Trailing-Edge
-
PDP-10 Archives
-
tops20tools_v6_9-jan-86_dumper
-
tools/sed2/sed.shf
There are 16 other files named sed.shf in the archive. Click here to see a list.
Edit history for VAX version of SED beginning 4-Aug-1983
1 4-Aug-1983 If a file was being edited that would have the
attributes changed on exit (as a VFC format file
would), the attribute-change flag was not being
cleared when the <SET-FILE> was done to another
file. Module affected: SED1FI
2 4-Aug-1983 The test for the /AGAIN switch in the /RCUR
processor was backward. It said the two couldn't
be mixed if /AGAIN was off instead of if it was on.
Module affected: SED1SW
3 4-Aug-1983 If windowing was in effect, doing a <SET-FILE> with
the /RCUR qualifier displayed the file in the wrong
window. Module affected: SED1FI
4 11-Aug-1983 The display pointer, slide value and position were
being lost for the old window when <SET-FILE> was
done with an RCUR qualifier. Modules affected:
SED1DT, SED1FI
5 2-Sep-1983 If previous data was in the close buffer, successive
<CLOSE-LINE> commands followed by an <ENTER><PUT>
command would cause extra characters to be put into
the file. The number of extra characters would be
equal to the number of succesive <CLOSE-LINE>
commands that were given. The count of characters
closed was incremented one time too many for each
<CLOSE-LINE> command done. Remove the extra increment
at CLSLN4: + 4 lines. Modules affected: SED1CM
6 2-Sep-1983 Certain files containing a large number of very short
lines would cause an access violation when the file
was being saved. This was caused by a design
deficiency in the TRAILL routine. TRAILL has to
allocate space at the beginning of the file, so it
will have room to create the RMS record lengths, so
it estimates the amount of space to allocate by
dividing the total size of the file by 10. In certain
cases this was too small, and the destination pointer
would overrun the source pointer. The cure: at the
end of each line, check to see that the overrun is not
imminent. If so, allocate some more null space by
using the same algorithm. Module affected: SED1SU
7 6-Sep-1983 Change the way form-feeds are output to the file.
Previously form-feeds were output as they are by
TECO: each one is in its own record. Change it to
the way EDT does it, with the form-feed imbedded
in the record, instead of in a separate record. The
compilers don't seem to mind it this way, and it is
more straightforward. Module affected: SED1SU
8 29-Sep-1983 Make <INSERT-MODE> print message when windowing is
in effect. Also make /DTABS display highlighted
"T" followed by dots, instead of just dots. (These
changes courtesy of Michael Iglesias at UCI). Module
affected: SED1DS
9 29-Sep-1983 Make /NOMESSAGE suppress the 'Parm defined by cursor
movement' message. (Also from Michael Iglesias, UCI).
Module affected: SED1SU
10 29-Sep-1983 Implement the /NOFENCE qualifier. Modules affected:
SEDVAX, SEDITB, SED1SW, SED1DT, SED1DS
11 20-Jan-1984 If journal recovery is attempted, /SHOW is not in
effect, and the journal file does not exist, SED gets
an access violation in SED1DS. Module affected: SED1JO
12 20-Jan-1984 Add capability to abort journal recovery. Module
affected: SED1JO
13 20-Jan-1984 If the journal file is not found, the nominals for
roll-forward-lines etc. are not set up. Module
affected: SEDVAX
14 28-Feb-1984 The comment created for basic files when a new file
is being created caused errors when the file was
compiled by basic. Add a "1" at the beginning of
the string to correct that problem. This will,
however, cause further problems if the edit-history
qualifier is enabled, since all comments inserted
by SED will have the same label.
15 13-Mar-1984 When executing with /SHOW set, if the execute buffer
contains a <SUBSTITUTE> command, the two "enter"
delimiters are output to the screen, though they
shouldn't be. Module affected: SED1MV
16 16-Mar-1984 If the file has been modified, the sequence <ENTER>
<CURSOR-DOWN><ABORT> followed by a non-abort character,
to cancel the abort, followed by <PICK> will cause
any subsequent cursor movement command to abort with
an access violation. The <ABORT> routine was setting
the FLG bit and not subsequently clearing it, so that
the cursor movement routines thought they should finish
by doing an RSB instead of JMP, causing a stack
violation. Module affected: SEC1MV
17 2-Apr-1984 Implement a character on the screen that indicates
data extends beyond the end of the screen for the
current line. This is currently only implemented
for VT100s and is the diamond character, the same as
is implemented in EDT. Modules affected: SEDITB,
SEDBYU, SED1DS.
18 10-Apr-1984 Implement the /DELIM qualifier and enhance word-wise
tabs to stop at the next character after any user-
supplied delimiter. Also implement the /ID=a,b
qualifier to allow specification of the comment
characters before and after the comment line.
Modules affected: SED1MV, SED1DT, SEDITB, SED1SW,
SED1FI.
19 12-Apr-1984 Implement the PUSH command to spawn a subprocess and
(optionally) execute a specified DCL command. Modules
affected: SED1MV, SED1DT, SEDITB, SEDBYU.
20 12-Apr-1984 Remove the comment characters from the /ID switch and
create a /COMMENT switch to specify the comment
characters. That way the comment characters can be
defined without having to have the ID lines displayed.
Modules affected: SED1SW, SED1DT, SED1FI.
22 25-Apr-1984 Don't save PUSH commands in journal file; save HELP
command only if ENTER-PARAMETER has been done. Module
affected: SED1JO.
23 16-May-1984 Fix some minor problems:
A. A <PICK> followed by creation of a new file and
a <PUT> into that file sometimes caused extra
characters to be put into the file. Module
affected: SED1SU (MAKCHR routine).
B. Enable global passall mode unless the logical
SED$DEBUG is set to YES. Module affected: SEDVAX.
24 21-May-1984 Add .DTR to the file-type table for datatrieve files.
Also, undo all special terminal setups in effect when
the <PUSH> command is issued, then do them again when
it returns. This allows ^Y, ^T, xon & xoff to work
in the subprocess.
25 5-Jul-1984 Add code to support terminals having insert-mode. Also
add code to support VT102, which has insert-mode.
Modules affected: SED1MV, SEDVAX, SED1CM.
26 6-Jul-1984 Set the terminal width in the terminal table to match
the width given us by the system. (This is currently
only being done for VT100-style terminals). Module
affected: SEDBYU.
27 16-Jul-1984 Accept the system values for screen width and length
instead of using the built-in values. Module affected:
SEDBYU.
28 17-Jul-1984 Add code to support SED on a Macintosh running
Macterminal without a numeric keypad. The terminal
responds to VT100 command sequences but needs to use
the "generic" layout of escape functions in place of
the keypad keys and function keys.