Amiga Elysian Archive/AmigaElysianArchive.iso/comm/c342e19.lha/update.e19

Text File | 1993-02-06 | 17KB | 352 lines

 -- 3.41.E1 Changes --

 1) Change the screen/window on startup to be front and active.

 2) Changed the outside editor routine to bring the citadel window
   and screen to front and be active.

 3) Changed the chat request to active window and move Citadel to front.

 4) Compiled and linked with D1(line numbers) to help track problems.

 5) Turned on stack checking since people might be trashing the stack.

 -- 3.41.E2 Changes --

 1) Changed the Read Forward to allow a parameter that is number of
    days as a substitute for date.

 2) Added a CallLog.sys error message when an external command cannot
    be executed.

 3) took out the stack checking, since it did not show anything and
    made the image larger....

 -- 3.41.E3 Changes --

 1) fixed a couple of minor bugs in the hope one would fix the other
    problems...  No such luck...

 2) fixed the error code in the Calllog.sys messages for bad external
    commands.

 -- 3.41.E4 Changes --

 1) Fixed the bug that trashed the netlog.  Citadel was writing error
    messages to whatever the netlog file unit was.  It was closed at
    the time.

 2) Added a window/screen to front and activate if the sysop uses ^T
    to grab the system.  Citadel will now pop to the front for him/her.

 3) added some error messages to the baud rate detection code.  It is
    not detecting the user speed correctly.

 -- 3.41.E5 Changes --

 1) added an extra check and some debug messages to the find_baud routine
    to determine why it is not networking.
 2) added the "B" for backup command, this is identical to the "U", for
    ungoto, but is the K2NE command.  I prefer this myself to UNGOTO which
    seems anti-intuitive(sorry Hue!).  Anyway, both are there and do the
    same thing.
 3) Rewrote the way the baud detection routine so that it will check for
    a net session alittle sooner than before.  The old way was skipping
    a net session on my system.  This fixed the networking problem.

 -- 3.41.E6

 1) Fixed the file size display for downloads.  The SAS/C was assuming an
    integer return on the lseek that was used for sizing.  Due to the IBM
    compatibility issues(ints are 16 bits), this cause a truncation and then
    a sign extention and trashed the file size displayed.  Fortunately, this
    was just a cosmetic bug since files are read to EOF when transfered.

 -- 3.41.E7

 1) In the process of messing with the Find_Baud routine I discovered a
    few cases where the start of a net session could be bypassed.  This
    would totally depend on whether or not your modem and Citadel connected
    properly.  At any rate, I changed the routine so that you needed 2
    crs in a row for a user to connect.  I have tested it and it seems to
    be a little more reliable than before.  This may be something I introduced
    in .E5, anyway, I think it is fixed now!

 -- 3.41.E8

 1) I have made some changes in the Find_Baud routine that will make the
    connection be a little quicker.  I also setup the routine so that if
    a user does not hit return, it will default to the user.  This seems
    to be alot better than before(.E7 and earlier).  It might connect a
    little differently and probably needs a little more testing before I
    release it on an unsuspecting public!
 2) I am also looking at the changes Hue put in his version.  Apparently
    there are some bug fixes I may wish to incorporate...

 -- 3.42.E0

 1) Nothing major here...  Hue updated to add a feature where certain
    aide room messages go into the "Net Aide" room.  Also fixed a bug
    between fast transfers and virtual rooms.

 -- 3.42.E1

 1) Added support for multiple copies of Citadel running at the same time
    with at least unique ARexx Port names that will not conflict.  The
    first name will be "Citadel_68K".  All subsequent names will have the
    CLI number appended to the end.
 2) Discovered a minor bug in the Auto Baud routine which I fixed.  It might
    just explain the occasional strange behavior.

 -- 3.42.E2

 1) Again I adjusted the AutoBaud to make it connect faster.  It seems to
    work better for user connects and doesn't affect net sessions.

 -- 3.42.E3

 1) Another timing adjustment in the AutoBaud.  Net sessions work ok, now
    I have to wait for a few callers...  Ah... a caller or two..  This
    Should trim a fair amount of time off the connection...
 2) LHA support added to the ".RAD"  since zip is not used much in the
    amiga world, I replace it with LHA.  This is crude, I have to do a
    better job, but will require a re-write of several routines.

 -- 3.42.E4

 1) Well, I think I finally fixed the "banging" return key syndrome Citadel
    was experiencing.  I added the capability to use the Escape key plus
    made the count of CRs global and not reset on each pass.
 2) Added Logedit to the list of utilities converted.
 3) Added a new file to the Audit Area: debug.sys.  All disk output from
    debug commands will go into that file instead of the callog.sys.
    I also added debug messages to record all external commands
    This will allow you to debug setup commands by turning on debug, shutting
    down, then bringing Citadel up again.
  4) Added some code to allow BPS=nnnn on the command line to either be
     the baud rate or character rate.  This means you can put 2400 or 240
     and it will be correct either way.  It should also report the correct
     baud rate of a citadel run as a "door" also.
  5) Add support for a new user door function.  This function
     will automatically execute a door type "newuser".  If you confg a
     door with this type, all new users will execute the door before
     connecting with Citadel.  The parameter is in the docs for #door.

 -- 3.42.E5

 1) Added some additional debug messages that should show what is being
    read for protocols and archivers.

 -- 3.42.E6

 1) Added some additional debug messages and the output of the current
    directory to the debug.sys file.  All input lines will be output
    to the debug.sys file as they are eaten.
 2) added MSGOUT, MSGADD, and VERIFY to the utilities archive.
 3) created the XPR library for Zmodem, a CTDLPROT.SYS, and a read.me file
    for P342.lha.  This will allow the use of XPR for Zmodem plus Kermit!
    in this example.

 -- 3.42.E7

 1) Changed the autobaud timers for faster logins.
 2) Fixed a bug in the NEWUSER door entry, now you only need it once.
 3) Added a new type of AutoDoor.  If you use the user name of Citadel
    in an AutoDoor, every user will execute the autodoor upon login.

 -- 3.42.E8

 1) Fixed a bug in the Autodoor interface.  It was ignoring the flags
    on an autodoor, a door that was modem only would still be run from
    the console.  Now flags are checked on all autodoors(including the
    new user door).
 2) Added code to open NIL: as the file handles for the Execute calls.
    This may be needed for 1.3 compatibility.
 3) added some code changes for stack checking.
 4) Adjusted the timings to the AutoBaud Routine.  Added a minor
    correction to the Check_CR routine, added a debug dump of the
    last eight characters seen to help detect possible problems.

 -- 3.42.E9

 1) Made two timing setups for autobaud detection, one for lock-ported
    which will be fast, and one for regular which will be slower.
    If you have a high speed modem, Citadel will respond much faster.
 2) added back the stack checking.
 3) replace some calls to CitSystem with direct calls to the JSystem
    routine.  They are supposed to be functionally equivalent except
    Citsystem formats.  This seems to have fixed my Zmodem bug.  I
    still have to figure out why!
 4) If a User drops carrier while being asked a Yes/No question, Citadel
    would get trapped in an infinite loop attempting to get an answer.
    I fixed it to assume "Yes", so that "abort account creation" type
    questions would be answered.
 5) Fixed an invalid User name which could allow a duplicate looking user
    to be created.  The duplicate has spaces at the on the end.

 3.42.E10

  1) I re-wrote the door interface code to attempt to fix a bug that
     does not allow room based doors.  There are several types of doors
     now, I am in the process of testing each type to see that they
     work correctly.  The types are:
       a) Auto Doors - These automatically execute when a user logins in
          and are tested.  The User Citadel can be used to do one for
          all users.
       b) New User   - This door will run for a new user before the
          account is created.  ***this one is not working yet, I broke
          it, now it fails the autodoor check I  will fix it soon...
       c) Regular    - This type will run with the D command, it may
          have restrictions on the door(Sysop only, Aide, Time limits,
          Console and Modem).  There also is the capability of having
          a door execute only from a certain room.
     All type should work now, even the room restricted types.  The
     re-write was needed because the newuser and other doors restricted
     to the modem only were being executed from the system console.  Also
     there was a lot of duplicate checks(all in error) that were not
     correct.  They all ignored the room or checked it incorrectly.
  2) Added some code to protect against a carrier lose in the input routine
     and the chat.  If a user should drop carrier while in a chat, it
     should exit the chat(or you can exit with escape).  This should fix
     the "Sleeping" message.

 3.42.E11

 1) fixed the New User autodoor.  Now all the doors work as documented.
 2) Reversed the order the baud rate are checked in the autobaud routine.
    This was changed back in .E6.  The difference is it starts from 0
    up to #SysBaud, not from #SysBaud down to zero.  This means it checks
    the high baud rate last.  The user will greatly speed up this type of
    connect by hitting escape.  If you have a modem that supports the
    locking of the baud rate, use the #Lock-Port parameter, it fixes most
    of the complaints about AutoBaud.
 3) Increased the timings of the AutoBaud checks, it will slow down the
    connect, but also give a little more time for the net recognition.
 4) added another check for dropped carrier to prevent a hung system.
 5) Fixed an order dependancy between the #CLOCK and #SERIAL_7WIRE, if you
    had #CLOCK after the #SERIAL_7WIRE, the CTS/RTS handshaking was not
    enabled.  This would cause garbage in for those with high speed modems.
 6) Changed the .ECD timing constant to be .1 milliseconds instead of
    milliseconds.  Also adjusted the ascii dump to not do a delay unless
    you had a delay specified.
 7) defined a symbol that caused a change in the conditional assembly
    of the message formatter.  This seems to have fixed several formatting
    bugs.  There was a recursive call in the old code that the symbol's
    definition got rid of!  the ".KI" is now correct and working!!!!
 8) Added the baud rates 38.4K, and 56.7K baud to Citadel plus tweaked
    the serial port code to be a little bit faster.  This was related to
    the problem discovered in #5 above.  We now use RAD_BOOGIE!

 3.42.E12
 1) Some additional re-structuring of the autobaud code, an adjustment
    of the timings.  Hopefully, this will connect better with all modems.
    I did run this for a day and had users connect without the #lock-port
    (runs a whole different set of code).  Both ways work fine here.


 2) I introduced a bug in the autobaud code, if the result.sys did not
    match, Citadel would set the baud rate to -1, this might have resulted
    in a garbaged connect.  I had one report of a system connecting only
    the first time, then it would not connect.  It may be that this was
    the cause.

 3.42.E13
 1) found two bugs I introduced, neither should be serious if you have
    the results.sys.  The one would cause Citadel to hang up if the user
    did not hit return twice or escape.  I accidentally deleted a line
    which I replaced.  This should work a bit better.

 3.42.E14

 1) Rewrote the routine for setting the baud rate to make sure it is
    always correctly setting the modem speed.
 2) Fixed an enforcer hit that caused a read of low memory.
 3) adjusted the sync timing on a net call.
 4) fixed a typo in the top baud rate.  57.6K not 56.7K.  also fixed
    a couple of occurances where the max baud rate was not setup
    correctly and was still limited to 19.2K.
 5) converted CTDL and CONFG over to the SAS/C version 6.1 compiler.  There
    seems to be a problem and optimization was removed to get it to run.

 3.42.E15

 1) Found a window closed twice error in the Citadel Dialer that was causing
    the dialer to crash if you connected and then exited the dialer.
 2) fixed the Citadel Screen title, was a compiler dependancy with the
    SAS/C 6.1 compiler and how I got rid of a warning.
 3) re-wrote the routine to size files.  Citadel used this alot and the
    lseek routine in the SAS/C compiler doesn't seem to be able to position
    back to the beginning of the file.
 4) re-structured the Fast Transfer menu, fixed a bug in the code, corrected
    the display routine.  Lha, Lzh, Zoo, and Arc are now available.
 5) Fixed a bug when a message was posted in the Aide/Net Aide room that
    caused it to change to another room incorrectly.
    There may still be some cases of this occuring, but the .AED is fixed
    «aide-edit directory» would end up in the base room.
 6) added some debug code to give an english language reason you are netting
    (i.e. calling a system).  Most of the time this is useless, but on
    occasion, I would have liked to know why.  This displays in the netlog
    when debug is turned on, or in debug.sys if you do not have +netlog on
    the command line.
 7) removed the opens of NIL: to see if this fixes the 1.3 external program
    interface.
 8) added some code to "jinx" the optimized in aideMessage.  It seemed to
    be losing the room id when an aide message was generated.  Hope this
    fixes it.

3.42.e16

 1) added code to flush characters when NULLs are recieved.  If a USER
    does a break, Citadel will see a character of '\0'.  When this occurs
    Citadel will eat all characters until the buffer is empty.  This
    should make recovery much quicker on a noise hit too..

 2) added a open of NIL: with a different call to execute to see if it
    fixes the 1.3 problems.  This was suggested by C= CATS's R. Jessup.

3.42.e17

 1) Working on the problem of a zmodem upload of a message not working.
    I think I have it fixed.
 2) Attempted to fix an enforcer hit in calllog processing.  Not sure of
    the call, but it may be just that a static variable was not initialized
    when a modem caller occured.  As it happened, the system was started and
    the caller called.  Usually the Sysop gets on and the variable gets
    initialized.  In this case the sysop never logged in from the console.
 3) made another attempt at the enforcer hit.  It still occurs, but is
    harmless other than the fact that it is a hidden bug.
 4) tried removing the peep hole optimizer to see if some problems are fixed.

3.42.E18

 1) reduced the stack requirements down to 20,000.  This is because I
    measure a stack usage of about 2K.  I am sure it probably uses more
    in some cases.  The bottom line is you can always increase the size
    of your stack, and this is the minimum.
 2) Added code so that when debug is on, the stack usage is tracked and
    the values recorded to debug.sys.
 3) fixed the enforcer hit.  I found a null pointer in the system independant
    code and informed Hue Jr.
 4) turned off optimization, added in the catch plus set debug to line
    and created a special image with version 3.42.D18.  This is to find
    the problems where people crash.

3.42.E19

 1) Fixed a bug in the stack size collection routine.  I also made it
    do the stack info collection all the time.
 2) added the stack usage to the I command in the Sysop menu.
 3) Fixed a bug in the initialization code that setup the Audit Area
    name.  If you did not have one defined, there were severa error
    messages that could try to use the audit area as repository of the
    information.  Now it goes to the home area by default.
 4) Put debug==line on, and include the info in the image.  This one
    is smaller, and is ok for regular usage.

Known Bug:  Deleting a Message(or moving it) can trash the setup of the
    base room.  This is probably a compiler dependancy and I will squash
    it real soon..