Port a few test_shell_tests. Now it runs 45 of them....

Port a few test_shell_tests. Now it runs 45 of them.

I removed some files from test_files because scons complains (warns)
when a file is both in test_files and input_files. So the files removed from
test_files also exist in input_files and at some pointed will be ported anyway.

I also made some linking stubs.

BUG=3674

[Mark Mentovai, 2008-10-24 17:12:55]

Nice job

http://codereview.chromium.org/7974/diff/1/5
File webkit/port/platform/GKURL_unittest.cpp (right):

http://codereview.chromium.org/7974/diff/1/5#newcode38
Line 38: #pragma warning(push, 0)
Use the compiler_specific.h macros here.

http://codereview.chromium.org/7974/diff/1/5#newcode44
Line 44: // This evil preprocessor hack allows us to run both the original KURL
and our
Ew, evil indeed.  OK...

http://codereview.chromium.org/7974/diff/1/5#newcode289
Line 289: // TODO(port): Port this test. The problem is probably with different
UChar
This may also be masking a behavior problem that we don't want to persist.  I'd
rather have this figured out "for sure" instead of just "probably" before we #if
this to not run on other platforms.

http://codereview.chromium.org/7974/diff/1/2
File webkit/tools/test_shell/image_decoder_unittest.h (right):

http://codereview.chromium.org/7974/diff/1/2#newcode4
Line 4: 
Does this file need #include guards?

[please use my chromium address, 2008-10-25 14:26:43]

After some recent updates more of the files link, so after few stubs more (and
few no longer relevant removed) we can run 45 tests. GKURL_unittest.cpp is now
ported properly.

http://codereview.chromium.org/7974/diff/1/5
File webkit/port/platform/GKURL_unittest.cpp (right):

http://codereview.chromium.org/7974/diff/1/5#newcode38
Line 38: #pragma warning(push, 0)
On 2008/10/24 17:12:55, Mark Mentovai wrote:
> Use the compiler_specific.h macros here.

Done.

http://codereview.chromium.org/7974/diff/1/5#newcode289
Line 289: // TODO(port): Port this test. The problem is probably with different
UChar
On 2008/10/24 17:12:55, Mark Mentovai wrote:
> This may also be masking a behavior problem that we don't want to persist. 
I'd
> rather have this figured out "for sure" instead of just "probably" before we
#if
> this to not run on other platforms.

Okay. After further investigation I managed to make these tests pass. Hopefully
it doesn't break on Windows.

http://codereview.chromium.org/7974/diff/1/2
File webkit/tools/test_shell/image_decoder_unittest.h (right):

http://codereview.chromium.org/7974/diff/1/2#newcode4
Line 4: 
On 2008/10/24 17:12:55, Mark Mentovai wrote:
> Does this file need #include guards?

C++ style guide says every file should have them, so I added the guards.

[Evan Martin, 2008-10-27 19:54:10]

looking now

[Evan Martin, 2008-10-27 22:48:42]

r4031
(I reordered some of your #includes to alphabetical)