Imported Optional from Chromium with some modifications.

Imported Optional from Chromium with some modifications.

Specifically I reimplemented the OptionalStorage class it uses with a
union, rather than their AlignedMemory class. This means less code has
to be moved over and it's also a much more standard solution for this
type of thing.

[ossu, 2016-06-17 14:30:18]

Patchset #3 (id:40001) has been deleted

[ossu, 2016-06-17 14:51:48]

ossu@webrtc.org changed reviewers:
+ kwiberg@webrtc.org

[ossu, 2016-06-17 14:51:49]

Import of Chromium's Optional.
There seems to be a troll running about, because for some reason the default
constructor gets implicitly deleted, but only on some builds. Maybe it's a fault
with the compiler used in those builds or maybe I missed something. Please take
a look to see if you can make sense of it!

https://codereview.webrtc.org/2071003003/diff/1/webrtc/base/optional.h
File webrtc/base/optional.h (right):

https://codereview.webrtc.org/2071003003/diff/1/webrtc/base/optional.h#newcode1
webrtc/base/optional.h:1: // Copyright 2016 The Chromium Authors. All rights
reserved.
I kept the original copyright message, since I don't know what to do with it.
Should I keep both or just replace it with ours and a mention this file is from
Chromium?

https://codereview.webrtc.org/2071003003/diff/1/webrtc/base/optional_unittest.cc
File webrtc/base/optional_unittest.cc (right):

https://codereview.webrtc.org/2071003003/diff/1/webrtc/base/optional_unittest...
webrtc/base/optional_unittest.cc:1: // Copyright 2016 The Chromium Authors. All
rights reserved.
Same with this copyright message.

https://codereview.webrtc.org/2071003003/diff/1/webrtc/base/template_util.h
File webrtc/base/template_util.h (right):

https://codereview.webrtc.org/2071003003/diff/1/webrtc/base/template_util.h#n...
webrtc/base/template_util.h:126: // This hacks around libstdc++ 4.6 missing
stuff in type_traits, while we need
These are necessary to implement the partially specialized versions of
OptionalStorage. If we don't care enough about propagating trivial
destructability with optionals (I'm guessing it's useful if I use Optional a
lot) I can remove this as well. For now, I think it should remain in.

[ossu, 2016-06-17 14:54:01]

Some context to my last comment:
The error I'm talking about can be seen in
https://build.chromium.org/p/tryserver.webrtc/builders/ios_api_framework/buil...

At that point, the compiler doesn't see a default constructor, but many other
builders got past it without a problem. Intriguing!

[kwiberg-webrtc, 2016-06-18 05:24:58]

https://codereview.chromium.org/2080003002

Because it makes little sense for us to fork Chromium's Optional. If union is a
good idea for us, it probably is for Chromium too.

[tommi, 2017-02-21 16:50:33]

On 2016/06/18 05:24:58, kwiberg-webrtc wrote:
> https://codereview.chromium.org/2080003002
> 
> Because it makes little sense for us to fork Chromium's Optional. If union is
a
> good idea for us, it probably is for Chromium too.

should we close this issue?

[ossu, 2017-02-21 16:55:48]

On 2017/02/21 16:50:33, tommi (webrtc) wrote:
> On 2016/06/18 05:24:58, kwiberg-webrtc wrote:
> > https://codereview.chromium.org/2080003002
> > 
> > Because it makes little sense for us to fork Chromium's Optional. If union
is
> a
> > good idea for us, it probably is for Chromium too.
> 
> should we close this issue?

I'm not sure. It's there as a vague reminder I should do something about this at
some point. I believe having a more conforming Optional, with nullopt for
example, would be a good thing. It's just not been a priority for me lately.