This is a one-shot program. It can be linked from DRAGON, with the
first file name e.g. HOURLY or DAILY for the desired frequency. For
this to work, the file CSTACY; GUNNER LOG must exist.
The @ TECO binary is from the . directory on many backup tapes going
all the way back to 1971. The LPT 11 and MACTAP F68 sources come
unchanged from Peter Samson's "RandomSystem" DECtape. The TECO DUMMY
file is close to the undated SYSENG; TECODM 2 file.
The source files can be assembled with MIDAS 73, and linked with the
Muddle version of STINK. The link order is TECO DUMMY, LPT, MACTAP,
TECO.
This results in a binary that is identical to the original @ TECO with
a few exceptions:
- @ TECO has a loader in 20-37, and something in 776000-777777.
- MOBY is set to 174000 for 64K core, which results in MACDMP being
1777400 and MACCR being 177777. The @ TECO binary has it set to
DSKDMP at 777700.
- The buffer size in MEMSIZ is not the same.
- CONSTANTS has one additional zero at the end, shifting the rest
of the memory image up by one word.
- Update MDL 54 interpreter to look in MUDSAV;SAV 54FILE and MUDSAV;FIXUP 54FILE
for pure library.
- Update MKSVFL utility to create pure library in new location.
- Update muddle.tcl to run MKSVFL to create dummy pure library.
The TS RBYE binary starts Zork from CFS; MADADV SAVE. The assembly
info says the source file was AR1: CFS; ZORK 2. Tape backup records
show this archive file was updated on 1979-02-13, which is shortly
before the TS RBYE timestamp.
As a convenience to users, TS ZORK is linked to RBYE from SYS1 on the
assumption users will probably play the newest Zork. The older SYS2;
TS ZORK which starts the MADMAN; MADADV SAVE file is renamed to OZORK.
When a new directory is created, its QSALLO needs to be set. Failing
to do so will result in a random value, which will confuse the code to
select a dick pack for new files. Setting to -1 means the directory
has no allocation, so files can be created on any pack.
The loop is off by one, so it clears the first word in the next
process block. This turns out to be PRBFRE, the process block free
word. Since it's cleared, it's no longer considered free, so INTPRS
is forced to skip to the next block. This makes SYSSPR run out of
space for processes.
