Taylor, thanks for the help! I had install mit-scheme with homebrew and will file an bug with them. After installing x86-64 binary you linked to (and noting that the command-line binary lived at /Applications/MIT:GNU\ Scheme\ 10.1.10.app/Contents/Resources/mit-scheme), the mechanics.com install went smoothly and i was able to use the mechanics code from edwin.
- Chas On Sat, Apr 11, 2020 at 4:43 PM Taylor R Campbell <campb...@mumble.net> wrote: > > Date: Sat, 11 Apr 2020 14:12:13 -0700 > > From: Chas Leichner <c...@chas.io> > > > > Sorry if this isn't an appropriate place to ask this question, I asked > > about this on #scheme and they directed me here. > > > > I'm working on reading Functional Differential Geometry and ran into a > > problem installing scmutils. I'm using a mac (Mojave 10.14.5 with MIT > > Scheme 10.1.10) which the scmutils docs say is supported, but get an > > architecture not supported error when I try to run it. > > [...] > > File has: compiled-code interface 3; architecture 14. > > Expected: compiled-code interface 3; architecture 13. > > It appears you are using the svm1 version of MIT Scheme (compiled-code > architecture 13), which is based on a byte code interpreter, rather > than the x86-64 version of MIT Scheme that gjs used to create the > mechanics.com band (compiled-code architecture 14), which uses native > x86-64 machine code. You correctly ascertained that one cannot use > another's bands. > > So either you'll need to use the x86-64 version of MIT Scheme, which > is probably what you want anyway; or if for some reason you > particularly want to use svm1, you'll need to build scmutils from > source -- maybe ask gjs how to do that since I don't know offhand. > > How did you install MIT Scheme? From > > https://www.gnu.org/software/mit-scheme/ > > you'll need either the `Unix binary' or `macOS binary' distribution; > the `Portable SVM' one will not work with the band gjs provided. If > you installed from Homebrew or MacPorts or something, they should be > changed to use the x86-64 version of MIT Scheme; there's no advantage > to the svm1 one. >
