Cleanup Registry API: part 3.

Cleanup Registry API: part 3.

- Remove the default arguments from the constructor.
- Add a empty constructor to handle the case of the default arguments were used.

BUG=44644
TEST=trybots

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=56123

[tfarina, 2010-08-12 14:37:54]

Marc, how can I fix the failures in rlz (it's a third_party library)? Pulling it
on my tree won't work right? Because we ignore it on .gitignore.

[M-A Ruel, 2010-08-12 15:02:28]

Pulling in Roger would could do the fix for you first.

lgtm once rlz is fixed.

[Roger Tawa OOO till Jul 10th, 2010-08-12 15:17:30]

http://codereview.chromium.org/3172009/diff/4001/5002
File base/registry.h (right):

http://codereview.chromium.org/3172009/diff/4001/5002#newcode24
base/registry.h:24: RegKey(HKEY rootkey, const wchar_t* subkey, REGSAM access);
Why not add a third ctor like this:

  // Use this constructor for read-only access to key.
  RegKey(HKEY rootkey, const wchar_t* subkey);


Then you don't need to change win_utils.cc and plugin_list_win.cc, and rlz won't
break.

[M-A Ruel, 2010-08-12 15:25:24]

http://codereview.chromium.org/3172009/diff/4001/5002
File base/registry.h (right):

http://codereview.chromium.org/3172009/diff/4001/5002#newcode24
base/registry.h:24: RegKey(HKEY rootkey, const wchar_t* subkey, REGSAM access);
On 2010/08/12 15:17:32, Roger Tawa wrote:
> Why not add a third ctor like this:
> 
>   // Use this constructor for read-only access to key.
>   RegKey(HKEY rootkey, const wchar_t* subkey);
> 
> 
> Then you don't need to change win_utils.cc and plugin_list_win.cc, and rlz
won't
> break.

In theory I agree but I like the expressiveness that the key is open read-only
and will be more coherent with Create() if its default argument is removed. I
don't really mind either way.

[(NOT FOR CODE REVIEWS), 2010-08-12 15:30:08]

OK.  Let me make a CL to fix rlz.

MA: can you help me setup code reviews for the rlz project?  Or maybe
just point me to the instructions?

Thanks,
Roger

-



On Thu, Aug 12, 2010 at 11:25,  <maruel@chromium.org> wrote:
>
> http://codereview.chromium.org/3172009/diff/4001/5002
> File base/registry.h (right):
>
> http://codereview.chromium.org/3172009/diff/4001/5002#newcode24
> base/registry.h:24: RegKey(HKEY rootkey, const wchar_t* subkey, REGSAM
> access);
> On 2010/08/12 15:17:32, Roger Tawa wrote:
>>
>> Why not add a third ctor like this:
>
>>   // Use this constructor for read-only access to key.
>>   RegKey(HKEY rootkey, const wchar_t* subkey);
>
>
>> Then you don't need to change win_utils.cc and plugin_list_win.cc, and
>
> rlz won't
>>
>> break.
>
> In theory I agree but I like the expressiveness that the key is open
> read-only and will be more coherent with Create() if its default
> argument is removed. I don't really mind either way.
>
> http://codereview.chromium.org/3172009/show
>

[Roger Tawa OOO till Jul 10th, 2010-08-12 16:00:34]

I have updated the RLZ code to specify the access argument always.  To get the,
change in chrome, modify http://src.chromium.org/viewvc/chrome/trunk/src/DEPS to
get revision @8 of the rlz code.  Let me know how it goes.

[tfarina, 2010-08-13 00:37:39]

On 2010/08/12 16:00:34, Roger Tawa wrote:
> I have updated the RLZ code to specify the access argument always.  
Thanks Roger for fixing it quickly.

>To get the,
> change in chrome, modify http://src.chromium.org/viewvc/chrome/trunk/src/DEPS
to
> get revision @8 of the rlz code.  Let me know how it goes.

Sent I patch for review to update the deps in order to get these fixes.

[tfarina, 2010-08-13 00:38:25]

On 2010/08/13 00:37:39, tfarina wrote:
> On 2010/08/12 16:00:34, Roger Tawa wrote:
> > I have updated the RLZ code to specify the access argument always.  
> Thanks Roger for fixing it quickly.
> 
> >To get the,
> > change in chrome, modify
http://src.chromium.org/viewvc/chrome/trunk/src/DEPS
> to
> > get revision @8 of the rlz code.  Let me know how it goes.
> 
> Sent I patch for review to update the deps in order to get these fixes.

s/I/a

[tfarina, 2010-08-13 13:08:41]

Marc, I'm going to commit this.