Issue #21371 has been updated by mame (Yusuke Endoh). Status changed from Open to Assigned Assignee changed from ioquatix (Samuel Williams) to ngoto (Naohisa Goto) ---------------------------------------- Misc #21371: Proposal to Remove SPARC Architecture Support from Ruby https://bugs.ruby-lang.org/issues/21371#change-113420 * Author: ioquatix (Samuel Williams) * Status: Assigned * Assignee: ngoto (Naohisa Goto) ---------------------------------------- We propose removing support for the SPARC architecture from Ruby. SPARC has seen little to no community usage in recent years, and its continued presence in the codebase, particularly through special cases like `rb_sparc_flush_register_windows` adds complexity without clear benefit. Removing SPARC support will reduce maintenance overhead and simplify Ruby’s internals. ## Proposal We propose removing all SPARC-specific code from the Ruby source tree, including `rb_sparc_flush_register_windows` and related build system conditionals. This change would be noted in the release documentation and targeted for inclusion in Ruby 4.0. -- https://bugs.ruby-lang.org/

Issue #21371 has been updated by ioquatix (Samuel Williams). Are there any official distributions other than OpenBSD/sparc64 that still support SPARC? There appears to be a somewhat active (unofficial?) Debian port, but I couldn't find anything else. SPARC appears to be a dead end. No new CPUs since 2017, Oracle ceased SPARC processor development after the M8, and Fujitsu has shifted its focus to ARM-based architectures, planning to discontinue SPARC production by 2029. Nothing prevents users of SPARC from continuing to use an older release of Ruby (e.g. 3.4). But since hardware is hard to come by, it's difficult to develop for. I also imagine it's a lot of work to maintain our testing environments and the cost/benefit ratio is probably not that great. Finally, I imagine the toolchain for compiling is only going to get (relatively speaking) more and more antiquated. So it may eventually hold back our usage of more advanced C language features, etc. My main exposure to this is through maintenance of coroutines, which are a bit more complicated because of shadow registers. No other architecture has as much special cased behaviour as SPARC, so it sticks out as a high cost low value architecture. If we think Ruby 4.0 is too soon to drop support, maybe we can at least plan on a later date. ---------------------------------------- Misc #21371: Proposal to Remove SPARC Architecture Support from Ruby https://bugs.ruby-lang.org/issues/21371#change-113422 * Author: ioquatix (Samuel Williams) * Status: Assigned * Assignee: ngoto (Naohisa Goto) ---------------------------------------- We propose removing support for the SPARC architecture from Ruby. SPARC has seen little to no community usage in recent years, and its continued presence in the codebase, particularly through special cases like `rb_sparc_flush_register_windows` adds complexity without clear benefit. Removing SPARC support will reduce maintenance overhead and simplify Ruby’s internals. ## Proposal We propose removing all SPARC-specific code from the Ruby source tree, including `rb_sparc_flush_register_windows` and related build system conditionals. This change would be noted in the release documentation and targeted for inclusion in Ruby 4.0. -- https://bugs.ruby-lang.org/

Issue #21371 has been updated by ioquatix (Samuel Williams). Thanks for that list. What do you think a realistic deprecation/removal timeline might look like for SPARC? ---------------------------------------- Misc #21371: Proposal to Remove SPARC Architecture Support from Ruby https://bugs.ruby-lang.org/issues/21371#change-113424 * Author: ioquatix (Samuel Williams) * Status: Assigned * Assignee: ngoto (Naohisa Goto) ---------------------------------------- We propose removing support for the SPARC architecture from Ruby. SPARC has seen little to no community usage in recent years, and its continued presence in the codebase, particularly through special cases like `rb_sparc_flush_register_windows` adds complexity without clear benefit. Removing SPARC support will reduce maintenance overhead and simplify Ruby’s internals. ## Proposal We propose removing all SPARC-specific code from the Ruby source tree, including `rb_sparc_flush_register_windows` and related build system conditionals. This change would be noted in the release documentation and targeted for inclusion in Ruby 4.0. -- https://bugs.ruby-lang.org/

Issue #21371 has been updated by ioquatix (Samuel Williams). If that's the case, we need to ensure (1) we have CI running and (2) we have resources available for development/testing. At present, I don't think we have either, but I'm happy to be corrected. ---------------------------------------- Misc #21371: Proposal to Remove SPARC Architecture Support from Ruby https://bugs.ruby-lang.org/issues/21371#change-113426 * Author: ioquatix (Samuel Williams) * Status: Assigned * Assignee: ngoto (Naohisa Goto) ---------------------------------------- We propose removing support for the SPARC architecture from Ruby. SPARC has seen little to no community usage in recent years, and its continued presence in the codebase, particularly through special cases like `rb_sparc_flush_register_windows` adds complexity without clear benefit. Removing SPARC support will reduce maintenance overhead and simplify Ruby’s internals. ## Proposal We propose removing all SPARC-specific code from the Ruby source tree, including `rb_sparc_flush_register_windows` and related build system conditionals. This change would be noted in the release documentation and targeted for inclusion in Ruby 4.0. -- https://bugs.ruby-lang.org/

Issue #21371 has been updated by kosaki (Motohiro KOSAKI). As long as the SPARC Platform Maintainer is actively maintaining SPARC code, it should not be removed. I think. ---------------------------------------- Misc #21371: Proposal to Remove SPARC Architecture Support from Ruby https://bugs.ruby-lang.org/issues/21371#change-113439 * Author: ioquatix (Samuel Williams) * Status: Assigned * Assignee: ngoto (Naohisa Goto) ---------------------------------------- We propose removing support for the SPARC architecture from Ruby. SPARC has seen little to no community usage in recent years, and its continued presence in the codebase, particularly through special cases like `rb_sparc_flush_register_windows` adds complexity without clear benefit. Removing SPARC support will reduce maintenance overhead and simplify Ruby’s internals. ## Proposal We propose removing all SPARC-specific code from the Ruby source tree, including `rb_sparc_flush_register_windows` and related build system conditionals. This change would be noted in the release documentation and targeted for inclusion in Ruby 4.0. -- https://bugs.ruby-lang.org/