[TUHS] Fortran question for Unix System-5 r3

[Paul Winalski]

[already answered in COFF, but since it is Unix-related I'll answer
here as well]

On 3/26/23, KenUnix <ken.unix.guy at gmail.com> wrote:
> Fortran question for Unix System-5 r3.
>
> When executing fortran programs requiring input the screen will
> show a blank screen. After entering input anyway the program
> completes under Unix System V *r3*.
>
> When the same program is compiled under Unix System V *r1* it
> works as expected.
>
> Sounds like on Unix System V *r3* the output buffer is not being flushed.
> I tried re-compiling F77. No help.

The bug is almost certainly in the Fortran run-time library (RTL), not
in the compiler, so rebuilding the compiler won't help.

As a compiler engineer for many years, this regression bug got me
curious as to the software development processes in place for Unix
System V.  How was testing and release qualification done for f77 and
other utilities released with System V?  Was an automated test system
used?  Were reported and fixed bugs turned into regression tests?

Apropos this regression bug in particular, what testing of interactive
functionality was done, and how was it done?  Automating interactive
testing can be cussedly difficult.

-Paul W.