FBTF: Remove unneeded headers from base/ (part 9)

FBTF: Remove unneeded headers from base/ (part 9)

BUG=none
TEST=none

Committed: http://src.chromium.org/viewvc/chrome?view=rev&revision=66873

[Lei Zhang, 2010-11-19 01:25:32]

Please take a look at the base/lock*.h part. All the other changes are not as
interesting.

[jar (doing other things), 2010-11-19 01:45:13]

Looks like a lot of nice cleanup, Thanks!

Please be SURE this compiles on all platforms, as sometimes there are subtle
platform differences.  I didn't see try bot results... and that would make me
feel a lot better.

There is also one suggestion below, that I'd be curious to hear Brett's
perspective on.

After I see (positive) try bot compile results, and with the one nit listed
below, I'd be glad to approve this.

http://codereview.chromium.org/5091005/diff/1/base/lock.h
File base/lock.h (right):

http://codereview.chromium.org/5091005/diff/1/base/lock.h#newcode11
base/lock.h:11: #include "base/platform_thread.h"
IMO, using conditional inclusion depending on debug is a bad choice.  I'd rather
see the includes and suffer the compile time than add such preprocessing ifs
like this.  

I'll be curious to hear Brett's perspective.

[Lei Zhang, 2010-11-19 19:06:16]

http://codereview.chromium.org/5091005/diff/1/base/lock.h
File base/lock.h (right):

http://codereview.chromium.org/5091005/diff/1/base/lock.h#newcode11
base/lock.h:11: #include "base/platform_thread.h"
On 2010/11/19 01:45:14, jar wrote:
> IMO, using conditional inclusion depending on debug is a bad choice.  I'd
rather
> see the includes and suffer the compile time than add such preprocessing ifs
> like this.  
> 
> I'll be curious to hear Brett's perspective.

I can definitely see this causing release vs. debug build problems, so I removed
the #ifdef. All the bots except Windows passed on patch set 1, and Windows
passed on patch set 5.

[jar (doing other things), 2010-11-19 21:38:32]

LGTM

Side suggestion: When you land, you'll also probably induce a complete re-build,
which would greatly slow the bots during the day (peak checkin times).  Also,
with big changes like this, it is often difficult to land due to conflicts.  You
may have better luck landing at night or over the weekend... when the tree is
not changing fast.  ...and you won't cause the tree to enter into a "catch up /
sheriff induced closure."