Fix leaks in embedder test's FlateEncode() usage and in FlateEncode().

Fix leaks in embedder test's FlateEncode() usage and in FlateEncode().

For FlateEncode(), error handling code leaked memory.

R=tsepez@chromium.org

Committed: https://pdfium.googlesource.com/pdfium/+/1962d61b28df03284e3e5c6de6a19f397a066e68

[Lei Zhang, 2015-05-15 06:46:46]

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

[Lei Zhang, 2015-05-15 06:46:46]

https://codereview.chromium.org/1144603002/diff/1/core/src/fxcodec/codec/fx_c...
File core/src/fxcodec/codec/fx_codec_flate.cpp (right):

https://codereview.chromium.org/1144603002/diff/1/core/src/fxcodec/codec/fx_c...
core/src/fxcodec/codec/fx_codec_flate.cpp:766: if (useOldImpl) {
I took the original code and make 2 copies, then removed the dead code in each
of the two branches.

[Tom Sepez, 2015-05-15 15:23:39]

we might want to ping Nico about how far off is a std::unique_ptr to replace our
nonstd:: one, since it can take a destructor, no?  Or consider adding one to our
impl if time -> inf.

https://codereview.chromium.org/1144603002/diff/1/core/src/fxcodec/codec/fx_c...
File core/src/fxcodec/codec/fx_codec_flate.cpp (right):

https://codereview.chromium.org/1144603002/diff/1/core/src/fxcodec/codec/fx_c...
core/src/fxcodec/codec/fx_codec_flate.cpp:759: if (!guess_buf)
nit: I'd expect this one line earlier, with line 758 moved after 761.

https://codereview.chromium.org/1144603002/diff/1/core/src/fxcodec/codec/fx_c...
core/src/fxcodec/codec/fx_codec_flate.cpp:802: if (ret != Z_OK) {
check for error before FlateGetAvailOut? as in line 769?

https://codereview.chromium.org/1144603002/diff/1/core/src/fxcodec/codec/fx_c...
core/src/fxcodec/codec/fx_codec_flate.cpp:817: for (FX_INT32 i = 0; i <
result_tmp_bufs.GetSize(); i++)
nit: would like { } despite the one-line body.  I have nightmares about
mis-indented goto fails ...

https://codereview.chromium.org/1144603002/diff/1/core/src/fxcodec/codec/fx_c...
core/src/fxcodec/codec/fx_codec_flate.cpp:830: for (FX_INT32 i = 0; i <
result_tmp_bufs.GetSize(); i++)
ditto

[Lei Zhang, 2015-05-15 22:29:49]

https://codereview.chromium.org/1144603002/diff/1/core/src/fxcodec/codec/fx_c...
File core/src/fxcodec/codec/fx_codec_flate.cpp (right):

https://codereview.chromium.org/1144603002/diff/1/core/src/fxcodec/codec/fx_c...
core/src/fxcodec/codec/fx_codec_flate.cpp:759: if (!guess_buf)
On 2015/05/15 15:23:38, Tom Sepez wrote:
> nit: I'd expect this one line earlier, with line 758 moved after 761.

Compiler says, "crosses initialization of ‘unsigned char* cur_buf"

Alternatively, I coul make two |cur_buf| variables scoped to the if/else blocks
below.

https://codereview.chromium.org/1144603002/diff/1/core/src/fxcodec/codec/fx_c...
core/src/fxcodec/codec/fx_codec_flate.cpp:802: if (ret != Z_OK) {
On 2015/05/15 15:23:38, Tom Sepez wrote:
> check for error before FlateGetAvailOut? as in line 769?

The error handling code uses |avail_buf_size|, so I left the code as is.

https://codereview.chromium.org/1144603002/diff/1/core/src/fxcodec/codec/fx_c...
core/src/fxcodec/codec/fx_codec_flate.cpp:817: for (FX_INT32 i = 0; i <
result_tmp_bufs.GetSize(); i++)
On 2015/05/15 15:23:38, Tom Sepez wrote:
> nit: would like { } despite the one-line body.  I have nightmares about
> mis-indented goto fails ...

Done.

https://codereview.chromium.org/1144603002/diff/1/core/src/fxcodec/codec/fx_c...
core/src/fxcodec/codec/fx_codec_flate.cpp:830: for (FX_INT32 i = 0; i <
result_tmp_bufs.GetSize(); i++)
On 2015/05/15 15:23:38, Tom Sepez wrote:
> ditto

Done.

[Tom Sepez, 2015-05-15 22:34:32]

lgtm

[Lei Zhang, 2015-05-15 23:09:16]

Committed patchset #2 (id:20001) manually as
1962d61b28df03284e3e5c6de6a19f397a066e68 (presubmit successful).