Future maintenance of RubyGems packages with native extensions

[Daisuke Fujimura]

Hello all,

I would like to discuss the future maintenance of Ruby-related
packages in Cygwin and gather feedback from the community.

Currently, I maintain not only the core Ruby packages, such as Ruby
itself, RubyGems, and RDoc, but also a number of RubyGems-based
packages that include native extensions.

However, the maintenance cost of these packages has been increasing
over time. Each Ruby upgrade typically requires rebuilding and
updating packages with native extensions to follow ABI/SOVERSION
changes. In addition, upstream changes in gem packaging, build
systems, or dependencies often require package-specific investigation
and adjustments.

These packages are also tightly coupled to Ruby releases. After a Ruby
upgrade, packages that have not yet been rebuilt can leave users in a
broken or partially unusable state, which means timely updates are
often required.

To be honest, I personally use only Ruby itself and the core
libraries. When I need additional gems, I typically install them
directly using RubyGems rather than through Cygwin packages. As a
result, I do not have much real-world usage or operational experience
with many of the packaged gems.

My impression is that many Ruby users today may follow a similar
workflow and install gems directly from RubyGems, rather than through
Cygwin Setup. Therefore, I suspect that the number of users relying on
packaged gems from the Cygwin distribution may be relatively small.
However, this is only my perception, and I would be interested in
hearing from others about actual usage and expectations.

Because of these factors, I intend to continue maintaining Ruby itself
and the core Ruby packages. However, I do not believe I can continue
providing the same level of active maintenance for RubyGems packages
with native extensions.

I have considered whether transferring some of these packages to other
maintainers would help. In practice, however, such packages still need
to track Ruby upgrades closely and often require coordinated updates
around Ruby releases. Given the communication and synchronization
overhead involved, I am not convinced that this would significantly
reduce the overall maintenance burden.

Unless there is clear demand from users, I am inclined to treat these
packages on a best-effort basis going forward and focus my maintenance
efforts on the core Ruby packages.

I would appreciate any feedback, concerns, or suggestions from the community.

Best regards,

Daisuke Fujimura