Antoine Pitrou
Antoine Pitrou
@OliLay Does using a [`BufferedOutputStream`](https://github.com/apache/arrow/blob/524a463207fdb799b2cd784b7ef95052165882ec/cpp/src/arrow/io/buffered.h#L38) fix the problem for you? Though I agree this should probably be handled transparently inside the Azure FS implementation. @felipecrv
Issue resolved by pull request 43096 https://github.com/apache/arrow/pull/43096
A side effect of providing GH releases would be that one can can get notified using the right GH "Watch" settings for the repository. I've enabled Watch settings for new...
> I don't think we have an explicit status or result benchmark We do, but for some reason I had stuffed it in `type_benchmark.cc` https://github.com/apache/arrow/blob/1b4080032a86cacd9b21988514438ac3d5696d81/cpp/src/arrow/type_benchmark.cc#L545-L555
Yes, the code changes probably trigger random variation in code placement in these benchmarks, that could explain the false regressions.
@github-actions crossbow submit -g cpp
CI failures are unrelated, I'll merge.
@rok I'm taking a look. Thanks for being persistent :-)
@rok Github claims there are conflicts below ("This branch has conflicts that must be resolved"). Can you take a look?