(almost) rewrite tcopy(1) to teach it new tricks.

[phk]

Summary

This is almost a rewrite of tcopy(1) to add new functionality.

This is a continuation of D51442, where Warner's comment made me decide to go all in and rewrite the "tape dev" code as C++ classes for clarity.

I also changed the url to open-simh and removed the prompt characters from the examples as Alexander suggested.

I made SIMH-TAPFILE the default for non-tape-devices, and added -r instead of -t.

I added -l logfile option.

I appended duration and throughput to the "total" line.

I added SIGINFO support.

The exit status was and still are all over the place. I have used EX_* in new code, but left the old alone. Given that the exit status was undocumented before, I feel it would be safe to sanitize it all to EX_* values, but would like feedback on this point ?

In -c mode the two tape drives are rewound sequentially, and short of spawning a pthread, there seems to be no way to do it in parallel ?

(Please ignore the 'tst' and 'flint' targets added to the make file, I'll remove them before committing.)

[imp]

Looks mostly sane and the issues I remember flagging before have been corrected.

usr.bin/tcopy/tcopy.cc

44

In new code, it's been suggested that we include the cXXXXX variant for many of these things... Is there a reason to not do that? Or is this not 'new code' but old C code transported into a C++ class hierarchy? if the latter, never mind

432

Can we use __unused in the function definition prototype instead? Here and elsewhere?

465

Extra newlines before this line, need one after.

560

Do we need to fflush(stderr) since this is a signal handler? Also, any concern about using stdio in a signal handler? Or is that only on pedantically POSIX systems.. I forget.

574

char* -> char *

[phk]

Committed with a few more cleanups.