FileSink now closes the underlying writer.

[samkaufman]

This PR modifies FileSink to, on close, close its inner FileStreamer's owned writer. At the moment, the inner AsyncWrite's close method is not called when the FileSink is closed.

This is a minor backwards-incompatible change. While it won't affect cases where the FileSink wraps an actual file, which is generally closed when dropped, it may affect behavior when wrapping an AsyncWrite which expects close to be called. (In my case, that was a handle to an S3 object.)

[codecov]

Codecov Report

Merging #1213 (90d1193) into main (3b9d86b) will increase coverage by 0.02%.
The diff coverage is 100.00%.

@@            Coverage Diff             @@
##             main    #1213      +/-   ##
==========================================
+ Coverage   83.48%   83.50%   +0.02%     
==========================================
  Files         356      356              
  Lines       37081    37082       +1     
==========================================
+ Hits        30957    30967      +10     
+ Misses       6124     6115       -9     

Impacted Files	Coverage Δ
src/io/parquet/write/sink.rs	72.41% <100.00%> (+0.23%)	⬆️
src/bitmap/utils/slice_iterator.rs	97.56% <0.00%> (-1.22%)	⬇️
src/io/ipc/read/reader.rs	96.32% <0.00%> (ø)
src/array/utf8/mod.rs	83.64% <0.00%> (+0.92%)	⬆️
src/array/binary/mod.rs	90.12% <0.00%> (+1.23%)	⬆️
src/io/ipc/read/array/boolean.rs	98.14% <0.00%> (+7.40%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[samkaufman]

The "Integration IPC / Flight / Test" failure seems to be a result of this recent commit: apache/arrow@6d1bc62. Perhaps merge should hold off until that's resolved (by a separate PR?).

[jorgecarleitao]

Thanks a lot @samkaufman 🙇 Yes, ignore the integration test - I need to fix that as something upstream changed.