Date: Tue, 9 Sep 2025 07:57:06 -0400 From: Josef 'Jeff' Sipek <jeffpc@josefsipek.net> To: Vadim Goncharov <vadimnuclight@gmail.com> Cc: Norman Gray <gray@nxg.name>, Mark Liam Brown <brownmarkliam@gmail.com>, freebsd-hackers@freebsd.org Subject: Re: Git haas gone wild (Rust), freebsd-update Message-ID: <aMAWErNE7apNtZGo@satis> In-Reply-To: <20250909131346.1e1011ea@nuclight.lan> References: <00202803-6a1a-44ca-b110-9f1404d2c9bc@FreeBSD.org> <CAN0SSYwGC1G8s5Ygb6rKqX2yPSoCCeJAyFk0gscBW0b94BwWRA@mail.gmail.com> <39D21672-603B-42A6-8820-F274FCC1191D@nxg.name> <20250909131346.1e1011ea@nuclight.lan>
next in thread | previous in thread | raw e-mail | index | archive | help
On Tue, Sep 09, 2025 at 13:13:46 +0300, Vadim Goncharov wrote: > On Tue, 09 Sep 2025 09:39:20 +0100 > Norman Gray <gray@nxg.name> wrote: > > > Mark, hello. > > > > On 9 Sep 2025, at 9:18, Mark Liam Brown wrote: > > > > > Or switch to Mercurial, https://www.mercurial-scm.org/ and declare git > > > as obsolete > > > > I think that's an excellent idea, for both technical and social reasons! > > > > (It's worth noting, though, that the Mercurial developers have, for the > > past couple of years, been incrementally and _very_ carefully replacing > > performance-critical Python components with Rust equivalents > > <https://www.mercurial-scm.org/help/topics/rust>. The words > > ‘incremental’ and ‘careful’ are pretty much hallmarks of > > Mercurial's development history.) > > So again, no actual alternative Part of the 'carefully' is that the rust is *not* required and you can build without it. The parts that are in rust all have fall-back code written in python and often have an equivalent in C as well. Put another way, for many years now, hg has been python + optional C perf-improving code. They are adding python + optional rust. (So, at the moment you can build in 3 different ways: pure, with C, with rust.) I don't know what the long-term plan of the C code, but it'll no doubt consider the userbase's needs. The hg devs understand (and this is my paraphrasing/observation) that they are good at writing python and not so good at writing C. To oversimplify things, they don't trust themselves to maintain C code of sufficient quality for an scm. I think this realization is a good thing, and I applaud them for it. That said, switching scms is a *lot* of work, so this discussion is (IMO) moot. Jeff.
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?aMAWErNE7apNtZGo>