Keep installing all-users Start Menu and Desktop shortcuts on system-level installs.

Keep installing all-users Start Menu and Desktop shortcuts on system-level installs.

Also, pin the system-level shortcut if pinning was requested when creating the user-level shortcut, but it isn't created because its system-level match is present.

BUG=164655
TEST=
Below "all shortcuts" refers to: Desktop, Quick Launch, Start Menu (Start Screen on Win8), and taskbar shortcuts.

Have 5 users (setup this way BEFORE installing system-level Chrome -- might be worth to save a VM with this state before installing system-level Chrome...):
User A => M23 user-level Chrome installed (he will also be the user installing system-level Chrome)
User B => M23 user-level Chrome installed, but logged out.
User C => M23 user-level Chrome installed, but stays logged in.
User D => No user-level install.
User E => No user-level install.

From user A, install a system-level Chrome including this CL.

User A should have both types of shortcuts (per-user + all-users); invoking any of the per-user shortcuts should self-destruct the user-level Chrome (deleting its shortcuts in the process).

Login to User B:
He should also have both types shortcuts and self-destruct should happen when invoking user-level Chrome.

Go to user C (should have been logged in and stayed logged in since the system-level install):
User C will have both types of shortcuts (he will however be missing the system-level taskbar pin; and that until either next login or system-level first run). Self-destruct should happen the same way.

Login to user D:
He should have all system-level shortcuts.

**Delete the all-users Desktop shortcut.**

Login to user E:
He should have all system-level shortcuts (the Desktop shortcut will be a per-user shortcut pointing to system-level Chrome) :)! This is a nice side-effect of the previous work I did (will not work for users who later go through self-destruct or who have already ran system-level Chrome, but still a nice side-effect feature!).

Committed: https://src.chromium.org/viewvc/chrome?view=rev&revision=172223

[gab, 2012-12-07 19:29:18]

Sir, PTAL.

Cheers,
Gab

[robertshield, 2012-12-10 20:58:28]

lg, couple of nits

https://codereview.chromium.org/11465025/diff/1/chrome/installer/setup/instal...
File chrome/installer/setup/install.cc (right):

https://codereview.chromium.org/11465025/diff/1/chrome/installer/setup/instal...
chrome/installer/setup/install.cc:377: // TODO(tommi): Change this function to
use WorkItemList.
since we don't generally roll back on failure, you can remove this comment.

https://codereview.chromium.org/11465025/diff/1/chrome/installer/util/shell_u...
File chrome/installer/util/shell_util.cc (right):

https://codereview.chromium.org/11465025/diff/1/chrome/installer/util/shell_u...
chrome/installer/util/shell_util.cc:1274: }
clean up nit: we build and populate both user_shortcut_path and
system_shortcut_path but only ever use one of them. Would it not be better to
first decide which one we need, then build it? Doing so could considerably
shorten this method.

[gab, 2012-12-10 21:50:32]

Removed TODO, PTAL.

https://codereview.chromium.org/11465025/diff/1/chrome/installer/setup/instal...
File chrome/installer/setup/install.cc (right):

https://codereview.chromium.org/11465025/diff/1/chrome/installer/setup/instal...
chrome/installer/setup/install.cc:377: // TODO(tommi): Change this function to
use WorkItemList.
On 2012/12/10 20:58:28, robertshield wrote:
> since we don't generally roll back on failure, you can remove this comment.

Done.

https://codereview.chromium.org/11465025/diff/1/chrome/installer/util/shell_u...
File chrome/installer/util/shell_util.cc (right):

https://codereview.chromium.org/11465025/diff/1/chrome/installer/util/shell_u...
chrome/installer/util/shell_util.cc:1274: }
On 2012/12/10 20:58:28, robertshield wrote:
> clean up nit: we build and populate both user_shortcut_path and
> system_shortcut_path but only ever use one of them. Would it not be better to
> first decide which one we need, then build it? Doing so could considerably
> shorten this method.

While it's true that we do not need the user-level when chosing the system-level
one; the opposite is not true (i.e. when using the user-level one we need to
make sure the system-level one doesn't exist) -- the check is only if operation
== SHELL_SHORTCUT_CREATE_IF_NO_SYSTEM_LEVEL so I agree that this could be
optimized, but I don't this belongs in this CL which needs to be merged in M24.

[robertshield, 2012-12-10 22:16:05]

lgtm

[commit-bot: I haz the power, 2012-12-10 22:16:33]

CQ is trying da patch. Follow status at
https://chromium-status.appspot.com/cq/gab@chromium.org/11465025/8001

[commit-bot: I haz the power, 2012-12-11 01:53:52]

Change committed as 172223