Add permissions buts for Pepper plugins.

Add permissions buts for Pepper plugins.

This patch doesn't actually hook anything up, but it plumbs them in for the laces we'll need it.

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=146519

[brettw, 2012-07-07 05:08:48]

Here is a basic permissions checking class for Pepper so I can implement "user
gesture" checking in my other patch you pointed out.

[bbudge, 2012-07-07 22:26:21]

https://chromiumcodereview.appspot.com/10735011/diff/2001/ppapi/shared_impl/p...
File ppapi/shared_impl/ppapi_permissions.h (right):

https://chromiumcodereview.appspot.com/10735011/diff/2001/ppapi/shared_impl/p...
ppapi/shared_impl/ppapi_permissions.h:23: };
I get the feeling we'll quickly have more than 32 permissions, since there will
be API permissions. Perhaps a non-flag enum would be better. That makes it a bit
of a pain to initialize, but I'm not sure how much of a pain it would be in
practice, especially since permissions would be set based on parsing a manifest
file. Or do you have another idea in mind for API permissions?

[brettw, 2012-07-08 03:24:25]

https://chromiumcodereview.appspot.com/10735011/diff/2001/ppapi/shared_impl/p...
File ppapi/shared_impl/ppapi_permissions.h (right):

https://chromiumcodereview.appspot.com/10735011/diff/2001/ppapi/shared_impl/p...
ppapi/shared_impl/ppapi_permissions.h:23: };
I'm actually more optimistic about this being enough bits. We only have 24
permissions for Chrome extensions which probably does more than we will support.
And one bit will likely do much more than one interface assumign we follow a
similar model:
http://code.google.com/chrome/extensions/manifest.html#permissions
(for example, "direct socket access" would enable both TCP and UDP sockets, and
I would imagine also unlock all http origin and header restrictions (since you
could do http over sockets yourself if you wanted)).

In any case, I'm trying to get "something" so we can make progress and this
seemed like a good start. It should be easy to change this later should the need
arise since there are only a few places where these are initialized: the
checkers of permissions, of which there will be a lot, don't know about the
implementation).

[bbudge, 2012-07-08 15:10:12]

LGTM

[commit-bot: I haz the power, 2012-07-08 22:57:07]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/brettw@chromium.org/10735011/2001

[commit-bot: I haz the power, 2012-07-08 22:57:13]

Presubmit check for 10735011-2001 failed and returned exit status 1.

Running presubmit commit checks ...

** Presubmit Messages **
If this change has an associated bug, add BUG=[bug number].

If this change requires manual test instructions to QA team, add
TEST=[instructions].

** Presubmit ERRORS **
Missing LGTM from an OWNER for files in these directories:
    chrome/browser/component_updater

Presubmit checks took 1.2s to calculate.

[dmichael (off chromium), 2012-07-09 14:41:35]

drive-by lgtm

http://codereview.chromium.org/10735011/diff/2001/ppapi/shared_impl/ppapi_per...
File ppapi/shared_impl/ppapi_permissions.cc (right):

http://codereview.chromium.org/10735011/diff/2001/ppapi/shared_impl/ppapi_per...
ppapi/shared_impl/ppapi_permissions.cc:19: return !!(permissions_ &
static_cast<uint32>(perm));
optional: You could DCHECK that perm has exactly one flag set (so that you can
make sure it's changeable in the future, if we need more than 32). I say this
partly because it would be fun to actually have a practical use for the old:
(x & (x - 1))
trick for detecting if x is a power of two. :-)

http://codereview.chromium.org/10735011/diff/2001/ppapi/shared_impl/ppapi_per...
File ppapi/shared_impl/ppapi_permissions.h (right):

http://codereview.chromium.org/10735011/diff/2001/ppapi/shared_impl/ppapi_per...
ppapi/shared_impl/ppapi_permissions.h:31: PpapiPermissions(uint32 perms);
explicit, please

[brettw, 2012-07-13 04:28:13]

TBR=cpu for component_updater