⚠️ GitHub issue #37102 has been automatically assigned in GitHub to PR creator.

@mapleFU Does this solve the issue in the linked issue? Can you take the test from the issue and add it as a test to confirm this solves the problem here?

I think it fix the problem. I've add a test here.

I have a long time didn't using Golang, so maybe I didn't follow the best style, and It's a bit late in UTC-8, so feel free to edit this patch directly. @zeroshade

@mapleFU Thanks much! I'll take a look. at first glance it looks good

I'm afraid thats a bit hard for me.

The reason is that RLE also has this problem (see RleEncoder and FlushLiteralRun in cpp code). So almost io.WriterAt should be refactor to the writer. It's ok but I've a long time didn't use golang, so maybe I will break the interface. You can edit it directly.

You can edit this patch directly or close it and create one :-)

Hi Matt, I found Reserve is confusing in our project. Some time it like C++ stl reserve(k): make sure elements upto k elements, some times it means ensure k more slots. Would it be confusing here?

And thanks so much for your help!!!

I've no problem now. Though it might get more memory than expected, the implement is right. So it LGTM, thanks Matt!

After merging your PR, Conbench analyzed the 5 benchmarking runs that have been run so far on merge-commit 6b810cb.

There were no benchmark performance regressions. 🎉

The full Conbench report has more details. It also includes information about possible false positives for unstable benchmarks that are known to sometimes produce them.