Re: [Thought experiment] Bringing swift into an experimental branch?
- In reply to: Rob Wing : "Re: [Thought experiment] Bringing swift into an experimental branch?"
- Go to: [ bottom of page ] [ top of archives ] [ this month ]
Date: Mon, 16 Jun 2025 23:47:13 UTC
On Mon, Jun 16, 2025, 5:33 PM Rob Wing <rob.fx907@gmail.com> wrote: > can someone (who was around at the time) school us on why perl was > removed from base? Short summary: Evolved too quickly, so stable branches were hard; A lot of work to keep up to date; A real ecosystem that should have had parts imported, but that was impractical; and conflicts with port installed versions. and why introducing language XYZ won’t suffer the same issues/fate? > Indeed. That's always the question. But we have external toolchain support now that would help. Warner On Monday, June 16, 2025, Tomek CEDRO <tomek@cedro.info> wrote: > >> On Tue, Jun 17, 2025 at 12:05 AM Jordan Hubbard >> <jordan.hubbard@gmail.com> wrote: >> > I know that the topic of “bringing additional languages” into OSes is >> controversial and, frankly, we made a bad call with macOS when we brought >> the kitchen sink in (everything from Tcl to Python) and regretted it >> afterwards, so this is just a thought experiment, as noted in the subject. >> > >> > All that said, I think a more modern language than C++ being readily >> available has a useful way of shifting development and generating some >> excitement, and a modern language that is also strongly supported by your >> existing clang / LLVM compiler toolchain represents an increment along the >> same evolutionary path. It’s also a language that needs a number of >> components to be tightly in sync (libdispatch, clang, llvm, various >> compilation tools and libraries, etc) before any 3rd party developer can >> even compile it for FreeBSD, so saying “it’s a ports collection problem” >> may ultimately work, but not without considerable pain, as evidenced by the >> fact that ports/lang/swift does not exist. >> > >> > Why swift, if anything? >> > >> > Python changes too often in incompatible ways and Rust is still going >> through its Rust foundation stewardship changes (that didn’t form until >> 2021) plus if you bring Rust in, you’ll be immediately enmeshed in the >> “should we just rewrite the kernel in rust?!” topics that seem to plague >> Linux. Swift, on the other hand, is an application and service >> development language and sits nicely on top of what FreeBSD already has - >> nobody’s going to push to rewrite the kernel or base utilities in Swift, >> but you do get access to a very active community and, of course, the >> backing of a major player like Apple and the open source swift ecosystem. >> > >> > I’m told that Swift also just announced support for FreeBSD, so if >> there is ever a time, this might be it? >> > - Jordan >> >> There is nothing wrong with resurrecting lang/swift in the ports [1], >> this may bring Swift lovers to FreeBSD make it base for their servers >> / platforms / devices (maybe even porting macOS/iOS applications???), >> just as it is with any other programming language already here (i.e. >> Python), people will have choice whether use it or not with zero >> impact on the kernel / base :-) >> >> The problem was with "dump old bad C and rewrite everything in Swift >> just to get modern" because we clearly do not want to rewrite kernel / >> base in Swift :D :D >> >> [1] https://www.freshports.org/lang/swift/ >> >> -- >> CeDeROM, SQ7MHZ, http://www.tomek.cedro.info >> >>