Add ostream helpers for FX String classes.

Add ostream helpers for FX String classes.

This allows integration with the gtest EXPECT_* macros.

R=brucedawson@chromium.org

Committed: https://pdfium.googlesource.com/pdfium/+/948095edfc04f35a69f15cd7d9926844ee0ac624

[Tom Sepez, 2015-01-06 23:50:02]

tsepez@chromium.org changed reviewers:
+ brucedawson@chromium.org, thestig@chromium.org

[Tom Sepez, 2015-01-06 23:50:02]

Lei, Bruce, please review.

[brucedawson, 2015-01-07 00:35:24]

https://codereview.chromium.org/837843002/diff/1/core/include/fxcrt/fx_string.h
File core/include/fxcrt/fx_string.h (right):

https://codereview.chromium.org/837843002/diff/1/core/include/fxcrt/fx_string...
core/include/fxcrt/fx_string.h:21: typedef unsigned char value_type;
Should this be FX_CHAR to match the declaration of m_String[1]? Ditto for
CFX_ByteString.

https://codereview.chromium.org/837843002/diff/1/core/include/fxcrt/fx_string...
core/include/fxcrt/fx_string.h:478: typedef wchar_t value_type;
Should this be FX_WCHAR to match the declaration of m_String[1]? Ditto for
CFX_WideString.

https://codereview.chromium.org/837843002/diff/1/testing/fx_string_testhelper...
File testing/fx_string_testhelpers.cpp (right):

https://codereview.chromium.org/837843002/diff/1/testing/fx_string_testhelper...
testing/fx_string_testhelpers.cpp:14: out << std::hex << std::setfill('0') <<
'"';
I assume that the goal is just to produce a reasonable representation of the
string, not necessarily one that can be parsed back into the original data?

I'm not familiar with Google's conventions for logging of data so I just wanted
to verify.

https://codereview.chromium.org/837843002/diff/1/testing/fx_string_testhelper...
testing/fx_string_testhelpers.cpp:22: if (c < 0x100) {
Are there any plausible platforms where this check is needed, given the sizeof()
check above?

https://codereview.chromium.org/837843002/diff/1/testing/fx_string_testhelper...
testing/fx_string_testhelpers.cpp:41: return output_string<CFX_ByteStringC>(out,
str);
The explicit template parameter in this call and in the other three is not
needed -- the compiler can infer the templatization parameter from the second
argument (tested with VC++). It's harmless (a mismatch would produce an error)
but doesn't feel like it adds value.

[Tom Sepez, 2015-01-07 00:42:24]

https://codereview.chromium.org/837843002/diff/1/core/include/fxcrt/fx_string.h
File core/include/fxcrt/fx_string.h (right):

https://codereview.chromium.org/837843002/diff/1/core/include/fxcrt/fx_string...
core/include/fxcrt/fx_string.h:21: typedef unsigned char value_type;
On 2015/01/07 00:35:24, brucedawson wrote:
> Should this be FX_CHAR to match the declaration of m_String[1]? Ditto for
> CFX_ByteString.

Done.

https://codereview.chromium.org/837843002/diff/1/core/include/fxcrt/fx_string...
core/include/fxcrt/fx_string.h:478: typedef wchar_t value_type;
On 2015/01/07 00:35:24, brucedawson wrote:
> Should this be FX_WCHAR to match the declaration of m_String[1]? Ditto for
> CFX_WideString.

Done.

https://codereview.chromium.org/837843002/diff/1/testing/fx_string_testhelper...
File testing/fx_string_testhelpers.cpp (right):

https://codereview.chromium.org/837843002/diff/1/testing/fx_string_testhelper...
testing/fx_string_testhelpers.cpp:14: out << std::hex << std::setfill('0') <<
'"';
On 2015/01/07 00:35:24, brucedawson wrote:
> I assume that the goal is just to produce a reasonable representation of the
> string, not necessarily one that can be parsed back into the original data?
> 
> I'm not familiar with Google's conventions for logging of data so I just
wanted
> to verify.

Exactly.  There's no inverse operation provided.

https://codereview.chromium.org/837843002/diff/1/testing/fx_string_testhelper...
testing/fx_string_testhelpers.cpp:22: if (c < 0x100) {
On 2015/01/07 00:35:24, brucedawson wrote:
> Are there any plausible platforms where this check is needed, given the
sizeof()
> check above?

Done.

https://codereview.chromium.org/837843002/diff/1/testing/fx_string_testhelper...
testing/fx_string_testhelpers.cpp:41: return output_string<CFX_ByteStringC>(out,
str);
On 2015/01/07 00:35:24, brucedawson wrote:
> The explicit template parameter in this call and in the other three is not
> needed -- the compiler can infer the templatization parameter from the second
> argument (tested with VC++). It's harmless (a mismatch would produce an error)
> but doesn't feel like it adds value.

Done.

[brucedawson, 2015-01-07 01:35:59]

lgtm

[Tom Sepez, 2015-01-07 20:25:11]

Committed patchset #3 (id:40001) manually as
948095edfc04f35a69f15cd7d9926844ee0ac624 (presubmit successful).