home
***
CD-ROM
|
disk
|
FTP
|
other
***
search
/
SysOp's Arsenal
/
SysOp's Arsenal 1 (Arsenal Computer).ISO
/
tbbs
/
modfixes.doc
< prev
next >
Wrap
Text File
|
1995-02-05
|
21KB
|
545 lines
Field Mod status as of 05 FEB 95 (UPDATE.BIN):
Note: Problem repairs are presented by product category. The following
current release products have one or more field mods:
TBBS 2.3, ULEDIT 2.3, SYSOM 1.2, TDBS 1.2, QSO 1.0,
TIMS 1.2, and UltraChat 1.0
-------------------------------------------------------------------------
These TBBS 2.3 problems have been fixed via field mods:
-------------------------------------------------------------------------
MLTBBS MOD 1
------------
1. Tagging too many files in old style (non-full screen)
interface could cause a line of the system to crash.
2. When downloading a batch of files from the full screen
file system, the batch information is shown twice.
3. When doing <N>ew or <S>earch in the full screen file system,
the file owner information is lost.
4. When the user has a default protocol, and the menu item
limits protocol using /P:abc, the full screen file system
can put the hilite bar in the wrong place when the user
tries to change protocols.
5. After escaping out of a search for new files in the full
screen file system, it was sometimes possible to not have
the file screen redisplayed.
6. Batch Uploads to a pseudo-directory TYPE=47 using the /M
switch will say that uploads are already present if done
following a raw download.
7. On some systems the /M switch for multiple file uploads from
a TYPE=4 command will accept the files but never prompt for
the file descriptions.
8. When leaving the full screen file system's area display, a
user is left marked as not available for chat.
9. Internet addresses should be all lower case after the "@".
ULEDIT MOD 1
------------
1. If a user has selected more than 255 topic areas, a divide
overflow will occur in the <B>ackup function.
-------------------------------------------------------------------------
These SYSOM 1.2 problems have been fixed via field mods:
-------------------------------------------------------------------------
SYSOM 1.2 MOD 1
---------------
1. Pressing <F5> at the DOS emulator prompt line will crash
TBBS, often corrupting USERLOG.PTR
2. If you are in insert mode at the DOS emulator prompt, then
pressing F3 brings back remainder of line repeatedly instead
of only once as it should.
-------------------------------------------------------------------------
These QSO 1.0 problems have been fixed via field mods:
-------------------------------------------------------------------------
QSO 1.0 MOD 1
-------------
1. If QSO is called with a meaningless Opt Data field, it will
exit instantly with no error. This mod makes it ignore the
Opt Data field and go into normal QSO operation as though
there were no Opt Data at all.
2. If the name field entered by an off-line reader is upper/lower
case, QSO will indicate "not net status" and reject the reply
packet when it should accept it. The QWK standard specifies
that the names must be upper case, but apparently many readers
don't follow the standard in this area.
3. Topic scanning is slower than it should be in QSO.
4. The keyword file upload command does not allow ASCII protocols
to be selected if the protocol menu is presented. It should,
so that a key file may be typed online directly.
QSO 1.0 MOD 2
-------------
5. QSO does not always remove messages it reads from the user's
TBBS message waiting chain.
6. If a remote configuration command has certain types of syntax
errors it will crash QSO and TBBS.
7. If an uploaded message has text longer than 5.5k or in some
cases if it encloses a file, QSO will crash after displaying
garbage on the local console screen.
8. If an uploaded message or remote configuration command has a
message text > 14k, QSO can lock up a line or crash TBBS.
9. If services or bulletins are defined, QSO can lock up a line
on loss of carrier during scanning.
10. If there are more than 100 file areas in the .FAR file that
QSO uses for its new file scan, the title lines for areas
100 and up will display incorrectly in NEWFILES.DAT
QSO 1.0 MOD 3
-------------
11. If the .MSG file in a .REP packet is an exact multiple of
4096 bytes in size, QSO will process all messages (or parts
of messages) in the final 4k block twice. This can result in
either spurious "Invalid msg in packet" errors, or also in
messages being entered into TBBS multiple times.
12. QSO resets the logical counter on each message board causing
1STREADER to malfunction. This counter should only be reset
on the 1st message in the packet.
13. Invalid TYPE 201 strings in USERLOG.PTR can cause QSO to lock
up or worse. They should be discarded and the user treated
as a new user to recover.
QSO 1.0 MOD 4
-------------
14. If a caller drops carrier at exactly the right time in QSO
DOS file handles may be stranded. This can result in lost
clusters requiring CHKDSK /F to remove. On a system where
this happens repeatedly it may even result in spurious
"too many files open" error messages.
QSO 1.0 MOD 5
-------------
15. If the QSO temp directory disk becomes full, or a disk read
error occurs while QSO is adding either a bulletin or service
file to the QWK packet, QSO will crash TBBS.
16. QSO should not include *Encl distribution list enclosed file
control messages in QWK packets.
17. QSO should only send return receipts if they are TO: the
user, not receipts the user causes to be generated unless
the caller has net status (or is sysop) on this conference.
QSO 1.0 MOD 6
-------------
18. If an external conference is empty or mal-formed, QSO will
produce a corrupted scan display.
19. If the last message read pointer for an external conference
is set below the 1st message of the conference, QSO will
find no messages instead of finding all messages in the
conference as it should.
20. Certain error conditions can cause an external conference
file to be stranded open by QSO.
21. Trailing blanks on the QSO.CTL commands SYSID, WELCOME,
GOODBYE, and TEMPDIR can cause QSO problems.
QSO 1.0 MOD 7
-------------
22. QSO doesn't properly mark external conferences so MFSQZ can
tell them apart from TBBS internal conferences in the
201 USERLOG.PTR string.
QSO 1.0 MOD 8
-------------
23. When reading messages > 4k from external conferences, QSO can
put the incorrect text or "chop" text of themessage in the QWK
packet. Also, packing messages can take longer than it should.
24. QSO should mark any message tossed into an external conference
from an uploaded .REP packet as requiring echo so offline
scanners will pick it up. *Revised 6/13/94*
25. Selecting a conference past the highest configured will select
the last conference if it is internal. If the last conference
is external, QSO will generate many spurious error messages.
26. If the .FAR file that the NEWFILES command points to contains
a .DIR file that is in a directory that does not exist, QSO
will loop forever and crash TBBS.
27. If QSO packets being tossed with the /T command contain any
configuration commands, the configuration does not "stick".
Note: If QSO /T is run as a ghost task, all config will be
applied to the sysop record.
28. During certain operations, QSO "loads" TBBS more than it should.
QSO 1.0 MOD 9
-------------
29. Include files within restrict blocks were not properly handled
by QSO.
30. TBBS 2.3 does not require QSO to upper case the from, to and
subject fields of imported messages.
31. TBBS 2.3 allows using ~ as the first character of a line in a
.FAR file as a special form of comment for the full screen
file interface.
32. QSO was not setting the "local" bit on imported messages.
33. QSO now does case insensitive searches for name matches on
messages since TBBS 2.3 no longer stores everything as upper
case only.
-------------------------------------------------------------------------
These TDBS 1.2 problems have been fixed via field mods:
-------------------------------------------------------------------------
TDBSOM 1.2 MOD 1
----------------
1. RJUST() produces garbage if the argument is already fully
right justified.
2. If an intermediate results entry (IR) is left to be purged
at the end of the instruction, it is stranded instead. This
error makes some programs which run in TDBS 1.1 require
recompiling with a larger /GLOSS value to run under TDBS 1.2
because part of the glossary space is being wasted.
TDBSOM 1.2 MOD 2
----------------
3. The READ, CLEAR and CLEAR GETS commands erroneously generate
"Variable not found" errors when memvars currently in the GETPOOL
have been "hidden" by a PRIVATE command.
TDBSOM 1.2 MOD 3
----------------
4. When an indexed database which has multiple records with the
same key is in use by more than one user at the same time, if
a key is updated in the index other users may erroneously have
the file positioned to EOF on the next SKIP or GOTO.
5. The DIR command lists a "random" # of bytes remaining on disk.
6. CLOSE FORMAT doesn't "close" the format file. (It should work
like SET FORMAT TO with no file name specified).
TDBSOM 1.2 MOD 4
----------------
7. When an indexed database is updated in such a way that
the root index record is changed from one record to another,
a "[1147] Index file EOF hit (error during NDX processing)"
error may occur.
TDBSOM 1.2 MOD 5
----------------
8. TDBS erroneously ignored the SET UNIQUE ON setting which means
that if the "UNIQUE" keyword is not used with the "INDEX ON"
command, the new NDX file is ALWAYS created in "UNIQUE OFF" mode.
9. As is required by the xBase standard, TDBS defaults to "SET
ESCAPE ON" at program start. Unfortunately, this makes it
impossible for the programmer to prevent the user from ABORTING
out of a TDBS program by pressing ESC IMMEDIATELY, before the
first TDBS statement can be executed. This mod makes the following
changes to avoid this problem without losing xBase compatability:
o Until the first statement of the program has been executed,
"SET ESCAPE OFF" is the default "escape mode", preventing the
user from aborting the program via the ESC key.
o If the first statement of the program is a "SET ESCAPE ..."
command the "escape mode" is updated in the normal manner.
o If the first statement of the program is NOT a "SET ESCAPE"
command then "SET ESCAPE ON" becomes the default escape mode.
With this MOD installed, programmers can avoid the "unauthorized
abort" problem by making "SET ESCAPE OFF" the 1st program line.
TDBSOM 1.2 MOD 6
----------------
10. INKEY(0) ignores TBBS user inactivity and session timeout limits.
Note: This mod has been made separate in case anyone has come
to rely on this bug. It seems unlikely, but if so, you
can just back this mod out.
TDBSOM 1.2 MOD 7
----------------
11. FTIME() Returns invalid "seconds" field for hh:mm:32 - hh:mm:59
12. THE ENTIRE TBBS SYSTEM MAY HANG after an "[1151] Record is not in
index" error message is displayed if ALL of the following are true:
1. A record is updated by the simultaneous REPLACE of TWO or MORE
index fields of a DBF record and:
2. Both of the updated key fields are currently associated with
invalid or incomplete index records and:
3. One or more FLAT files were also open, with internal I/O
buffers assigned.
TDBSOM 1.2 MOD 8
----------------
13. Attempting to position a flat file 1-512 bytes past its physical EOF
prior to executing a FLREAD, may cause the program to LOOP FOREVER,
reading the last sector of the file instead of returning the
expected EOF status.
14. In the memo editor of TDBSOM the string "File already exits, overwrite
it? (Y/N) " should read: "File already exists, overwrite it?(Y/N) "
TDBSOM 1.2 MOD 9
----------------
15. If the updating of a field results in the deletion of 2 or more leaf
records in a .NDX file, the key fields of the remaining higher layers
of the index chain may erroneously retain the deleted key value instead
of being updated to the closest remaining key. This causes the index
file to subsequently "self destruct" with use getting keys out of order.
16. The "LOCATE / CONTINUE" sequence erroneously ABORTS on EOF, when a
CONTINUE begins on the last record of the DBF file (or at the EOF) and
SET DELETED ON or SET FILTER TO is active. (Should return EOF()=.T. and
FOUND()=.F. without abort).
TDBSOM 1.2 MOD 10
-----------------
17. An erroneous system fault may occur when a SINGLE USER is
sharing one or more index files across multiple work areas and
a data base update causes multiple index records to be altered.
18. When multiple users are updating or appending data to the same
shared file, the updated records may not always be shared with
all users of the file. This can lead to the corruption of the
data base, missed mailbox messages, etc.
19. The ** operator and EXP() function sometimes return erroneous
answers.
-------------------------------------------------------------------------
These TIMS 1.1 problems have been fixed via field mods:
-------------------------------------------------------------------------
TIMS 1.1 MOD 1
--------------
1. A NO-REQUESTS command in the TIMS.CTL can erroneously be
treated by TIMS as a NO-OUTBOUND command.
2. Using the TIMS Console to manipulate the outbound area will
sometimes change the AKA in use to be the node number of the last
node where mail was manipulated.
3. Monitoring a line while TIMS is receiving a file can sometimes
cause corrupted data in the received file.
4. Netmail SCAN will sometimes scan out packets with invalid headers
TIMS 1.1 MOD 2
--------------
5. NO-LOGONS did not work.
6. When attempting to not overwrite ARCmail TIMS would sometimes
loop forever trying to come up with another name.
7. The TIMS.CTL file HOLD statement did not default to the primary
zone when used without a zone number.
8. When using the TIMS control panel the error "Node is busy" would
appear if the proper outbound directory did not exist. After this
mod TIMS will create the proper outbound directory for the node.
9. During the one minute after midnight it was possible for TIMS to
halt the entire TBBS system.
TIMS 1.1 MOD 3
--------------
10. When using multiple DOTBBS commands it was possible to cause
TBBS to lock up.
11. The TIMS SCAN function when used with DELIVER statements would
properly route the mail to another node, but would also address
the individual message(s) to that node instead of to the actual
destination node.
12. The TIMS SCAN function would not properly put an ^AINTL line in
mail that was sent through a zonegate.
TIMS 1.1 MOD 4
--------------
13. It is possible for a TIMS in line-camping mode to sometimes lose
track of the fact it is line camping and hang the line
14. TIMS will occasionally report NO MAILER for a call that completes
successfully
15. TIMS does not put zone numbers in the packet header as it should
16. TIMS does not properly put the zonegate address in packet headers
when mail is gated
TIMS 1.1 MOD 5
--------------
17. The TIMS console does not look up addresses with points in
them properly.
18. The TIMS SCAN command will send messages that have been
marked deleted in the TBBS message base.
19. The TIMS SCAN command would either not properly send mail
through zone gates, or would not properly route mail with
the DELIVER statement, depending upon which TIMS mods were
installed and what the DELIVER command was.
20. Using the RESTRICT SESSION NOT-NODELISTED configuration
would sometimes cause properly nodelisted systems to be
treated as though they were not nodelisted.
TIMS 1.1 MOD 6
--------------
21. Using NO-OUTBOUND or NO-MAIL would sometimes still allow
calls to be made on a particular line.
TIMS 1.1 MOD 7
--------------
22. On Novell networks TIMS would truncate files after
sending them and then resend a 0 byte file which would
overwrite the original.
23. TIMS can sometimes get stuck trying to send a file for
forever if the other end sends back a Zmodem header that
TIMS does not understand.
24. TIMS would not properly create packed message headers for
zonegated messages.
TIMS 1.1 MOD 8
--------------
25. TIMS would often hog too much of the CPU when it could
have been making things better on other lines.
26. Using TIMS to toss echomail into the message base could
sometimes cause the message base to lose integrity.
27. TIMS will now set the bits that say that tossed messages
should not be scanned back out by FLAME as long as /H
is not used with the TOSS command. This is harmless if
FLAME is not being used as the mail processor.
-------------------------------------------------------------------------
These UltraChat 1.0 problems have been fixed via field mods:
-------------------------------------------------------------------------
UCHAT 1.0 MOD 1
---------------
1. Once a user on a line has been in "read-only" chat mode,
that line is never taken out of "read-only" mode, even
after the caller hangs up and another call comes in.
2. Users were sometimes placed in a very odd conference if
their first entry to chat was of the /E:4 type.
3. CALLERS#01_CMD did not respect a callers page length
setting.
4. Displaying %U3LINEBIO% would display a previous 3-line
bio if the user being shown did not have one.
5. The ANNOUNCE_CHARS_NUM length was being ignored in cases
where users typed the command and the announce string on
the same line.
6. Handles that were too long were not being truncated to
the proper length, and as a result, they would "bleed"
through to other lines.
7. Private channel passwords that were too long were not
truncated to the proper length and as a result, they
could cause other users to have passwords they did not
intend to use.
8. Entering KILL_LINE_KEY when changing a handle would crash
the entire system.
9. EXITBBS#xx_CMD calls did not properly save the clear
screen codes.
10. Ending times for links were not always being honored.
UCHAT 1.0 MOD 2
---------------
11. UltraChat was not aware of some TBBS 2.3 structures. As a
result it could generate spurious Software Check #4's when
a user logged on. This mod corrects the problem.