Google
 

Trailing-Edge - PDP-10 Archives - steco_19840320_1er_E35 - 10,5676/teco/source/decbug.ans
There is 1 other file named decbug.ans in the archive. Click here to see a list.
TECO 200A(1144) - This copy is mostly bug fixes.  The only new
things are the /YANK switch on ER/EB (I think I put this in after you got
the last copy), and the ^W/^U processing in video mode crossing line boundaries
(as suggested below).  It also has the correct fix for 100+%(q-reg).
We are also looking for comments on changing how space works: What do
you (everyone up there) think about space no longer being a plus operator,
instead having it just be a formatting character that is ignored like
CR and LF currently are?
Also, please see TOPS10.BUG for a couple problems that TECO has brought
out.  The first one is the only one that can be reliably reproduced at the
moment.

[Fixed]	1. Stu Grossman	29-Dec
	The command {FOO}U* [*]g gets an ill mem ref...

[Fixed]	2. Stu Grossman 29-Dec
	TECO does not set up my screen correctly. The .INI file contains only
	:ER/MODE:DUMP1,20e$(.)21,23e$(command-buffer)evon
	When I read in a screenful of data, the bottom line does not show up.
	However, if I type 1,20e$(.), the bottom line reappears.
	(See problem 6)

[Fixed plus user error: In doing the :ER/MODE:DUMP$, you are closing the
current input file (TECO had a bug in doing this).  If you wish to default
to /MODE:DUMP, you need to do an E. into a random Q-reg first (see TECO.TCO).]
	3. Stu Grossman 29-Dec
	This is a real goody, watch closely.
	1. Dismount the structure that your TECO.INI is on so that TECO can't
	find it. Ex: DISMOUNT DSKB:.
	2. Use the TECO monitor command to edit a file. Ex: TE FOO.BAR. TECO
	should just print a star and not go into video mode.
	3. Now use the EI command to read in and execute your TECO.INI.
	Ex: EIDSKB:TECO.INI$$.
	4. Don't even bother to edit the file.
	5. Use the EX command to exit.

	An ill mem ref should occurr after step 5.

	A variation on this theme is to read the .INI file with an ER command
	then to go HXAMA to execute it.

[Fixed]	4. Stu Grossman 29-Dec
	1. R TECO
	2. EVOFF$$
	3. EVON$$
	4. ?Ill mem ref (or some other flavor of malfunction)

	The TECO.INI file contains only the line in one of the prevoius
	complaints.

[Fixed]	5. Stu Grossman 30-Dec
	Ill mem ref from E? command. Make a TECO.INI that contains only the
	command EC. Then run teco, and type E?Q$I$$.

[Fixed]	6. Stu Grossman 30-Dec
	The expression :ER/MODE:DUMP$1,20E$(.) gets evaluated as 1,19E$(.)
	because the expression :ER/MODE:DUMP$ has a value of -1. I don't
	recall TECO ever doing anything with the form <expr>n,m.

[Confusion about what ^Z is supposed to do.  The ^Z command exits to
monitor, closing all files, WITHOUT writing out any data from the buffer(s).]
7.
30-Dec-81  9:00:02,725;000000000001
Date: 30-Dec-81  9:00:02
From: TW <Wachs at KL1026>
To: Grossman at KL1026
Reply-to: Wachs at KL1026
DTN: 8-231-6455
Subject: STECO bug
Message-ID: <"MS5(2045)+GLXLIB1(1067)" 11788189701.23.8.44520 at KL1026>

If I say TECO FOO.BAR[10,675]/INPLACE, later decide I want to
exit without doing anything and type ^Z, the file becomes 0 blocks
long in [10,675]!
The file doesn't even have to be in a different area - ^Z to
a file being edited /INPLACE nakes the file 0 blocks long.

	It took me 42 minutes to recover the file that STECO deleted
on me (JMF would have been PISSED if the file had been completely
wiped out so I did some magic which an ordinary mortal would not
have been able to do)
   --------

[Haven't yet figured out why you want to load TECONC with DDT.  Field
image VMDDT works quite well with it.  Also, in order to be able to do
this nicely, it would require LINK support in the form of a /SEGMENT:PSECT:xxx
switch, so that DDT could be loaded where we wanted it, not into .LOW.]
8. Stu Grossman 31-Dec
	Make TECONC normally load with DDT.REL, but teach it to get rid of DDT
	if the user says he doesn't want a debugging version.
	Addendum: teach it to keep DDT if it was loaded via the /TEST:DDT
	switch.

[I would assume that all that is needed to fix this is a "J" command in
the macro which is setting up Q-reg STATUS.]
9. Tony Wachs  31-Dec
	If Q-reg STATUS gets larger than 80 characters (for VT52) the
	first 80 chars are not displayed. I'd prefer to see the chars past
	the 80th not displayed instead
	Addendum: This may not be a bug, just a misinterpretation.../Stu

[Fixed]	10. Tony Wachs 31-Dec
	If I say .TECO FOO.BAR, mess up the edit and say ER$$, TECO
	has forgotten (never knew?) the filespec of the input file.

[Couldn't make this problem occur in any copy of TECO I have.  If you
still have this problem with this copy please give me an exact example.]
11. Tarl Neustaedter 31-Dec-81
	S^E[a,b]$ doesn't work. This works in standard teco and texas teco.

[Deferred for the moment.  We have an idea for doing this but have not
worked out all the implications as yet.]
12. Tarl Neustaedter 31-Dec-81
	I need a way to UN-define an immediate command - Specifically, I
	don't want ^U or ^W or space to have an immediate meaning. I want
	them to behave as they would normally in a command string.

[Implemented.  ^U will now delete the previous line if it is typed when
the previous character is a line terminator.  ^W will now delete line
terminators (it treats them like any non-space punctuation character, and
treats CRLF like a single character).]
13. Tarl Neustaedter 31-Dec-81
	Suggestion: If ^U is typed while at the beginning of a line (of a
	multiple line command), have it delete the previous line. This would
	allow backtracking several lines with only one key. Also, allow ^W to
	cross line boundaries.

[Eventually. It is a bit of a pain to write the .EXE files with the core image
the way TECO needs it. TECO does not write out an exact core image,
but instead avoids writing some of the data that would be nonsense when
the EE'ed file is run.]
14. Tarl Neustaedter 31-Dec-81
	EE produces a .SAV file. Since under 701 and 702 you can no longer
	create these by SAVE commands, I'd like the code updated to .EXE files.

[This is on the list.  The command format will be E?(q-reg)DISPLAY-BOUNDS$
and will return the positions of the first and last characters displayed,
once I figure out a reliable method of figuring out the first and last
characters displayed if a user stops the screen update with type-ahead.]
15. Tarl Neustaedter 31-Dec-81
	There is no way to find out what part of a q-register is on the
	screen - If "." is not in the window, the E? returns 0.