Hi Sebastien,

Given that Cargo is so much less secure than Go (no
equivalent of a module proxy), it might
be safer to let the Rust project be the "lead" or "top level"
project, and have the Rust build script; these

https://doc.rust-lang.org/stable/cargo/reference/build-scripts.html

pull in any Go dependencies. 

Otherwise automatic Rust invocation from a Go package install 
would appear/seem to be introducing alot of security issues for
Go packages, right?

As least that how it appears at the moment... but I've only been learning 
Rust 
for a week, so I have pretty limited perspective.

Jason

On Wednesday, October 29, 2025 at 9:14:36 AM UTC Sebastien Binet wrote:

hi there, 

Right now, wrapping C, C++ and Fortran code is somewhat easy: just slam 
some .c, .cxx or .f files within your Go module and the go command will 
pick them up and compile those as part of the final binary. 
This allowed "go-get"-ability for, e.g., mattn/go-sqlite3. 

It doesn't seem like the same thing is possible for Rust code. 

Right now, the perhaps "easiest" thing is to ship the compiled Rust code 
(.a, .so or .syso) with your Go project, but you get all the complications 
with shipping possibly large binaries with your VCS (`cmd/go` and 
proxy/sumdb support for Git-LFS is a bit sketchy). 

Am I the only one who wants to wrap Rust code ? 
Are there enough people who'd like to do just that to warrant some kind of 
support from the Go ecosystem ? 
Also, if we were to add support for Rust, do we use rustc as a compilation 
vehicle or Cargo ? (the latter might open a rather big can of worms with 
its ability to execute arbitrary code via the 'build.rs' file, but... Cargo 
is now *the way* to build Rust code, AFAICT) 

WDYT ? 

cheers, 
-s 

-- 
You received this message because you are subscribed to the Google Groups 
"golang-nuts" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion visit 
https://groups.google.com/d/msgid/golang-nuts/cbb05218-556c-4295-8f0c-15ee48bd0d42n%40googlegroups.com.

Reply via email to