hcrosse opened a new pull request, #1617: URL: https://github.com/apache/arrow-go/pull/1617
### Rationale for this change View builders can produce arrays that pass `ValidateFull` but panic on read because the end offset overflows `int32`. ### What changes are included in this PR? `BinaryView.Value`, `StringView.Value`, `ViewHeader.getBufferBytes` and `validateUtf8View` compute the end in `int`, so valid arrays from earlier releases read correctly. The builder also stops filling a block at `MaxInt32` bytes, so earlier releases can read its output. Calling `SetBlockSize` with a value above `MaxInt32` could otherwise produce negative offsets that no reader change can fix. ### Are these changes tested? Yes. I added tests for block boundaries, oversized blocks, reading view values, equality and UTF8 checks during casts. ### Are there any user-facing changes? Yes. View values whose end is past `MaxInt32` in a data block now read correctly. There is no API change. Closes #1614. -- This is an automated message from the Apache Git Service. To respond to the message, please log on to GitHub and use the URL above to go to the specific comment. To unsubscribe, e-mail: [email protected] For queries about this service, please contact Infrastructure at: [email protected]
