00100	*****THE TRUTH ABOUT PUB*****
00200	
00300	ENTRY 9-JAN-73 LE03	FILLER:BUG FIX
00400		FOUND REASON [<E>]LABEL DIDN'T WORK: IN FILLER.SAI,
00500	IN PROCEDURE TEXTLINE, CASE 1,
00600		LEN←E("5",0)  SHOULD HAVE BEEN  LEN←CVD(E("5",0)).
00700		I FIXED IT IN FILLEK AND FILLER.
00800	
00900	
01000	ENTRY 29-DEC-72 PK01	FILLEK.SAI AND WHY IT EXITS
01100		IN AN ATTEMPT TO GET SOME  SORT  OF  HYPHENATION  BUILT  INTO
01200	FILLER  SO  THAT BLANKS WOULD NOT EXPAND SO MUCH, I HAVE ATTEMPTED TO
01300	IMPLEMENT AN "ALGORITHM" GIVEN TO ME BY RAJ.   I  CREATED  FILLEK.SAI
01400	AND PUBDFK.SAI SO THAT I WOULD NOT WRITE OVER ANY GOOD STUFF.
01500	
01600	ENTRY 21-SEP-72	LE03	PUB.PUB AND FRIENDS.
01700		PUB.PUB IS LATEST FREEBED VERSION.  HAS CMU SWITCH  ON  FIRST
01800	LINE WHICH IS YET UNTESTED (I.E. HAVE ONLY MADE A STANFORD VERSION SO
01900	FAR).
02000		PUBO.PUB IS VERSION WE USED FOR EARLIER CMU MANUAL-- PROBABLY
02100	SHOULD BE DELETED.  PUBS.PUB IS WHAT  STANFO  JUST  SENT  US  AND  IS
02200	INCORPORATED IN PUB.PUB.
02300		PUB.PUB REQUIRES PUBHDR.PUB, FOR NO GOOD REASON.
02400	
02500	
02600	ENTRY 08-JUL-72 RJ14	MANY MORE THINGS WORK!
02700	
02800		1) CHANGED THE INTERFACE TO THE XGP.  NOW PUB USES THE NORMAL
02900	↓_ TO SIGNAL CHANGE OF KSET.  XGP CHARACTERS ARE QUOTED CORRECTLY.
03000		2) SIGNAL RESPONSES WITH  MULTIPLE  ARGUMENTS  NOW  WORK.  (I
03100	DOUBT THAT ANYONE KNEW THEY DIDN'T!).
03200		3) EVEN, ODD, MAX, AND MIN NOW WORK...MOD STILL UNDEFINED.
03300		4) DEFAULT OUTPUT EXTENSION IN XCRIBL MODE IS XGO.
03400		5) ALPHABETIZING  ON  RECEIVE  NOW  WORKS.   I  REPLACED  THE
03500	QUICKERSORT  ALGORITHM WITH A BRUTE FOREOCE ALGORITHM AND EVERY TH[NG
03600	WAS FINE.
03700		6)  SENDING  PAGE NUMBERS DOESN'T WORK RIGHT, BUT IT'S PRETTY
03800	EASY TO GET AROUND.  JUST ASSIGN THE VALUE TO  ANOTHER  VARIABLE  AND
03900	SEND THAT.
04000	
04100	
04200	ENTRY 04-JUN-72	RJ14	MERGED XCRIBL & STANDARD VERSION ON CUSP!!!
04300	
04400		ONE OF THE TWO ERRORS MENTIONED IN MY ENTRY OF 24-MAY WAS  IN
04500	THE  WRONG  PLACE!   I  CORRECTED  THE SYMPTOM RATHER THAN THE CAUSE.
04600	CORRECT FIX IS  NOW  IN  COMBINED  XCRIBL/STD  VERSION.  THE  CHANGES
04700	REQUIRED TO MAKE FOOTNOTES WORK WERE:
04800	
04900		1) CHANGE MAIN TEST IN FIND_ROOM TO  MOVE  PINE  OUTSIDE  THE
05000	    PARENTHESIZED EXPRESSION.
05100		2) ADDED CODE IN FIND_ROOM TO SET FOOTTOP TO TRUE IF PAGE  IS
05200	    OVERFLOWED WHILE DOING FOOTNOTES.   THIS LETS FOOTSEP GET PRINTED
05300	    ON THE NEW PAGE.
05400		3)  CHANGE  TEST  FOR  OUTPUT  OF  FOOTSEP SO THAT FOOTTOP IS
05500	    TESTED THERE (TO CATCH VALUE SET BY CHANGE 2).
05600	
05700		4) ADD A CALL TO ASSURE AREA IN SEND JUST BEFORE  SENDING  TO
05800	    FOOT.   SOMETIMES  THERE  WAS  NO  AREA OPEN AND AN ILLEGAL ARRAY
05900	    INDEX RESULTED.
06000		THE XCRIBL AND THE STANDARD VERSIONS HAVE BEEN COMBINED.  PUB
06100	MAY BE RUN IN XCRIBL MODE BY USING A /X SWITCH IN THE COMMAND  STRING
06200	OR  BY  SAYING ".DEVICE XGP" IN THE MANUSCRIPT. (NOTE THAT THE DEVICE
06300	COMMAND IS NOT  DOCUMENTED  IN  OUR  CURRENT  MANUAL.)  THE  INTERNAL
06400	VARIABLE  XCRIBL  WILL BE TRUE WHEN PUB IS RUNNING IN XCRIBL MODE AND
06500	CAN BE TESTED BY PUB CONDITIONALS.  THE VALUE IS READ ONLY.
06600		SEVERAL  MINOR  CHANGES RELATING TO THE HANDLING OF PPN'S AND
06700	LOWER CASE ANSWERS HAVE BEEN IMPLEMENTED IN BOTH PASS  ONE  AND  PASS
06800	TWO.
06900	
07000	
07100	ENTRY 24-MAY-72	RJ14 & LE03	NEW PUB ON CUSP!!
07200		1) OLD EVERYTHING FROZEN ON TAPE SP7271.
07300		2) CHANGED PUBSTD.DFS DEFAULTS TO BE:
07400			PAGE FRAME 55 HIGH.
07500			HEADING 1 TO 3.
07600			TEXT 4 TO 53
07700			FOOTING 55.
07800		3) EVEN, ODD, AND MOD STILL DON'T SEEM TO WORK?
07900	
08000	
08100	ENTRY 24-MAY-72	RJ14
08200		CORRECTED TWO LOGICAL ERRORS IN PUB.SAI, FOOTNOTES WORK  NOW.
08300	CHANGES  ARE  ON  PAGES  32  AND 33 OF THE SOURCE AND ARE MARKED WITH
08400	"CMU".   ORIGINAL  NEW   VERSION   OF   PUB.SAI   IS   DSKD:PUBA.SAI.
08500	DSKD:PUB.SAV HAS THE CHANGES.
08600	
08700	
08800	ENTRY 23-MAY-72	LE03&RJ14	?
08900		FOOTNOTES DON'T SEEM TO WORK, STILL.
09000		PUBX.PUB & PUBX.DOC ARE THE MANUAL  FROM  11-APR-72  COMPILED
09100	USING   THE  PUB  OF  20-MAY,  WITH  A  BASTARDIZED  VERSION  OF  THE
09200	PUBSTD.DFS.(I.E. THE CURRENT VERSION OF PUBSTD.DFS.)
09300	
09400	
09500	ENTRY 20-MAY-72	LE03	LOTS OF WORK & SUCCESS.
09600		MADE NEW PUB FROM "NEX:" - FROM STANFORD 12-MAY.  THINGS
09700	SEEM TO WORK PRETTY WELL.  HAVE NOT PUT ON CUSP YET, THOUGH.
09800		NEW SOURCE IS ON DSKD.  COR FILES:
09900		COR	ORG-DSKB(CURRENT--DEC-71).
10000		CORX	NEW-NEX
10100		CORJ	ORG-NEX
10200		CORK	COR,CORJ
10300		CORL	EDITED VERSION OF CORK.
10400	
10500	
10600	ENTRY 5-3-72	LE03 (HB)	WORKING ON NEW PUB
10700		CREATED NEW SOURCE FILE FROM DCS#4 USING SOUP.  DOESN'T  WORK
10800	YET.
10900		NEW FILES ARE ON DSKD, WITH MOSTLY THE SAME  NAMES.   ON  THE
11000	SAVE  MAG  TAPE,  THEY HAVE THE SAME NAMES TOO, THEY SHOULD BE AT THE
11100	END.
11200	
11300	
11400	ENTRY 4-17-72	LE03	RECOMPILED--ON CUSP
11500		BECAUSE  OF  NEW SAIL ROUTS, HAD TO RELOAD.  PUB.SAV
11600	TO CUSP.
11700	
11800	
11900	ENTRY 28-DEC-71	LE03	SAIL BUG FOUND, PROGRAMMED AROUND.
12000		SAIL    IS   NOT   HANDLING   DPB(<STRING>,<MUMBLE>)
12100	PROPERLY,  WHICH  PUB  WAS  DOING  AT  5000/34  &  5100;   I
12200	PROGRAMMED  AROUND  IT BY ASSIGNING THE STRING TO AN INTEGER
12300	(THEREBY ABSOLUTELY FORCING THE GETCH TO TAKE PLACE).
12400		NOTICE  THAT THIS LOCATION IS THE INITIALIZATION FOR
12500	EVEN, ODD,MAX, AND MIN.    I  HAVEN'T  TESTED  ANYTHING  OUT
12600	(OTHER  THAN TO ASCERTAIN THAT THE RIGHT CODE WAS COMPILED),
12700	BUT I EXPECT THAT THOSE OPERATORS WILL NOW EXIST.
12800		HAPPY NEW YEAR.
12900	
13000	
13100	ENTRY 22 DEC 71		JN11
13200	
13300	THE FOLLOWING TWO MACROS  HAVE  BEEN  ADDED  TO  PUBSTD.DFS:
13400	"FLIP  HEADING  (L,C,R)" WILL PRINT L-C-R ON EVERY EVEN PAGE
13500	AND R-C-L ON EVERY ODD PAGE.  "FLIP FOOTING(L,C,R)"  IS  THE
13600	CORRESPONDING FOOTING OPERATIONS.
13700	
13800	THE  "LOCAL" DECLARATION OF VARIABLES IN PUB DOESN'T SEEM TO
13900	WORK.   I USED IT IN A MACRO AND GOT "UNDEFINED  INDENTIFIER
14000	LOCAL" JUST ABOVE (OR ON) <LINE WHERE MACRO WAS CALLED>. THE
14100	IDENTIFIER  FOLLOWING  THE  "LOCAL"  DECLARATION  WAS   ALSO
14200	FLAGGED  AS  UNDEFINED.   THEN  PUB STATED IT COULD NOT MAKE
14300	SENSE OUT OF THE REMAINING MACRO BODY ("NOFILL, TURN ON "\",
14400	TABS 40; ONCE CENTER).
14500	
14600	A  MORE  SOPHISTICATED  "PINDEX"  MACRO  WAS  WRITTEN  WHICH
14700	ASSIGNS  THE  VALUE  OF  "INAME"  TO  ANOTHER  STRING:  ⊂ IF
14800	"INAME"≠NULL THEN GARBLE←"INAME" ELSE GARBLE←"ARG";  <<OTHER
14900	TESTS>> GARBLE; "∞.\"[]ARG ⊃ THIS WORKS. DON'T ASK ME WHY.
15000	
15100	
15200	ENTRY 17 DEC 71		JN11
15300	
15400	THE FOLLOWING MACRO DOESN'T SEEM TO WORK:
15500	
15600	.MACRO PINDEX(ARG,INAME) ⊂ IF "INAME"≠NULL THEN "INAME" ELSE
15700	"ARG"; "∞.\"[]ARG ⊃
15800	
15900	FOR INAME THE STRING "Non-ARDS OUTPUT".  THE WORD "Non" GETS
16000	DROPPED  .   REPLACING  HYPHEN  BY α- OR BLANK DOESN'T HELP,
16100	EITHER.  IT WORKS FINE FOR OMITTED INAMES.
16200	
16300	AFTER USING THE ASDUMP PROGRAM, WE FIND THAT PUB2 OUTPUTS  A
16400	CR-CR-LF  TRIPLE  FOR EACH LINE! THIS SEEMS TO SLOW DOWN THE
16500	LPT.     PRESUMABLY THIS ALLOWS  FOR  UNDERSCORING,  BUT  IT
16600	SHOULDN'T ALL GO OUT.
16700	
16800	CORRECTION  TO  A  COMMENT MADE BELOW:  THE "AT NULL" PHRASE
16900	DOES WORK CORRECTLY.  HOWEVER, REDUNDANT BLANK LINES ARE NOT
17000	REPEAT NOT RETAINED.
17100	
17200	AN ATTEMPT TO REDEFINE THE WORD "EVEN" TO BE THE NULL STRING
17300	RESULTED IN AN ERROR MESSAGE TO THE EFFECT THAT THE RESERVED
17400	WORD  EVEN  COULD NOT BE REDEFINED.  SO IT CAN FIND THE WORD
17500	IN THE SYMBOL TABLE AND GET ITS ATTRIBUTES.
17600	
17700	A GLITCH HAS BEEN ADDED TO ALLOW TESTS FOR EVEN AND ODD. THE
17800	MACRO  EVENX(ARG) HAS BEEN ADDED TO PUBSTD.DFS.    THIS SETS
17900	TWO GLOBAL VARIABLES, "WASEVEN"  AND  "WASODD"  TO  INDICATE
18000	WHETHER  THE  RESULT  OF THE TEST WAS AN EVEN OR ODD NUMBER.
18100	THESE VARIABLES MAY THEN BE USED IN  AN  "IF"  STATEMENT  TO
18200	CONTROL  ACTIONS.  THE "STANDARD TITLES" MACRO IN PUBSTD.DFS
18300	HAS BEEN REDEFINED IN TERMS OF EVENX AND NOW  WORKS.    THIS
18400	DOES NOT SOLVE OUR PROBLEM, BUT IT IS EASIER TO LIVE WITH.
18500	
18600	ENTRY 16 DEC 71		JN11
18700	
18800	THE    FORWARD    REFERENCE   ISSUE   CAUSES   PROBLEMS   IN
18900	JUSTIFICATION.  EVEN WHEN THE BRACKET SPECIFICATION OF WIDTH
19000	IS USED, THE JUSTIFICATION IS SCREWED UP.  SEE PAGE 4 OF THE
19100	FILE BLEA.DOC FOR AN EXAMPLE OF HOW BAD  THIS  CAN  BE.   IF
19200	BLEA.DOC DOESN'T EXIST, RUN PUB ON FILE BLEA.PUB.
19300	
19400	THE  OUTPUT  FROM  PUB  SEEMS  TO  TAKE MUCH LONGER TO PRINT
19500	(ACCORDING TO REPORTS).  LET'S LOOK AT PUB2 TO SEE IF IT  IS
19600	PUTTING  IN NULLS OR DOING SOMETHING ELSE FUNNY TO SLOW DOWN
19700	THE LPT.
19800	
19900	THE BLANK-LINE BREAK (WHICH IS INITIALIZED IN PUBSTD.DFS  BY
20000	AN  "AT  NULL"  REQUEST) DOESN'T WORK.   ALSO, THE "ODD" AND
20100	"EVEN" UNARY OPERATORS DON'T WORK.
20200	
20300	SEE SLIGHT CORRECTION TO ABOVE COMMENT IN ENTRY  OF  17  DEC
20400	71.
20500	
20600	AFTER  TESTING THE "MAX" AND "MIN" BINARY INFIX OPERATORS, I
20700	FOUND THAT AT LEAST THE "MIN"  OPERATOR  BOMBS  THE  PROGRAM
20800	WITH  "INVALID  INDEX  NO. 1 FOR ARRAY TEXTTYPE" BUT I CAN'T
20900	FIND ARRAY TEXTTYPE IN THE CREF!
21000	
21100	AFTER FURTHER  TESTING,  THE  "MOD"  OPERATOR  DOESN'T  WORK
21200	EITHER! F---!
21300	
21400	SEE FILE "EVEN.PUB" FOR A BUNCH OF NON-WORKING EXAMPLES!
21500	
21600	ENTRY 14 DEC 71		JN11
21700	
21800	I  JUST  SAW A COMMENT IN "PARSER" (5000/11) ABOUT RECURSIVE
21900	STRING PROCEDURE BUGS IN SAIL, WITH A COMMENT TO SEE  "DCS".
22000	ANOTHER WINNER?
22100	
22200	THE TEST LINE IN PARSER.SAI AT 600/4 (ROUTINE RDENTITY) MUST
22300	BE CHANGED TO COMPENSATE FOR THE CMU  BRACE/ALTMODE  CHANGE.
22400	CURRENTLY THIS TEST ELIMINATES THE OCTAL CONSTANT DESCRIBING
22500	RIGHT CURLY BRACKET BUT ALLOWS THE ALTMODE  CONSTANT  TO  BE
22600	INSERTED.   CONFUSION AND MADNESS WILL ENSUE IF SOMEONE USES
22700	THIS.
22800	
22900	THE "EVEN" AND "ODD" UNARY OPERATORS DON'T SEEM TO WORK.   I
23000	DON'T YET KNOW WHY
23100	
23200	LET'S  THINK ABOUT ADDING CCL LINKAGES FROM THE COMPIL CUSP.
23300	SOME OF THE STUFF IS ALREADY THERE.  WE PROBABLY HAVE TO ADD
23400	TEMPCOR GOODIES.
23500	
23600	
23700	ENTRY 13 DEC 71		JN11
23800	
23900	HEAR YE, HEAR YE! TWO BUGS HAVE BEEN FIXED  IN  PUB  ITSELF!
24000	THE  CREATION  OF A BYTE POINTER BY THE "SPECIES" MACRO USED
24100	TOO MANY QUOTES (LIKE 5) IN  THE  ARGUMENT  AND  CAUSED  THE
24200	COMPILER TO PRODUCE THE WRONG INSTRUCTIONS (""/"").  WE TOOK
24300	OUT TWO ON EACH SIDE AND GOT IT TO WORK.
24400	
24500	THE SAIL RUNTIME ROUTINES WERE FIXED BY LE03  TO  ALLOW  THE
24600	AUXILIARY  GARBAGE  COLLECTOR  ROUTINE  (SA!PROVIDE)  TO  BE
24700	CALLED WITH THE CORRECT PARAMETERS.  IT SEEMS NOW  THAT  ALL
24800	IS  WELL,  BUT MORE TESTING WILL FOLLOW.    PERHAPS THE REAL
24900	"PUBSTD.DFS" NOW WORKS!  (MORE IN THIS  EXCITING  COMMENTARY
25000	LATER).
25100	
25200	ENTRY 09 DEC 71		JN11
25300	
25400	WITH  THE  AID  OF  LE03,  THE  "PUB"  MAIN PROGRAM HAS BEEN
25500	ALTERED AS FOLLOWS:
25600		1) THE LINKAGE TO "PUB2" NOW WORKS (RUN UUO)
25700		2) THE DESCRIPTORS "SYSPPN" AND "SYSDEV" HAVE
25800		BEEN ADDED AS DEFINES.   THESE DESCRIBE THE
25900		VALUES OF THE SYSTEM PPN AND DEVICE FOR FINDING
26000		PUBSTD.DFS AND PUB2.SAV.
26100		3)   FOR   DEBUGGING,   SYSPPN   AND   SYSDEV    ARE
26200	[A700PU00],DSK:
26300	
26400	ENTRY 09 DEC 71	JN11
26500	
26600	THE FOLLOWING DFS FILES HAVE THE FOLLOWING PROPERTIES:
26700	
26800		PUBNEW.DFS	WORKS WITH "BLEA.PUB"
26900		PUBSTX.DFS	THE "MASTER COPY" OF PUBSTD.DFS
27000		PUBSTD.DFS	WHATEVER THE LAST EXPERIMENT WAS
27100	
27200	ANY .DFS FILE WITH A "BEFORE" OR "PLACE" IN IT SEEMS TO BLOW
27300	UP.     PUB ATTAINS PHENOMENAL SIZE (105K OR MORE) AND DIES.
27400	WHAT THE HELL IS HAPPENING?
27500	
27600	ENTRY 09 DEC 71		JN11
27700	
27800	THERE  IS  A  CONTROL CHARACTER, ?6 (RIGHT ARROW) WHICH DOES
27900	NOT PRINT IN THE LPT128 LISTING OF PUB.DOC.     THIS IS  THE
28000	"RIGHT  FLUSH" OPERATOR DESCRIBED ON PAGE 2-5 OF THE CURRENT
28100	(5 OCT 71) VERSION OF THE DOCUMENT.
28200	
28300	ENTRY 08 DEC 71		JN11 (WHO ELSE?)
28400	
28500	ALL  '176  CHARACTERS  (STANFORD  RIGHT  BRACE)  HAVE   BEEN
28600	REPLACED BY '175 CHARACTERS (CMU RIGHT BRACE).
28700	
28800	ALL  DEFINITIONS  OF  "ALTMODE" OR ANY OTHERS WHICH HAVE THE
28900	SAME VALUE (STANFORD '175) HAVE BEEN CHANGED SO THAT ALTMODE
29000	IS '176.     THIS INSURES CONSISTENCY WITHIN PUB.  '176 IS A
29100	FLAG CHARACTER TO BOTH PUB AND PUB2 WHICH INDICATES  SPECIAL
29200	ACTIONS.
29300	
29400	A  "CMU"  MACRO HAS BEEN ADDED TO PUBDFS.SAI AND TO PUB2.SAI
29500	(WHICH DOES NOT USE PUBDFS).  USE IT TO FLAG ANY CHANGE MADE
29600	AT CMU.  IT IS EQUIVALENT TO "COMMENT".