[BUG] Memcheck error found in JSON_TEST JsonReaderTest.ErrorStrings

[davidwendt]

Describe the bug
The nightly builds memcheck run found an error in the JSON_TEST JsonReaderTest.ErrorStrings which was introduced by PR #13803

Steps/Code to reproduce bug

# compute-sanitizer --tool memcheck gtests/JSON_TEST --gtest_filter=JsonReaderTest.ErrorStrings --rmm_mode=cuda
========= COMPUTE-SANITIZER
Note: Google Test filter = JsonReaderTest.ErrorStrings
[==========] Running 1 test from 1 test suite.
[----------] Global test environment set-up.
[----------] 1 test from JsonReaderTest
[ RUN      ] JsonReaderTest.ErrorStrings
========= Invalid __global__ write of size 1 bytes
=========     at 0x1670 in void thrust::cuda_cub::core::_kernel_agent<thrust::cuda_cub::__parallel_for::ParallelForAgent<thrust::cuda_cub::for_each_f<thrust::counting_iterator<int, thrust::use_default, thrust::use_default, thrust::use_default>, thrust::detail::wrapped_function<cudf::io::json::detail::string_parse<thrust::transform_iterator<cudf::io::json::detail::to_string_view_pair<char>, thrust::zip_iterator<thrust::tuple<const int *, const int *, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type>>, thrust::use_default, thrust::use_default>>, void>>, int>, thrust::cuda_cub::for_each_f<thrust::counting_iterator<int, thrust::use_default, thrust::use_default, thrust::use_default>, thrust::detail::wrapped_function<cudf::io::json::detail::string_parse<thrust::transform_iterator<cudf::io::json::detail::to_string_view_pair<char>, thrust::zip_iterator<thrust::tuple<const int *, const int *, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type>>, thrust::use_default, thrust::use_default>>, void>>, int>(T2, T3)
=========     by thread (2,0,0) in block (0,0,0)
=========     Address 0x7faf25a02e05 is out of bounds
=========     and is 1 bytes after the nearest allocation at 0x7faf25a02e00 of size 5 bytes

[wence-]

Backtrace with symbols:

========= Invalid __global__ write of size 1 bytes
=========     at 0x5280 in /home/wence/Documents/src/rapids/cudf/cpp/src/io/utilities/data_casting.cu:205:process_string<first_type>
=========     by thread (2,0,0) in block (0,0,0)
=========     Address 0x7f10dd202e05 is out of bounds
=========     and is 1 bytes after the nearest allocation at 0x7f10dd202e00 of size 5 bytes
=========     Device Frame:/home/wence/Documents/src/rapids/cudf/cpp/src/io/utilities/data_casting.cu:770:cudf::io::json::detail::string_parse<thrust::transform_iterator<cudf::io::json::detail::to_string_view_pair<char>, thrust::zip_iterator<thrust::tuple<const int *, const int *, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type>>, thrust::use_default, thrust::use_default>>::operator ()(int) [0xaa40]
=========     Device Frame:/home/wence/Documents/src/rapids/cudf/cpp/build/cuda-11.8.0/branch-23.10/release/_deps/thrust-src/thrust/detail/function.h:125:operator()<size_type> [0x2610]
=========     Device Frame:/home/wence/Documents/src/rapids/cudf/cpp/build/cuda-11.8.0/branch-23.10/release/_deps/thrust-src/thrust/detail/function.h:58:operator()<int> [0x26a0]
=========     Device Frame:/home/wence/Documents/src/rapids/cudf/cpp/build/cuda-11.8.0/branch-23.10/release/_deps/thrust-src/thrust/system/cuda/detail/for_each.h:97:consume_tile<false> [0x27e0]
=========     Device Frame:/home/wence/Documents/src/rapids/cudf/cpp/build/cuda-11.8.0/branch-23.10/release/_deps/thrust-src/thrust/system/cuda/detail/parallel_for.h:118:entry [0x2860]
=========     Device Frame:/home/wence/Documents/src/rapids/cudf/cpp/build/cuda-11.8.0/branch-23.10/release/_deps/thrust-src/thrust/system/cuda/detail/core/agent_launcher.h:76:void thrust::cuda_cub::core::_kernel_agent<thrust::cuda_cub::__parallel_for::ParallelForAgent<thrust::cuda_cub::for_each_f<thrust::counting_iterator<int, thrust::use_default, thrust::use_default, thrust::use_default>, thrust::detail::wrapped_function<cudf::io::json::detail::string_parse<thrust::transform_iterator<cudf::io::json::detail::to_string_view_pair<char>, thrust::zip_iterator<thrust::tuple<const int *, const int *, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type>>, thrust::use_default, thrust::use_default>>, void>>, int>, thrust::cuda_cub::for_each_f<thrust::counting_iterator<int, thrust::use_default, thrust::use_default, thrust::use_default>, thrust::detail::wrapped_function<cudf::io::json::detail::string_parse<thrust::transform_iterator<cudf::io::json::detail::to_string_view_pair<char>, thrust::zip_iterator<thrust::tuple<const int *, const int *, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type, thrust::null_type>>, thrust::use_default, thrust::use_default>>, void>>, int>(T2, T3) [0x2880]
=========     Saved host backtrace up to driver entry point at kernel launch time