[C++][Packaging] Leaner package for basic geoparquet functionality

[simonff]

Describe the enhancement requested

Hi everyone,

I work on the Google Earth Engine team. We are trying to use GDAL with GeoParquet on our internal Linux machines based on Ubuntu, but we are running into problems because Arrow is currently not included into the main distro.

Debian/Ubuntu have been hesitant to carry full Arrow C++ because of size, optional features, and release cadence. I'm not an expert, but it seems that a lean, blessed profile (core + Parquet, with a few codecs; optionally Dataset) should be feasible and stable enough for distros, while keeping maintenance manageable (or at least that's what LLMs tell me).

If I understand the process correctly, Debian would own this process, but they would need advice from someone on the Arrow team on the best build options.

Does anyone have time/interest in doing this? Is there something we can do to help? E.g., I don't mind volunteering for some glue work, but I'd need to find an expert to guide me.

Thanks,
Simon Ilyushchenko

Component(s)

Packaging

[kou]

Could you join the https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=970021 thread, summarize the current status and clarify the next step? I think that it's the first step for us.

FYI: We provide our own APT repository: https://arrow.apache.org/install/

[simonff]

I emailed 970021@bugs.debian.org to join the thread.

Good to know about the custom APT repo, but I think it would be good to fix the issue globally.