Google
 

Trailing-Edge - PDP-10 Archives - BB-H138B-BM - 4-documentation/edit.tco
There are 14 other files named edit.tco in the archive. Click here to see a list.
TOPS20 Change Order Number 4.1876

Written by: DBELL		 9-Nov-77 14:26:45

Edit checked: NO			   Document: NO
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT

Related TCO's:
Related SPR's: 	 11116


Problem: 
EDIT COMPLAINS ABOUT AN OUTPUT FILE ALREADY EXISTING WHEN IT
REALLY ISN'T THERE.  AN EXAMPLE IS IF THE USER HAD DEFINED
HLP: AS DSK:,HLP:.  THE COMMAND EDIT HLP:SORT.HLP SORT.HLP
WOULD THEN COMPLAIN WHEN IT REALLY SHOULDN'T.

Diagnosis: 
WHEN CHECKING TO SEE IF THE OUTPUT FILE ALREADY EXISTS, THE
MONITOR IS TOLD TO SEARCH THROUGH ALL LOGICAL NAMES, NOT
ONLY THE FIRST ONE.

Solution: 
SET THE GJ%NS BIT IN THE GTJFN AT END2+2, SO THAT THE MONITOR
WILL ONLY LOOK FOR THE FILE IN THE FIRST DEFINITION
TOPS20 Change Order Number 4.1877

Written by: DBELL		10-Nov-77 15:55:57

Edit checked: NO			   Document: NO
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT

Related TCO's:
Related SPR's: 	 10645


Problem: 
IF COMMANDS ARE BEING READ FROM A FILE, AND WE DO AUTO-SAVES,
THEN EDIT BLOWS UP WITH THE ERROR "JFN NOT ASSIGNED" WHEN THE
AUTO-SAVE IS COMPLETED.

Diagnosis: 
A RESET IS DONE AS PART OF THE SAVE THE WORLD ROUTINE, WHICH
WIPES OUT THE INDIRECT FILE'S JFN.

Solution: 
AT ENDEND+7, DELETE THE GENERAL CLOSF AND CALL A SUBROUTINE CLSALL
WHICH WILL EITHER DO THE GENERAL CLOSF, OR IF THE INDIRECT FILE
IS BEING USED, SEPARATE CLOSFS FOR ALL OTHER FILES.  ALSO REMOVE
THE RESET AT RESTRT.
TOPS20 Change Order Number 4.1878

Written by: DBELL		17-Nov-77 02:10:50

Edit checked: NO			   Document: YES
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT
   Routines Affected:	PHYSIO

Related TCO's:
Related SPR's: 	 10963


Problem: 
PROTECTIONS OF FILES GET CHANGED BY EDITING THEM.  IF NO OLD
DELETED GENERATION OF A FILE EXISTED, THE NEW FILES ALWAYS GOT
STANDARD PROTECTION.  IF A DELETED FILE DID EXIST, THE NEW FILE
GOT THE OLD FILE'S PROTECTION.

Diagnosis: 
EDIT MAKES NO ATTEMPT TO KEEP THE PROTECTIONS NICE.

Solution: 
REMEMBER THE PROTECTION OF THE ORIGINAL FILE, AND SET THE NEW FILE'S
PROTECTION TO IT ALSO.  ALSO SET THE .Q?? FILE'S PROTECTION TO THE
SAME VALUE.  THIS INVOLVES CODE AT END1B+4, END4+10, AND IN TSTNEW.
TOPS20 Change Order Number 4.1915

Written by: HESS		23-May-78 16:56:17

Edit checked: NO			   Document: NO
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT

Related TCO's:
Related SPR's: 


Problem: 
ESC not echoed on first input in new file creation.

Diagnosis: 
ALTECH flag not set until first command given.

Solution: 
Set ALTECH in create initialization.
TOPS20 Change Order Number 4.1922

Written by: DBELL		15-Jun-78 15:47:04

Edit checked: NO			   Document: NO
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT

Related TCO's:
Related SPR's: 	 11644


Problem: 
TRYING TO RUN A PROGRAM WHOSE FILE SPEC WAS AN EXACT MULTIPLE OF 5
CHARACTERS FAILED.

Diagnosis: 
THE PRARG TO WRITE THE SPEC FOR THE EXEC DIDN'T MAKE SURE A NULL BYTE
ENDED THE SPEC, THUS THE EXEC GOT CONFUSED WHEN READING IT.

Solution: 
ADVANCE THE BYTE POINTER PAST THE NULL BYTE, SO THAT IT IS ALWAYS
WRITTEN AS PART OF THE PRARG BLOCK.  AT CRFIT1-2, INSERT IBP T1.
TOPS20 Change Order Number 4.1923

Written by: DBELL		16-Jun-78 11:08:32

Edit checked: NO			   Document: NO
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT

Related TCO's:
Related SPR's: 	 11644


Problem: 
/RUN FEATURE LOSES BIG.  IT DOES NOT WORK AT ALL, AND WHEN IT IS GIVEN
A BAD FILE SPEC, EDIT GETS AN ILLEGAL INSTRUCTION.

Diagnosis: 
CODE DOESN'T KNOW DIFFERENCE BETWEEN TTY INPUT, STRING INPUT, OR
FILE INPUT (FOR THE CASES OF TERMINAL COMMAND, RESCANNED MONITOR COMMAND,
AND OPTION FILES).

Solution: 
FIX UP SETRUN AND RUNME CODE TO HANDLE ALL THESE CASES.
TOPS20 Change Order Number 4.2016

Written by: DBELL		20-Sep-78 15:14:41

Edit checked: NO			   Document: NO
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT

Related TCO's:
Related SPR's: 


Problem: 
On a copy command, if the lines get out of order, the message output is
ORD1=ORDER, which is not too informative.  It should be the standard
error message for this case.

Diagnosis: 
Code is a kludge.

Solution: 
If we get an order error, set a flag to remember it, and when the copy
is all done, if the flag is set then give the out of order message.
Also output nothing for ORD1= if out of order.
TOPS20 Change Order Number 4.2031

Written by: DBELL		 4-Oct-78 11:56:12

Edit checked: NO			   Document: NO
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT

Related TCO's:
Related SPR's: 


Problem: 
Disk full interrupts are not handled properly.

Diagnosis: 
The DSKTRP code assumes that the PC returned on a trap pointes to the
next instruction, but in fact it points to the offending instruction
itself.  Thus its "CAI" check never succeeds.

Solution: 
Look at the right instruction.
TOPS20 Change Order Number 4.2040

Written by: DBELL		10-Oct-78 16:47:48

Edit checked: NO			   Document: YES
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT

Related TCO's:
Related SPR's: 	 11698


Problem: CREATE command using a temporary file spec doesn't preserve
the temporary atribute.  Thus  CREATE FOO.BAR;T  create a file with
a high generation number, but is not temporary.

Diagnosis: 
The bit goes away.  However, creating a temporary file should be
illegal, since generation numbers mean things to a temporary file, and
we should not be randomly bumping the generation number.  This will
happen on AUTO-SAVE actions.

Solution: 
Give an error if the user is trying to create a temporary file.
TOPS20 Change Order Number 4.2042

Written by: DBELL		12-Oct-78 13:59:23

Edit checked: NO			   Document: NO
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT

Related TCO's:
Related SPR's: 	 12085


Problem: 
If no tabs is specified for a terminal, and a file is edited, the
first line of a "P" command has a space missing between the line
number and the first character of the line.  Similarly, when a <LF>
command is given, the line is missing the character.

Diagnosis: 
On display terminals, EDIT tries to be fancy and overprint the asterisk
typed on the command line.  It does this by sending the cursor up sequence
to the terminal.  The monitor takes the characters sent as motion
characters, and so on the next tab simulates fewer spaces.

Solution: 
Output carriage returns after the cursor up sequences, so the monitor
will reset its line position to the front of the line.
TOPS20 Change Order Number 4.2127

Written by: DBELL		20-Dec-78 16:07:28

Edit checked: NO			   Document: NO
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT
   Routines Affected:	DELOUT

Related TCO's:
Related SPR's: 	 12612


Problem: 
A CONTROL-C WHEN ANSWERING THE QUESTION "DO YOU WANT TO SUPERCEDE FILE"
FOLLOWED BY A QUIT COMMAND BLOWS EDIT UP WITH ILLEGAL DEVICE DESIGNATOR
ERROR.

Diagnosis: 
THE QUIT CODE CALLS THE DELOUT ROUTINE.  BUT AT THIS POINT IN TIME,
NO JFN EXISTS FOR THE OUTPUT FILE.  THUS OCRJFN CONTAINS 0, AND THE
CLOSF FAILS.

Solution: 
AT DELOUT, RETURN IF NO OUTPUT JFN EXISTS.
TOPS20 Change Order Number 4.2145

Written by: DBELL		 2-Jan-79 16:52:02

Edit checked: NO			   Document: NO
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT

Related TCO's:
Related SPR's: 	 12679


Problem: 
After an illegal insert command, such as "IX", the "P^" command
gives the error "no such lines" instead of typing the first line.

Diagnosis: 
Flag UPTOPF gets set by the insert command, and is usually cleared
afterwards.  But on an insert error, it gets left on.

Solution: 
Clear the flag on errors also.  In the routine CKIND, also clear UPTOPF.
TOPS20 Change Order Number 4.2236

Written by: GRADY		19-Apr-79 14:26:26

Edit checked: NO			   Document: NO
  TCO Tested: YES		Maintenance Release: NO
Hardware-related: NO

Program: EDIT

Related TCO's:	4.1894
Related SPR's: 	 12954


Problem: THE TRANSFER COMMAND CAUSES AN ILLEGAL INSTRUCTION TRAP
ON AN "% ILLEGAL TRANSFER DESTINATION" ERROR.

Diagnosis: THE COPY CODE SETS THE ISCOP FLAG IN THE FL FLAG WORD
TO INDICATE THAT LOCATION "COPDL" HAS THE STACK POINTER TO USE ON
ERROR RECOVERY.  "ISCOP IS SET IN 2 PLACES IN EDIT, COPY2+4 AND
OTHERF+112, BUT COPDL IS ONLY ASSIGNED A VALUE IN TE SECOND INSTANCE.
IN-CORE TRANSFER ERRORS BLOW UP, SINCE THE CODE NEVER REACHES OTHERF,
AND COPDL CONTAINS A 0.  THUS, THE STACK POINTER IS RESET TO 0.


Solution: WHEREVER ISCOP IS SET, BE SURE COPDL HAS A VALID STACK
POINTER.
TOPS20 Change Order Number 4.2319

Written by: GRADY		 3-Jul-79 12:34:52

Edit checked: NO			   Document: NO
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT
   Routines Affected:	help

Related TCO's:
Related SPR's: 	 13257


Problem: 
If device HLP: is defined as PS:<HELP.*>, or any reasonable string that includes
a wildcard, the EDIT 'H' command will fail to find HLP:EDIT.HLP.


Diagnosis: 
When EDIT gets a JFN on HLP:EDIT.HLP, it does not use the GJ%IFG bit to allow
for wildcards in the filespec, so in this case the GTJFN fails.

Solution: 
Set the GJ%IFG bit at location HELP+3, before the GTJFN, and then clear the
returned flag bits in the JFN word before saving it on the stack at location
HELP+7 : HRRZ T1,T1.
TOPS20 Change Order Number 4.2531

Written by: SIMMONS		18-Oct-79 13:36:45

Edit checked: NO			   Document: NO
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT

Related TCO's:
Related SPR's: 	 13405


Problem: WHEN IN ALTERMODE AND ONE DOES A ^C THE TERMINAL DOES NOT
ECHO WHAT YOU TYPE

Diagnosis: ECHO IS TURNED OFF

Solution: TURN ECHO ON BY CALLING DOECHO RTN IN CNCIN0 RTN,
AND THEN TURN IT BACK OFF BY CALLING NOECHO RTN IN CNCCON RTN
TOPS20 Change Order Number 4.2558

Written by: SIMMONS		29-Oct-79 10:42:37

Edit checked: NO			   Document: NO
  TCO Tested: NO		Maintenance Release: NO
Hardware-related: NO

Program: EDIT

Related TCO's:
Related SPR's: 	 13609


Problem: WHEN A "P" COMMAND WITH NO ARGUMENTS IS FOLLOWED BY ANOTHER
"P" COMMAND, ALSO WITH NO ARGUMENTS, THEN THE LAST LINE OF THE FIRST
"P" COMMAND IS PRINTED TWICE.

Diagnosis: "P" COMMAND DOES NOT RECALL THAT THE LAST LINE PRINTED
WAS THE CURRENT LINE.

Solution: ADD A VARIABLE "PNOARG" SO THAT IN THE OUTLIN RTN
IT IS SET TO -1, AND IN THE PRNT5 RTN SET "PNOARG: TO A +1.