lib/57867: Why does cbreak or raw affect terminal output differently?

[achowe%snert.com@localhost]

>Number:         57867
>Category:       lib
>Synopsis:       Why does cbreak or raw affect terminal output differently?
>Confidential:   no
>Severity:       non-critical
>Priority:       low
>Responsible:    lib-bug-people
>State:          open
>Class:          support
>Submitter-Id:   net
>Arrival-Date:   Sat Jan 20 15:15:00 +0000 2024
>Originator:     Anthony Howe
>Release:        9.3
>Organization:
>Environment:
NetBSD elf.snert.com 9.3 NetBSD 9.3 (GENERIC) #0: Thu Aug  4 15:30:37 UTC 2022  mkrepro%mkrepro.NetBSD.org@localhost:/usr/src/sys/arch/amd64/compile/GENERIC amd64

>Description:
In an effort to create a basic regression test suite for a program by capturing screen output for a given limited `TERM` (ansi-mini, vt52, adm3a), I came across an oddity that I have not yet understood (even after looking at `libcurses/tty.c`).

Q: Why does `cbreak()` generate slightly different terminal output from `raw()` for the same `TERM`?

I figured it something with respect to `OPOST`, but have yet to grok what is happening and how I might generate a minimal custom `TERM` that forces consistent output regardless of `cbreak()` or `raw()` (noecho() and nl()).

Thank you for your time and attention.

>How-To-Repeat:
Currently do not have a simple demonstration of above.
>Fix: