Make Directory's code style a little more consistent.

Make Directory's code style a little more consistent.

This is a cleanup-only/documentation change, no behavior changes.

Add comments explaining how Directory locking works.

Move protected methods to private (there was no reason for them to be
protected).

Move private data members below private method declarations.

For methods that take a ScopedKernelLock, always pass by const reference
before non-const pointer parameters.

Give ClearDirtyMetahandles a ScopedKernelLock parameter so it's clear
that the caller must be holding a lock.

BUG=

Committed: https://crrev.com/edb7809b91c8b29404ae9ec9322b4e16edeabaa1
Cr-Commit-Position: refs/heads/master@{#296045}

[maniscalco, 2014-09-18 21:00:32]

Nicolas, would you please review this change?

[maniscalco, 2014-09-22 18:31:30]

maniscalco@chromium.org changed reviewers:
+ zea@chromium.org

[maniscalco, 2014-09-22 18:31:58]

Putting zea on the R= line for real this time.

[Nicolas Zea, 2014-09-22 19:46:26]

LGTM with a couple nits

https://codereview.chromium.org/585663002/diff/20001/sync/syncable/directory.h
File sync/syncable/directory.h (right):

https://codereview.chromium.org/585663002/diff/20001/sync/syncable/directory....
sync/syncable/directory.h:539: bool InsertEntry(const ScopedKernelLock& lock,
nit: be consistent about ordering of locking version vs non-locking one?

https://codereview.chromium.org/585663002/diff/20001/sync/syncable/directory....
sync/syncable/directory.h:546: bool ReindexParentId(BaseWriteTransaction* trans,
EntryKernel* const entry,
nit: newline above, since this isn't a pair of locking/not locking methods

https://codereview.chromium.org/585663002/diff/20001/sync/syncable/directory....
sync/syncable/directory.h:560: // Add each of |metahandle|'s attachment ids to
the index_by_attachment_id.
nit: newline above?

[maniscalco, 2014-09-22 19:54:59]

All applied.  Thanks for the review.

https://codereview.chromium.org/585663002/diff/20001/sync/syncable/directory.h
File sync/syncable/directory.h (right):

https://codereview.chromium.org/585663002/diff/20001/sync/syncable/directory....
sync/syncable/directory.h:539: bool InsertEntry(const ScopedKernelLock& lock,
On 2014/09/22 19:46:26, Nicolas Zea wrote:
> nit: be consistent about ordering of locking version vs non-locking one?

Done.

https://codereview.chromium.org/585663002/diff/20001/sync/syncable/directory....
sync/syncable/directory.h:546: bool ReindexParentId(BaseWriteTransaction* trans,
EntryKernel* const entry,
On 2014/09/22 19:46:25, Nicolas Zea wrote:
> nit: newline above, since this isn't a pair of locking/not locking methods

Done.

https://codereview.chromium.org/585663002/diff/20001/sync/syncable/directory....
sync/syncable/directory.h:560: // Add each of |metahandle|'s attachment ids to
the index_by_attachment_id.
On 2014/09/22 19:46:26, Nicolas Zea wrote:
> nit: newline above?

Done.

[maniscalco, 2014-09-22 19:55:38]

The CQ bit was checked by maniscalco@chromium.org

[commit-bot: I haz the power, 2014-09-22 19:57:27]

CQ is trying da patch. Follow status at
 https://chromium-cq-status.appspot.com/patch-status/585663002/40001

[commit-bot: I haz the power, 2014-09-22 20:45:16]

Committed patchset #3 (id:40001) as ff1f0f2a8dbe8d42207b0ec17ce1e06c175cb26a

[commit-bot: I haz the power, 2014-09-22 20:45:56]

Patchset 3 (id:??) landed as
https://crrev.com/edb7809b91c8b29404ae9ec9322b4e16edeabaa1
Cr-Commit-Position: refs/heads/master@{#296045}