El-Malek wrote:
> nit: please add a comment here

OK, comment added.