[BUG] Deflate Undefined Behavior with Data Size Below Manager Chunk Size

Describe the bug

nvcomp deflate manager runs into illegal memory access when compressing data smaller than manager chunk size.
Note that all other managers did not report error in the same test case.

Steps/Code to reproduce bug

Reproducer:
nvcomp-deflate-reproducer.cpp.txt

Compile with:

g++ -g -Wall -Wextra -Wall -std=c++17 -I/usr/local/cuda/include -I<path_to_nvcomp.h> nvcomp-deflate-reproducer.cpp -o reproducer -lcudart -lnvcomp

Run with:

compute-sanitizer ./reproducer

Expected behavior

The reproducer I attached compresses a buffer of size 33B with DeflateManager and recommended chunk size.
When ran through compute-sanitizer the tool will report illegal memory access.
You can run the same code with other managers (adjusting to the recommended chunk size) and there will be no error reported.

If this is intended, please add some documentation about it.

Environment details:

Bare-metal
Cuda Driver Version: 535.154.05
CUDA Version: 12.2
NVIDIA GeForce RTX 3050
nvcomp Version: 3.0.4 (no extensions)