Fix memory leak in deflate decoder

[gigony]

I did a silly mistake in deflate decoder which caused a memory leak when loading the Deflate-compressed TIFF image.

This patch will free allocated memory in decode_deflate() method.

For the next steps to prevent this kind of errors:

• Make use of a safe pointer (smart pointer) or scoped deleter for 3rdparty library/internal memory allocation.
• Run unit/integration tests with memory sanitizer/tracker (e.g., valgrind/asan)

[quasiben]

rerun tests

[gigony]

Build aborted due to timeout:

https://gpuci.gpuopenanalytics.com/job/rapidsai/job/gpuci-v0.20/job/cucim/job/prb/job/cucim-gpu-test/CUDA=11.0,GPU_LABEL=gpu,OS=ubuntu16.04,PYTHON=3.7/55/console

Examining conflict for rapids-build-env libarchive libxml2 icu:  77%|███████▋  | 71/92 [20:48<05:27, 15.62s/it]
Examining conflict for rapids-build-env cucim:  77%|███████▋  | 71/92 [21:08<05:27, 15.62s/it]                 
Examining conflict for rapids-build-env cucim:  78%|███████▊  | 72/92 [21:08<05:39, 16.95s/it]
Build timed out (after 40 minutes). Marking the build as failed.

[jakirkham]

Yeah this came up on Slack as well. Wondering if this is just due to the fact that we haven't fully moved to the new RAPIDS version scheme. If so, then it might be worth focusing on getting PR ( #41 ) in first and working with AJ to get cuCIM updated to the new version scheme

[gigony]

this change merged by other PR.

#close