Delete some tests from allocator_unittests.

Delete some tests from allocator_unittests.

The atomics and "die on OOM" bits are tested elsewhere.
The motivation is that this is the only file in all of Chromium that uses
`try` / `catch`.  Originally I wanted to remove just that bit, but when I tried
to compile the file I noticed that the target doesn't exist in most configs,
and the bots don't run the test, so I think the whole thing can go.

Note that no bots run this test. It's only compiled on Windows, and only
in static library builds, and only when building `all` (which means no devs
build it either). http://crbug.com/458309 tracks moving the tests that remain
in this file (which are useful) into base_unittests.

BUG=82385
Committed: https://crrev.com/a249805f376e40174998be288ed64c6463d69000
Cr-Commit-Position: refs/heads/master@{#316084}

[Nico, 2015-02-12 21:54:55]

thakis@chromium.org changed reviewers:
+ cpu@chromium.org

[Nico, 2015-02-12 21:54:55]

There is no try.

[cpu_(ooo_6.6-7.5), 2015-02-12 22:02:19]

cpu@chromium.org changed reviewers:
+ wfh@chromium.org

[cpu_(ooo_6.6-7.5), 2015-02-12 22:02:19]

I am for it but I believe wfh might have an opinion.

[Will Harris, 2015-02-12 22:02:33]

On 2015/02/12 21:54:55, Nico wrote:
> There is no try.

Can we not add these to the try bots and remove the try/catch.  I'm not sure all
the tests (e.g. the aligned malloc) are actually tested anywhere, and they did
catch stuff I was missing when I was e.g. working on
https://codereview.chromium.org/774683003 - my bad for not actually adding them
to any bots.

[Nico, 2015-02-12 22:12:54]

On 2015/02/12 22:02:33, Will Harris wrote:
> On 2015/02/12 21:54:55, Nico wrote:
> > There is no try.
> 
> Can we not add these to the try bots and remove the try/catch.  I'm not sure
all
> the tests (e.g. the aligned malloc) are actually tested anywhere, and they did
> catch stuff I was missing when I was e.g. working on
> https://codereview.chromium.org/774683003 - my bad for not actually adding
them
> to any bots.

Ok, if you want to add it to bots I can keep around some of the file. Which bits
do you think are useful? The OOM tests are elsewhere, the atomic tests are
elsewhere too (they're also dead code in this translation unit).

Does patch set 2 look fine to you?

(I kind of feel it's not worth having a dedicated binary for these tests. If you
think they're useful, maybe you should move them to base_unittests instead of
hooking up the new binary in all the bot configs. I'm not sure if that's easily
doable though.)

[Nico, 2015-02-12 22:13:14]

(rietveld still claims 'D' but that's a lie. Look at the side-by-side diff.)

[Will Harris, 2015-02-12 22:24:49]

sure if cpu is okay with leaving this for a while, I can move all these tests
into base_unittests and remove the target once I'm done.

[Will Harris, 2015-02-12 22:27:56]

On 2015/02/12 22:24:49, Will Harris wrote:
> sure if cpu is okay with leaving this for a while, I can move all these tests
> into base_unittests and remove the target once I'm done.

created a bug to track this work, so I don't forget - crbug.com/458309

[Nico, 2015-02-12 22:34:27]

On 2015/02/12 22:27:56, Will Harris wrote:
> On 2015/02/12 22:24:49, Will Harris wrote:
> > sure if cpu is okay with leaving this for a while, I can move all these
tests
> > into base_unittests and remove the target once I'm done.
> 
> created a bug to track this work, so I don't forget - crbug.com/458309

Ok, does patch set 2 lgty in the meantime?

[Will Harris, 2015-02-12 22:37:05]

ps3 lgtm

[Nico, 2015-02-12 22:42:19]

The CQ bit was checked by thakis@chromium.org

[Nico, 2015-02-12 22:42:22]

Thanks!

[commit-bot: I haz the power, 2015-02-12 22:43:13]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/922863002/40001

[commit-bot: I haz the power, 2015-02-12 23:34:25]

Committed patchset #3 (id:40001)

[commit-bot: I haz the power, 2015-02-12 23:34:59]

Patchset 3 (id:??) landed as
https://crrev.com/a249805f376e40174998be288ed64c6463d69000
Cr-Commit-Position: refs/heads/master@{#316084}