[kou]

Rationale for this change

Apache Arrow GLib uses g_byte_new_static() for zero-copy API. If owner objects are destroyed before GBytes, GBytess refer freed data.

What changes are included in this PR?

• Use g_byte_new_with_free_func() instead to refer owner objects.
• GParquetStatistics refers GParquetColumnChunkMetadata because GParquetColumnChunkMetadata owns min/max values.
• gparquet_{,fixed_length_}byte_array_statistics_get_{min,max}(): Use (transfer full) instead of (transfer none) to avoid circular reference. Direct Apache Parquet GLib uses must call g_bytes_unref() manually. This is a backward incompatible change. GObject Introspection users don't care about this. Because GObject Introspection cares about this automatically.
• Enable ASAN tests only on Linux.

Are these changes tested?

Yes.

Are there any user-facing changes?

This PR includes breaking changes to public APIs. The return value of the following functions must be freed by g_bytes_unref():

• gparquet_byte_array_statistics_get_min()
• gparquet_byte_array_statistics_get_max()
• gparquet_fixed_length_byte_array_statistics_get_min()
• gparquet_fixed_length_byte_array_statistics_get_max()

Was AI used for this PR?

In accordance to the AI generation guidelines, please disclose below whether and how AI was used in this PR.

PR code and description written by:

• Human
• AI

Reviewed before submission by:

• Human
• AI
• Not reviewed

• GitHub Issue: [Ruby] Memory behind Arrow::Buffer#data is freed by GC while the Bytes still refer to it #54687

[Copilot]

Copilot was unable to review this pull request because the user who requested the review has reached their quota limit.

[github-actions]

⚠️ GitHub issue #54687 has been automatically assigned in GitHub to PR creator.

[kou]

@himotoyoshi Could you try this?

[himotoyoshi]

I tried this on macOS (arm64) with Ruby 3.4.1, and it fixes the problem:

irb> bytes = Arrow::StringArray.new(["hello world, a long-ish string"]).data_buffer.data
irb> GC.start
irb> bytes.to_s
=> "hello world, a long-ish string"

Thanks!

[kou]

Thanks. I'll merge this.