Include chrome/SConscript.unit_tests in Linux build....

Include chrome/SConscript.unit_tests in Linux build.

Some minor fixes done to prevent compile failures when
http://codereview.chromium.org/2418 gets commited.

It first requires http://codereview.chromium.org/2890 to be applied
(building chrome/browser/).

BUG=2335

[please use my chromium address, 2008-09-16 14:15:57]

Will probably need more reviewers, please check.

[Mark Mentovai, 2008-09-16 14:32:37]

http://codereview.chromium.org/2891/diff/1/6
File chrome/browser/net/dns_host_info_unittest.cc (right):

http://codereview.chromium.org/2891/diff/1/6#newcode11
Line 11: 
Why is the blank line here?

http://codereview.chromium.org/2891/diff/1/4
File chrome/browser/safe_browsing/chunk_range_unittest.cc (right):

http://codereview.chromium.org/2891/diff/1/4#newcode51
Line 51: EXPECT_EQ(static_cast<int>(ranges.size()), 2);
Prefer static_cast<size_t> on the other term.

These are actually flipped, they should be written as EXPECT_EQ(expected,
actual).

http://codereview.chromium.org/2891/diff/1/3
File chrome/browser/safe_browsing/protocol_parser_unittest.cc (right):

http://codereview.chromium.org/2891/diff/1/3#newcode10
Line 10: #if defined(OS_WIN)
#include "base/basictypes.h" above, it's not clear that we'll have the OS_
macros defined otherwise.

http://codereview.chromium.org/2891/diff/1/3#newcode31
Line 31: EXPECT_EQ(static_cast<int>(chunks.size()), 1);
As before.

http://codereview.chromium.org/2891/diff/1/3#newcode464
Line 464: // TODO(paulg): Bug: http://b/1084719, skip this test on Windows 2000
until
Take this comment out, we don't care about w2k and already have the necessary
comment in hmac_win.cc.

http://codereview.chromium.org/2891/diff/1/3#newcode636
Line 636: #endif // defined(OS_WIN)
Two spaces before //.

http://codereview.chromium.org/2891/diff/1/5
File chrome/browser/safe_browsing/safe_browsing_util_unittest.cc (right):

http://codereview.chromium.org/2891/diff/1/5#newcode40
Line 40: EXPECT_EQ(static_cast<int>(hosts.size()), 2);
As before.

[please use my chromium address, 2008-09-16 16:04:53]

On 2008/09/16 14:32:37, Mark Mentovai wrote:
> http://codereview.chromium.org/2891/diff/1/6
> File chrome/browser/net/dns_host_info_unittest.cc (right):
> 
> http://codereview.chromium.org/2891/diff/1/6#newcode11
> Line 11: 
> Why is the blank line here?

I didn't know how to exactly break these include sections. I'm still not sure
whether this is correct.

> Prefer static_cast<size_t> on the other term.

Done.

> These are actually flipped, they should be written as EXPECT_EQ(expected,
> actual).

Sure, but should I change this too? It would require changes in other areas of
these files for consistency, and that could be beyond the scope of this patch.

> http://codereview.chromium.org/2891/diff/1/3#newcode464
> Line 464: // TODO(paulg): Bug: http://b/1084719, skip this test on Windows
2000
> until
> Take this comment out, we don't care about w2k and already have the necessary
> comment in hmac_win.cc.

Please check whether I've done what you intended.

All other fixes applied too.

[Mark Mentovai, 2008-09-16 17:41:58]

LGTM with a few more changes and another pair of eyes on the SConscript.

http://codereview.chromium.org/2891/diff/13/214
File chrome/browser/net/dns_host_info_unittest.cc (right):

http://codereview.chromium.org/2891/diff/13/214#newcode93
Line 93: }  // namespace anonymous
This is not the end of "namespace anonymous", the comment should just be "//
namespace".  (I know it was here already, but let's roll it in.

http://codereview.chromium.org/2891/diff/13/212
File chrome/browser/safe_browsing/chunk_range_unittest.cc (right):

http://codereview.chromium.org/2891/diff/13/212#newcode51
Line 51: EXPECT_EQ(ranges.size(), static_cast<size_t>(2));
You're right about reorganizing these, we can save the expected/actual ordering
fix here for another patch.

http://codereview.chromium.org/2891/diff/13/211
File chrome/browser/safe_browsing/protocol_parser_unittest.cc (right):

http://codereview.chromium.org/2891/diff/13/211#newcode12
Line 12: #include "base/win_util.h"
Oh, DOH!  The only reason win_util is brought into this file is to check for
Windows 2000.  We don't need to worry about that any more.  Let's take this
#include out (along with basictypes.h) and get rid of the Windows 2000 checks in
the tests.

http://codereview.chromium.org/2891/diff/13/211#newcode462
Line 462: // TODO: Enable this test when we have cross-platform HMAC working.
This should be TODO(port): too.

http://codereview.chromium.org/2891/diff/13/211#newcode565
Line 565: // TODO: Enable these tests when we have cross-platform HMAC working.
This too.

http://codereview.chromium.org/2891/diff/13/210
File chrome/test/unit/run_all_unittests.cc (right):

http://codereview.chromium.org/2891/diff/13/210#newcode9
Line 9: // TODO: This is not Windows-specific, but needs to be ported.
TODO(port) in this file as well.

[please use my chromium address, 2008-09-16 18:20:32]

On 2008/09/16 17:41:58, Mark Mentovai wrote:
> LGTM with a few more changes and another pair of eyes on the SConscript.
> 
> http://codereview.chromium.org/2891/diff/13/214
> File chrome/browser/net/dns_host_info_unittest.cc (right):
> 
> http://codereview.chromium.org/2891/diff/13/214#newcode93
> Line 93: }  // namespace anonymous
> This is not the end of "namespace anonymous", the comment should just be "//
> namespace".  (I know it was here already, but let's roll it in.

No problem.

> http://codereview.chromium.org/2891/diff/13/211#newcode462
> Line 462: // TODO: Enable this test when we have cross-platform HMAC working.
> This should be TODO(port): too.

Done. I also updated the comment (NSS HMAC landed, but VerifyMAC is not yet
ready).

Rest of fixes done too.

[sgk, 2008-09-17 00:07:16]

SCons lgtm

[tony, 2008-09-17 17:55:05]

Testing this patch on windows and will commit soon...

[tony, 2008-09-17 18:16:31]

Committed as @r2319.  Thanks!