starting point
starting point
with kermit rolling and files prepped over from raspi side to tops10 side, time this weekend to start digging into merlyn’s instructions from the file ‘changes’ in the repo.
CAN1.CMD is linker control file for building executable.
DECCMP.CMD full compile control file.
these two @com1, @com2 are splitting up and equiv to @deccmp. @can1 gets used for r link.
.compile @com1
.compile @com2
.r link
*@can1
.get decwar
.ssave
then create p,pn [1,27] and install DECWAR.EXE, DECWAR.GRP, DECWAR.HLP, DECWAR.NWS in that p,pn. then
.dir [1,27]
.r [1,27] decwar
the one already there decwar in [1,27] is definitely working. and looks like there’s also another older one already there in [5,30]
.dir [5,30]
.r [5,30] decwar
experiments with macro compile
here we’re user opr [1,2] and getting first rel files. looking in cmp.cmd, tried lofchk, warmac, cishng, msg, setmsg.
warmac, msg, setmsg seem to work.
.comp warmac
MACRO: WARMAC
% .SGNAM will be obsolete; please use GETTAB .GTRFN instead.
% .SGPPN will be obsolete; please use GETTAB .GTRDI instead.
% .SGDEV will be obsolete; please use GETTAB .GTRDV instead.
EXIT
.comp msg
MACRO: msg
EXIT
.comp setmsgmsg
MACRO: setmsg
EXIT
.dir
WARMAC REL 109 <057> 7-Sep-24
MSG REL 21 <057> 7-Sep-24
SETMSG REL 3 <057> 7-Sep-24
Total of 1541 blocks in 142 files on DSKB: [27,101,DECWAR]
lofchk and cishng not, but actually they seem not to appear much elsewhere. bypassing these two for now, probly not needed!
.compile lofchk
MACRO: lofchk
?MCRCFU CANNOT FIND UNIVERSAL SYSMAC
EXIT
.comp cishng
MACRO: cishng
?MCRCFU CANNOT FIND UNIVERSAL SYSMAC
EXIT
experiments with fortran compile
some files seem to go
.comp param
FORTRAN: PARAM
00193
%FTNNED No END statement in program
%FTNWRN MAIN. No fatal errors and 1 warning
.dir
PARAM REL 1 <057> 8-Sep-24
Total of 1543 blocks in 145 files on DSKB: [27,101,DECWAR]
some not
.comp decwar
FORTRAN: DECWAR
?FTNFWE Line:00239 Found STATEMENT END when expecting IDENTIFIER
00240*
?FTNNRC Line:00240 Statement not recognized
00241* + hfz, shpcon, shpdam, base, nbase, board, locpln, locr,
00242*
...snip............
?Internal Compiler Error
?Memory Protection Violation at location 453034 in Phase FORTB
?while processing statement 00870
.
would guess that because param.for has no ‘include’ statements, it’s simpler and working. other words, the ‘include’ statements are causing trouble!
learnings around fortran versions
.r fortra
*^C
.version
FORTRA 6(1144) +
.r fortra
*/h
FORTRA.HLP -- HELP file for FORTRAN-10 version 6 December 1981
tape files at https://pdp-10.trailing-edge.com/
FORTRAN-10 V6.0 1981 3.52 Mbyte 385
FORTRAN-10_V7wLink_Feb83 4.96 Mbyte 502
Fortran-10 Version 11 5.54 Mbyte 432
https://github.com/PDP-10/its/issues/170#issuecomment-414909112 philbudne commented on Aug 22, 2018 F10/FORTRA (FORTRAN-10) was FORTRAN-IV (FORTRAN 66) until version 7, which supported the FORTRAN-77 standard. I seem to recall finding version 1 of F10 in ITS’ DECSYS directory. I can’t IMAGINE anyone ever running v7 on ITS, and I wouldn’t be shocked if it used TOP-10 monitor calls not supported by DECUUO.
current pic is that we’re on dec fortran-10 v6 from 1981, which makes sense historically for decwar. this means a dec superset of fortran IV / f66. it meets the f66 ansi standard, and has dec extensions. we know for a fact that v5 did the same and already had the ‘include’ statement. highly likely that ‘include’ is one of the dec extensions to f66! here’s a nutshell of current learnings for decwar context
dec fortran-10 v5, 1977, fortran iv / f66, plus dec extensions
dec fortran-10 v6, 1981, fortran iv / f66, plus dec extensions
dec fortran-10 v7, 1983, f77, ?plus dec extensions, !probly!?
does decwar.com matter for compiling?
turns out, nope! this is a ‘comments’ file! com == comments. but understanding this did lead onwards to decwar.imp, where probably imp == important! decwar.com is additional comments, on top of decwar.imp and decwar.tap
decwar.imp has mucho learnings about ‘building decwar from tape’
this is clearly some important stuff. hence the ‘.imp’ extension. and note it’s first thing listed in decwar.tap file. in fact, decwar.imp and decwar.tap match up perfectly.
one useful thing could be that [10,30,DECWAR] was an important folder / sfd at some point, probably for the build from tape described in decwar.imp, so p,pn 10,30 and sfd [10,30,decwar] were important back then, and could still be! in fact it’s even possible that if we were 10,30 and in [10,30,decwar] then everything would magically work. this may be something to try asap.
for our possible ‘include’ troubles, note especially
Files INCLUDEd in DECWAR.FOR (extracted from WARMAC.MAC by TECO macros):
PARAM.FOR Parameters (constants).
HISEG.FOR High segment common block.
LOWSEG.FOR Low segment common block.
Tools used to build game (.SAV files generated from TECO macros):
PARAM.SAV Build PARAM.FOR from WARMAC.MAC.
HISEG.SAV Build HISEG.FOR from WARMAC.MAC.
LOWSEG.SAV Build LOWSEG.FOR from WARMAC.MAC.
EXTERN.SAV Build EXTERN.FOR from MSG.MAC.
SETEXT.SAV Build SETEXT.FOR from SETMSG.MAC.
TECO macros used to build game (require TECO 124):
PARAM.TEC Source for PARAM.SAV.
HISEG.TEC Source for HISEG.SAV.
LOWSEG.TEC Source for LOWSEG.SAV.
EXTERN.TEC Source for EXTERN.SAV.
SETEXT.TEC Source for SETEXT.SAV.
hmmm, don’t have teco 124, potential trouble? wasn’t for merlyn! but potentially could be for a ‘build from tape’?
.teco param.tec
[2K Core]
*^C
.version
TECO 24A(235) +