Adding additional comparison operators for GeneralizedTime.

Adding additional comparison operators for GeneralizedTime.

BUG=

Committed: https://crrev.com/fbb800feef57571197fc5c02d2ea5c19fa5d772e
Cr-Commit-Position: refs/heads/master@{#400255}

[davidben, 2016-06-16 19:51:44]

davidben@chromium.org changed reviewers:
+ davidben@chromium.org

[davidben, 2016-06-16 19:51:45]

lgtm

[svaldez, 2016-06-16 19:52:10]

The CQ bit was checked by svaldez@chromium.org

[commit-bot: I haz the power, 2016-06-16 19:52:43]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/2075783002/20001

[commit-bot: I haz the power, 2016-06-16 20:58:02]

Committed patchset #2 (id:20001)

[commit-bot: I haz the power, 2016-06-16 20:59:17]

Description was changed from

==========
Adding additional comparison operators for GeneralizedTime.

BUG=
==========

to

==========
Adding additional comparison operators for GeneralizedTime.

BUG=

Committed: https://crrev.com/fbb800feef57571197fc5c02d2ea5c19fa5d772e
Cr-Commit-Position: refs/heads/master@{#400255}
==========

[commit-bot: I haz the power, 2016-06-16 20:59:18]

Patchset 2 (id:??) landed as
https://crrev.com/fbb800feef57571197fc5c02d2ea5c19fa5d772e
Cr-Commit-Position: refs/heads/master@{#400255}

[Ryan Sleevi, 2016-06-16 21:37:13]

rsleevi@chromium.org changed reviewers:
+ rsleevi@chromium.org

[Ryan Sleevi, 2016-06-16 21:37:14]

FWIW, I don't think we should do this.

using std::rel_ops

should be what people use for this.

[davidben, 2016-06-16 22:04:05]

On 2016/06/16 21:37:14, Ryan Sleevi wrote:
> FWIW, I don't think we should do this.
> 
> using std::rel_ops
> 
> should be what people use for this.

Huh, never seen that before! Interesting.

Although we don't seem to use std::rel_ops anywhere else, while we do this
everywhere else, including base::Optional which is both in base and a recent
addition, so my interpretation is this is the most recent practice.

https://cs.chromium.org/search/?q=operator%3E%3D+-file:third_party&sq=package...

Perhaps propose to cxx@ or chromium-dev@ that we use std::rel_ops? Though the
internal repo seems also more fond of operator>= than std::rel_ops. One could
certainly argue that std::rel_ops, by not applying to just one type, causes more
weird action-at-a-distance issues.

[Ryan Sleevi, 2016-06-16 22:21:27]

On 2016/06/16 22:04:05, davidben wrote:
> Huh, never seen that before! Interesting.
> 
> Although we don't seem to use std::rel_ops anywhere else, while we do this
> everywhere else, including base::Optional which is both in base and a recent
> addition, so my interpretation is this is the most recent practice.

Up until very recently, the style guide banned overloading the operators at all.

> 
>
https://cs.chromium.org/search/?q=operator%3E%3D+-file:third_party&sq=package...
> 
> Perhaps propose to cxx@ or chromium-dev@ that we use std::rel_ops? Though the
> internal repo seems also more fond of operator>= than std::rel_ops. One could
> certainly argue that std::rel_ops, by not applying to just one type, causes
more
> weird action-at-a-distance issues.

It's C++. I'm not sure why you're suggesting a proposal - this isn't some
esoteric C++11 functionality - it's core to std::utility and, by proxy, most of
the STL.

It really depends on how much of the API contract you want to make your
comparators. The trade-offs between overload and rel-ops is that anyone who
wants to use a comparator other than ==/< has to "using std::rel_ops" at their
callsite, since ADL/Koenig won't find it otherwise. For something like this -
where there's only one need to compare GeneralizedTime, I lean towards the 'less
is more' approach.