Make ChromotingHost::ui_strings() immutable.

Make ChromotingHost::ui_strings() immutable. Use string16 for localized strings.

There were two issues:
 - ui_strings() were mutable which creates possibility for race
   condition. Made it immutable, and added SetUiString() which can be
   called only before session is started.
 - strings were loaded after host is started which sometimes leads to
   non-localized string being shown. Particularly I often saw empty 
   text on Disconnect button.

BUG=None
TEST=Disconnect button text is always localized.

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=97376

[Jamie, 2011-08-18 16:28:41]

Your first point is fine, but for the second, please see my comment about not
localizing before we have a username. If the dialog is appearing with no
Disconnect button, then we might need an alternative fix.

http://codereview.chromium.org/7669045/diff/1/remoting/host/plugin/host_scrip...
File remoting/host/plugin/host_script_object.cc (right):

http://codereview.chromium.org/7669045/diff/1/remoting/host/plugin/host_scrip...
remoting/host/plugin/host_script_object.cc:484: LocalizeStrings();
Do we know the username at this point? Your objection seems to be that we're
localizing too late, so I'm guessing this is an earlier point in the protocol,
which is no good. The initialization is where it is because that's the earliest
we can do it, given that we need the user name.

http://codereview.chromium.org/7669045/diff/1/remoting/host/ui_strings.cc
File remoting/host/ui_strings.cc (right):

http://codereview.chromium.org/7669045/diff/1/remoting/host/ui_strings.cc#new...
remoting/host/ui_strings.cc:20: disconnect_button_shortcut("Ctrl+Alt+Esc"),
If this CL goes through, can you remove this string? I'm pretty sure it's not
used any more, but I forgot to remove it.

[Sergey Ulanov, 2011-08-18 18:04:32]

I changed DisconnectWindow to substitute username. I believe that UiStrings
should not contain any substitutions, otherwise, in the future, it will be hard
to keep track of what needs to be substituted and when.

Also changed UiString to store strings in UTF16 instead of UTF8.

http://codereview.chromium.org/7669045/diff/1/remoting/host/plugin/host_scrip...
File remoting/host/plugin/host_script_object.cc (right):

http://codereview.chromium.org/7669045/diff/1/remoting/host/plugin/host_scrip...
remoting/host/plugin/host_script_object.cc:484: LocalizeStrings();
On 2011/08/18 16:28:41, Jamie wrote:
> Do we know the username at this point? Your objection seems to be that we're
> localizing too late, so I'm guessing this is an earlier point in the protocol,
> which is no good. The initialization is where it is because that's the
earliest
> we can do it, given that we need the user name.

I updated DisconnectWindow to substitute user name when dialog is being shown.

http://codereview.chromium.org/7669045/diff/1/remoting/host/ui_strings.cc
File remoting/host/ui_strings.cc (right):

http://codereview.chromium.org/7669045/diff/1/remoting/host/ui_strings.cc#new...
remoting/host/ui_strings.cc:20: disconnect_button_shortcut("Ctrl+Alt+Esc"),
On 2011/08/18 16:28:41, Jamie wrote:
> If this CL goes through, can you remove this string? I'm pretty sure it's not
> used any more, but I forgot to remove it.

Done.

[Jamie, 2011-08-18 18:44:26]

LGTM, but please remove the comment I've flagged.

http://codereview.chromium.org/7669045/diff/4005/remoting/host/ui_strings.cc
File remoting/host/ui_strings.cc (right):

http://codereview.chromium.org/7669045/diff/4005/remoting/host/ui_strings.cc#...
remoting/host/ui_strings.cc:15: // missing for remoting_simple_host
This comment can be removed now.

[Jamie, 2011-08-18 18:46:57]

On 2011/08/18 18:04:32, sergeyu wrote:
> I changed DisconnectWindow to substitute username. I believe that UiStrings
> should not contain any substitutions, otherwise, in the future, it will be
hard
> to keep track of what needs to be substituted and when.

I agree with you in principle, but I think it's too restrictive. It's really
just a specific case of a more general problem caused by the fact that the
messages are defined in one file and referenced in another, with equivalent of a
linker to check that they match up. I'm going to submit a CL the addresses part
of this, but it won't solve substitution issues.