Add tests to exercise FPDF_SaveAsCopy().

Add tests to exercise FPDF_SaveAsCopy().

[Lei Zhang, 2015-12-24 10:25:25]

thestig@chromium.org changed reviewers:
+ weili@chromium.org

[jun_fang, 2015-12-28 10:30:02]

On 2015/12/24 10:25:25, Lei Zhang wrote:

I checked the original file, weblinks.pdf. It seemed that 
someone changed 5 0 object to 4 0 object so that there
was missing 5 0 object in the original file. 5 0 object
was a font object. Without it, Evince didn't display 
the related text. However, some pdf readers like acrobat
and foxit phantom can display the texts because they use
default font even without 5 0 object.

[Wei Li, 2016-01-04 19:01:14]

https://codereview.chromium.org/1546063002/diff/1/fpdfsdk/src/fpdfsave_embedd...
File fpdfsdk/src/fpdfsave_embeddertest.cpp (right):

https://codereview.chromium.org/1546063002/diff/1/fpdfsdk/src/fpdfsave_embedd...
fpdfsdk/src/fpdfsave_embeddertest.cpp:38: buffer_.resize(buffer_.size() + size);
You may use these when buffer_ is empty as well.

https://codereview.chromium.org/1546063002/diff/1/testing/embedder_test.cpp
File testing/embedder_test.cpp (right):

https://codereview.chromium.org/1546063002/diff/1/testing/embedder_test.cpp#n...
testing/embedder_test.cpp:102: nRet = FPDFAvail_IsPageAvail(*out_data->avail, i,
out_data->hints);
I know this is from the original code, but worth checking. Will this enter
infinite loop if page load failed? For example, PreparePageItem may return false
while setting status as error (fpdf_parser_parser.cpp:3337).

[Tom Sepez, 2016-01-04 20:24:13]

tsepez@chromium.org changed reviewers:
+ tsepez@chromium.org

[Tom Sepez, 2016-01-04 20:24:26]

Now blocking https://codereview.chromium.org/1556033002/

[Tom Sepez, 2016-01-04 20:40:22]

https://codereview.chromium.org/1546063002/diff/1/fpdfsdk/src/fpdfsave_embedd...
File fpdfsdk/src/fpdfsave_embeddertest.cpp (right):

https://codereview.chromium.org/1546063002/diff/1/fpdfsdk/src/fpdfsave_embedd...
fpdfsdk/src/fpdfsave_embeddertest.cpp:53: std::vector<char> buffer_;
meh, might be happier if this were a string, so that we can examine it.  I'd
expect this doesn't get called on big files, etc.

[Tom Sepez, 2016-01-05 17:55:00]

On 2016/01/04 20:40:22, Tom Sepez wrote:
>
https://codereview.chromium.org/1546063002/diff/1/fpdfsdk/src/fpdfsave_embedd...
> File fpdfsdk/src/fpdfsave_embeddertest.cpp (right):
> 
>
https://codereview.chromium.org/1546063002/diff/1/fpdfsdk/src/fpdfsave_embedd...
> fpdfsdk/src/fpdfsave_embeddertest.cpp:53: std::vector<char> buffer_;
> meh, might be happier if this were a string, so that we can examine it.  I'd
> expect this doesn't get called on big files, etc.

As I thibk more about this, I'm not fond of the idea of checking the result by
re-parsing it.  Since we are producing bytes, we should be checking against
bytes, whether they come from an internal string literal or perhaps by slurping
in an expected txt file.

[Lei Zhang, 2016-01-07 02:50:58]

In any case, since tsepez already landed some CLs for testing, I'll close this
CL for now.

https://codereview.chromium.org/1546063002/diff/1/testing/embedder_test.cpp
File testing/embedder_test.cpp (right):

https://codereview.chromium.org/1546063002/diff/1/testing/embedder_test.cpp#n...
testing/embedder_test.cpp:102: nRet = FPDFAvail_IsPageAvail(*out_data->avail, i,
out_data->hints);
On 2016/01/04 19:01:13, Wei Li wrote:
> I know this is from the original code, but worth checking. Will this enter
> infinite loop if page load failed? For example, PreparePageItem may return
false
> while setting status as error (fpdf_parser_parser.cpp:3337).

Maybe. If you can create a test case that does that, you can change
FPDFAvail_IsPageAvail() to detect the infinite loop and return PDF_DATA_ERROR.
samples/pdfium_test.cc has this pattern as well.