[TUHS] What's In a Prompt String?

[segaloco via TUHS]

On Tuesday, April 28th, 2026 at 15:50, Steffen Nurpmeso via TUHS <tuhs at tuhs.org> wrote:

> segaloco via TUHS wrote in
>  <hy0Wz3aAx-kcLtMuf4ojqunmhwc5V_iry8qO2R7WaLU0nJ5PNOLToEUmNBwun3Y3bG7Hv13\
>  o91QOGmbHRUg9fnqZZ8xok80997Ywpqr_Ggc=@protonmail.com>:
>  |preserved "s1" sources of V3 are pretty close to my restored V2 work \
>  |except that they used the enhanced mathematical capabilities of the \
>  |11/45 directly (whereas the V1 and V2 programs use the memory-mapped \
>  |EAE with address symbols provided by the assembler).
> 
> I am not busy looping on this.  But you are not talking about the
> prompt of the shell?  (I in turn would be interested in evolutions
> of the mail program, McIlroy wrote ~"never satisfied with its
> exact behaviour" and mentioned everyone was hacking on it; in 2015
>
> --steffen
>

Just pointing out that there is a lot in common between V2 and V3 and the transition from EAE to PDP-11/45 arithmetic is a good tell for that transition.  What I couldn't tell you is if there was an assembly-but-45 version of the shell or if the version in V3 with a % instead of @ also represents the rewrite to C.

Basically if not then we may have all the known early shells on hand now with the USG shell being dumped in PG3, encompassing all in all:

Thompson Shell V1 (V1/V2 assembly shell)
Thompson Shell V2 (V4-V6, C)
USG Shell (Program Generic)
Mashey Shell (PWB)
Bourne Shell (V7 and UNIX/TS).

For me the main unknown is if there was a distinct V3 shell straddling the line or if it was just the first release of the C-based (not csh) shell seen then in various forms (USG/Mashey) until the Bourne shell is written.

- Matt G.