Add ZipReader::ExtractCurrentEntry with a delegate interface.

Add ZipReader::ExtractCurrentEntry with a delegate interface.

This change gives consumers of ZipReader a way to have the current entry
streamed to them via a Delegate interface. It also:

- Reduces duplication in the ExtractCurrentEntry* functions.
- Uses the heap rather than the stack for intermediate buffers.
- Changes ExtractCurrentEntryToFd to ExtractCurrentEntryToFile, making
  it cross-platform in the process.

BUG=462584
Committed: https://crrev.com/2919be01ff758875fc7161f6b41f4461518c1213
Cr-Commit-Position: refs/heads/master@{#320948}

Committed: https://crrev.com/98e6db00a6cc9cf6df682240d61f1b535b564cb8
Cr-Commit-Position: refs/heads/master@{#321205}

[grt (UTC plus 2), 2015-03-16 20:20:50]

grt@chromium.org changed reviewers:
+ hshi@chromium.org

[grt (UTC plus 2), 2015-03-16 20:21:15]

PTAL, thanks.

[hshi1, 2015-03-16 20:37:18]

On 2015/03/16 20:21:15, grt wrote:
> PTAL, thanks.

Please fix warnings:

../../third_party/zlib/google/zip_reader.cc: In destructor 'virtual
zip::FileWriterDelegate::~FileWriterDelegate()':
../../third_party/zlib/google/zip_reader.cc:521:14: error: unused variable
'success' [-Werror=unused-variable]
   const bool success = file_->SetLength(file_length_);
              ^

[grt (UTC plus 2), 2015-03-17 03:01:00]

Patchset #2 (id:20001) has been deleted

[grt (UTC plus 2), 2015-03-17 03:01:52]

Done, I think. I've also added a new test. Thanks for taking a look.

[grt (UTC plus 2), 2015-03-17 13:05:55]

On 2015/03/17 03:01:52, grt wrote:
> Done, I think. I've also added a new test. Thanks for taking a look.

FYI: the red tryruns all appear unrelated to this change (failures with and
without patch).

[hshi1, 2015-03-17 17:46:30]

lgtm

[grt (UTC plus 2), 2015-03-17 18:06:00]

The CQ bit was checked by grt@chromium.org

[commit-bot: I haz the power, 2015-03-17 18:06:12]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1014653002/40001

[commit-bot: I haz the power, 2015-03-17 18:53:44]

Committed patchset #2 (id:40001)

[commit-bot: I haz the power, 2015-03-17 18:54:50]

Patchset 2 (id:??) landed as
https://crrev.com/2919be01ff758875fc7161f6b41f4461518c1213
Cr-Commit-Position: refs/heads/master@{#320948}

[benwells, 2015-03-18 02:31:12]

A revert of this CL (patchset #2 id:40001) has been created in
https://codereview.chromium.org/1016813004/ by benwells@chromium.org.

The reason for reverting is: There are uninitialized reads on the DrMemory bots
in zlib ZipTests. They have very little information, and there is no history
available due to a bot problem.

Experimentally reverting this CL to see if it helps..

[grt (UTC plus 2), 2015-03-18 19:13:39]

I'm re-landing since I can't repro the failure locally. I have reason to believe
that a clobber has been done on the DrMemory builder so that future test runs
will contain readable stacks.

[grt (UTC plus 2), 2015-03-18 19:13:42]

The CQ bit was checked by grt@chromium.org

[commit-bot: I haz the power, 2015-03-18 19:14:15]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/1014653002/40001

[commit-bot: I haz the power, 2015-03-18 21:22:49]

Committed patchset #2 (id:40001)

[commit-bot: I haz the power, 2015-03-18 21:23:45]

Patchset 2 (id:??) landed as
https://crrev.com/98e6db00a6cc9cf6df682240d61f1b535b564cb8
Cr-Commit-Position: refs/heads/master@{#321205}