afalcone@mbunix.mitre.org (Falcone) (03/19/91)
TWIMC,
I have been having trouble running "at" in 32-bit Minix. Having
recompiled cron, atrun and at, I first attempted to use "at" with stdin
attached to the terminal (i.e., by typing in a script to be run at
some future time, and then terminating said script with the usual ^D).
However, when I completed the script, and wished to exit, typing the
^D did nothing. Furthermore, when I interrupted "at" in order to
regain control of the screen, I found a huge (> 700K) file in
/usr/spool/at/past with the correct kind of file name (i.e.,
91.075.2300.11), but which comprised the usual environment header,
followed by blanks (as I said, about 700K worth).
I next tried to put the script in a file, and do:
# at 2300 file
but this also did nothing but sit there until I interrupted. Once
again, a mammoth file was created in at/past, having the same form as
was described before.
Has anybody else seen this kind of thing? It is puzzling, since
one would think that "at" shouldn't care whether it is running in 16
or 32-bit Minix, as it is written in C, and the compiler should
sufficiently insulate it from the underlying change in processor mode.
I thought at first it may have something to do with the virtual
consoles stuff (I have four), but "at" runs perfectly well in 16-bit
Minix with four VC's.
Any clues would be appreciated.
Thanks,
Tony Falcone (afalcone@mbunix.mitre.org <Internet>)
MTS, The MITRE Corp.