Trailing-Edge
-
PDP-10 Archives
-
BB-L014E-BM
-
autopatch/apl2s2.d03
There is 1 other file named apl2s2.d03 in the archive. Click here to see a list.
EDIT DESCRIPTIONS FOR APLSF-20-V2
EDIT 520 FOR APLSF
[SYMPTOM]
Character editing on screens (VT52, VT100, etc.) doesn't
work the same way as on hard copy terminals (ASR33, LA36).
[DIAGNOSIS]
APL types a line, and then tries to re-position the cursor.
On APL terminals, it backspaces to where it wants to be, no
problem. On TTY's, it send CR (no LF) and spaces to where
it wants to be. This erases characters at the beginning of
the line on a screen.
[CURE]
For TTY's, instead of sending spaces, send the characters
that are already there, printing them again. This works
fine on VT's, and might look a bit strange on hard copy
tty's.
********************************************************************************
EDIT 521 FOR APLSF
[SYMPTOM]
APLSF sets the hard formfeed bit for key paired terminals.
[DIAGNOSIS]
APLSF seems to be making an assumption that all APL
terminals have hard formfeed.
[CURE]
APLSF should leave the formfeed bit alone.
********************************************************************************
EDIT 522 FOR APLSF
[SYMPTOM]
Doing )OFF or )OFF HOLD from an )INPUT file causes a TOPS-10
device error.
[DIAGNOSIS]
When the user signs off, bill is called, and IBUFF is used
as a buffer for the accounting file. If the )OFF comes from
an )INPUT file, IBUFF may have the buffer headers for the
)INPUT file. If the buffer headers get clobbered on the
-10, you get a device error.
[CURE]
Use INBUF for the accounting file buffer.
********************************************************************************
EDIT 523 FOR APLSF
[SYMPTOM]
.BXRENAME does not work if protections or PPN's are used as
part of the file specification. The creation and access
dates, and the data mode could be altered.
[DIAGNOSIS]
The protection and PPN fields were mixed up. The second and
third words of the LOOPKUP/RENAME argument block have the
file extension and protection. These words also contain
information about the creation and access dates, and the
data mode. APLSF was not preserving this information.
[CURE]
Fix protection and PPN fields. Make sure the dates and data
mode are saved.
********************************************************************************
EDIT 524 FOR APLSF
[SYMPTOM]
This example fails with a DOMAIN ERROR. Although the
assignment doesn't really do anything, it should not be an
error.
A_2 .RO 0
A[.IO 0]_''
[DIAGNOSIS]
The code that does type and length checking is wrong. It
gets the right and left sides mixed up.
[CURE]
Fix the code. Also, move edit 453 to see if the left
argument exists before doing any further checking on it.
********************************************************************************
EDIT 525 FOR APLSF
[SYMPTOM]
)DROP with wildcards doesn't work on TOPS-10.
[DIAGNOSIS]
The LOOKUP block is not being set up properly.
[CURE]
Make sure the LOOKUP block is correct. Also make it work if
the default path is an SFD.
********************************************************************************
EDIT 526 FOR APLSF
[SYMPTOM]
Compression of a zero by n matrix returns a length error
when the left argument is a vector.
[DIAGNOSIS]
APL was trying to use a shortcut algorithm when one or both
of the arguments had zero length. The shortcut algorithm
doesn't work properly for many cases.
[CURE]
Remove the shortcut algorithm. The normal code always does
it right.
********************************************************************************
EDIT 527 FOR APLSF
[SYMPTOM]
Assignment of an empty array to a subscripted empty array
fails with a LENGTH ERROR.
[DIAGNOSIS]
The length of the subscripted argument was being computed
incorrectly. It was getting 1,,0 instead of 0 for a zero
length argument.
[CURE]
When computing the length of a subscripted argument, don't
allow the value to overflow into the left half of the word.
Nothing can be that big anyway.
********************************************************************************
EDIT 530 FOR APLSF
[SYMPTOM]
When ending a function definition with the workspace almost
full, the line number becomes minus one, and all you can do
is use ^C to exit apl. The active workspace is lost.
[DIAGNOSIS]
At the end of function definition, APL garbages all the
lines in the function. Then it reclaims the lines still in
the function. The old lines go away when the garbage
collector is called. APL may call the garbage collector in
the middle of reclaiming the good lines. When it does, it
sets the line number to minus one as a warning.
[CURE]
Make sure there is enough memory available before garbaging
any of the function lines. Comment out the minus one line
number code. If there is not enough memory, return a
WS FULL error, and let the user expand core or delete some
of the function lines.
********************************************************************************
EDIT 004 FOR APLTAP
[SYMPTOM]
Assorted magtape errors from APLTAP.
[DIAGNOSIS]
APLTAP is not checking for end of file properly. It can
easily try to read past the end of the tape.
[CURE]
Make sure APLTAP checks for end of file correctly.
********************************************************************************
EDIT 005 FOR APLTAP
[SYMPTOM]
These characters are not being translated properly: $ .DM
.LB .RK .RB DEL
[DIAGNOSIS]
The translate table OURTAB is wrong.
[CURE]
Fix the table.
********************************************************************************
END OF APLSF-20-V2