fortr6.doc

TOPS-10
*******

Generally, when building a 1022 high segment for the HR system, 1022SA
adds routines from HL1022.REL onto the top of the existing high
segment (e.g., FOROTS or LIBO12), producing the 1022 version of that
high segment (1022T or 1022C2).  If this were done to FOROT6, as DEC
distributes it, the 1022 routines would occupy the area that DEC has
reserved for FOROT6's impure storage.  Since TOPS-10 high segments
must be contiguous, it isn't possible to move 1022's routines beyond
the FOROT6 impure area;  and putting 1022's routines into address
space below FOROT6 confuses Fortran.  Instead, we move the FOROT6
impure area, to make room for 1022's routines.  This is done
by relinking the FOROT6 REL files with a gap between the top
of the code (F.EES) and the start of the impure storage (F.DATA).
For 1022 version 115A, this gap must be at least 48 pages.

The loading procedure contains a sequence "GET FOROT6", "SAVE FOROT6".
This is done to set the left half of word 400003 to the high segment
length.  LINK apparently doesn't set this when loading into PSECT's.
1022SA uses this word to determine where to start loading the code
from HL1022.REL when it creates 1022T6.

TOPS-20
*******

It is possible to load a 1022 high segment for TOPS-20 similar to the one
used by TOPS-10;  however, since TOPS-20 high segments needn't be contiguous,
it isn't necessary.  The 1022T6 high segment is loaded between page 400 and
DEC's FOROT6.

If your site loads FOROT6 at some other location, you may need to relocate
1022T6 so that 1022 doesn't overlap FOROT6 or the impure area.  Also, the
impure area must be above the top of the stand-alone system (QN1500.EXE),
in order for DBEXEC to work properly.