The Design of a Self-Compiling C Transpiler Targeting POSIX Shell | Hacker News Reader

The Design of a Self-Compiling C Transpiler Targeting POSIX Shell

dl.acm.org

30 pointsfanf26 comments

Worth checking out https://github.com/fosslinux/live-bootstrap which starts from a handwritten hex bootloader and later a Scheme interpreter that implements a C compiler that compiles tcc

Full steps on https://github.com/fosslinux/live-bootstrap/blob/master/part...

I don't really understand why this would be better than using tcc, which also can be used to bootstrap gcc, as a trusted prebuilt binary. Is there any reason to think that posix shell is more trustworthy? I certainly would expect c code to be more human readable than shell for writing a compiler.

Isn't this useful for obscure embedded systems where you have a posix shell but no working compiler (that you have access to)?

This compiler is written in C though. You also have the option to distribute the "compiled" version which is a POSIX shell script. I can imagine instances where I can't grab a TCC binary or maybe I'm not allowed to set the exec bit.

There is a problem if you want to verify using the C source though (assuming the resulting shell script is as inscrutably as binary executable). Even if you use pnut.sh to compile pnut.c and see that you get the correct output, or even include GCC in the chain, you're still (afaict) susceptible to a "trusting trust" attack on the original pnut.sh .

Author here,

> assuming the resulting shell script is as inscrutably as binary executable

It's quite the opposite, pnut generates shell code that's close to the original C code to make it easy to audit the code. A useful way to see pnut is as a tool that rewrites C code to POSIX shell, without significantly changing the structure.

This means that even if GCC is required for the initial compilation of pnut (GCC compiles pnut, then pnut compiles itself and we get the pnut-sh.sh script), the script can be "sanitized" from trusting trust attacks by simply comparing the script to the C code and making sure GCC hasn't introduced any malicious code.

Page 10 of the SLE24 presentation has a tombstone diagram showing the compilation steps to go from pnut's C code to a GCC binary: https://github.com/udem-dlteam/pnut/blob/main/doc/presentati...

Awesome. I guess I should have looked at the code rather than just speculating, but I couldn't imagine that it works _and_ it's readable.

Reply on news.ycombinator.com