Closed AndreyNudko closed 2 years ago
Merging #396 (14eff7a) into master (2234592) will decrease coverage by
0.06%
. The diff coverage is100.00%
.
@@ Coverage Diff @@
## master #396 +/- ##
============================================
- Coverage 62.43% 62.37% -0.07%
+ Complexity 1250 1246 -4
============================================
Files 135 135
Lines 6640 6628 -12
Branches 516 514 -2
============================================
- Hits 4146 4134 -12
Misses 2219 2219
Partials 275 275
Impacted Files | Coverage Δ | |
---|---|---|
.../com/github/housepower/buffer/ByteArrayWriter.java | 100.00% <ø> (ø) |
|
...thub/housepower/buffer/CompressedBuffedWriter.java | 97.36% <ø> (-0.14%) |
:arrow_down: |
...m/github/housepower/buffer/SocketBuffedWriter.java | 100.00% <ø> (ø) |
|
...com/github/housepower/data/ColumnWriterBuffer.java | 100.00% <100.00%> (ø) |
|
.../com/github/housepower/serde/BinarySerializer.java | 94.73% <100.00%> (ø) |
Continue to review full report at Codecov.
Legend - Click here to learn more
Δ = absolute <relative> (impact)
,ø = not affected
,? = missing data
Powered by Codecov. Last update 2234592...14eff7a. Read the comment docs.
Instead of copying serialized column data from ByteBuffer via intermediate byte[], just write backing array directly. This somewhat couples ColumnWriterBuffer with ByteArrayWriter internals, however as the name implies ByteArrayWriter is expected to use heap buffers (and using direct buffers for block data wouldn't make much sense now anyway - I/O layer is operating on byte[], plus most JVMs impose limits on direct memory allocation). ByteArrayWriter and CompressedBuffedWriter use fixed-size buffers internally and chunk bytes already. SocketBuffedWriter may benefit from splitting large writes into smaller ones to minimize malloc/free in native code - but it shouldn't be worse than current situation (at least with compression) and likely to be dubious optimisation.