While writing about compiler-related things, I realised for the first time that Inferno is distributed with versions of awk and sed. I don't think I was too surprised by this because both of these may well be needed to build Inferno, and they're not exactly standard issue on Windows, for example.
The surprise was that it doesn't build by default on Linux, though that is easily explained by the presence of the missing95.c file and the Windows-specific configuration in the mkfile. Getting it to work is relatively straightforward: adding dependencies on math and bio to LIBS, fixing CFLAGS to include the required header files, renaming getline symbols to avoid a clash with the host's standard library, and using __ieee754_atan2 for the missing atan2 library function.
And it works, at least for the things it is needed for when I build the native ports I am working on. The next step might be to compile it into Inferno as a module so that it can be used directly, instead of calling out to the host's implementation.
Playing with awk was a distraction from writing documentation for the Cortex-M ports which I was otherwise focused on. This is slowly taking shape as I revisit the different mechanisms and reacquaint myself with them, something which was also motivated by the deadline for work-in-progress papers for IWP9.