Terminius Beta read.me file --------------------------- Please read all the information below, some of it supercedes the info that is present in the first draft of the user manual. === 1.910 ================================================================== More bugfixes in this one and some last minute changes to round out the script language. The program as it exists here should be considered "frozen", no more suggestions/additions will be considered unless they are an absolute must for the 2.0 release. From here on out, strictly bugfixes from now on. Please try to start from scratch with this release and try to exercise as much of the program as you can. I'm going to concentrate on finding holes in the low memory conditions that can occur with 1Meg systems. This thing can suck up a lot of memory if every bell & whistle is enabled as well as opening all the requesters, so I have to make sure that it doesn't exit unless a definate dead-end is reached. The one pending change is to add clipboard support to the main terminal display, this needs to be done since the posting of lines to the review buffer can get weirded out if the buffer is opened and closed several times while ANSI encoded text is being received. What will change is that the review buffer will only have text that has scrolled off of the display. I also need to flesh out some more example scripts, would appreciate some additional help in this area too. Please try to document common tweaks that you run into with newly created login scripts, need this to add to the user manual, which also needs to be worked on. Anyways, changed the phonebook requester, removed the DIAL gadget and changed the LOAD/SAVE gadgets for consistency with the other requesters that load and save data files. Now uses the same gadget pair as the console requester. The DIAL gadget became redundant due to OK now starting the dialer if there are any entries selected, so it was removed. Added a public screen option gadget to the console requester so that you can now open or close them without having to exit all sessions and restart the program. Added a gadget in the serial port requester to ignore the OwnDevUnit library too. Removed the two CLI command options that controlled these functions, no longer needed now. The PRINT, INPUT and ASK script commands have been changed. The REMOTE option is no longer valid for each command as they now check for DCD to determine if the output prompt or data needs to be sent. Input to the ASK and INPUT commands can now come from the remote user or local keyboard to satisfy the command. In order to allow local-only output with the PRINT command, use the keyword LOCAL as the first item in the command statement. Also added global integer and string variables. The single letter variables, A-Z and A$-Z$ (case is not significant) are reserved as global to the program now. Once assigned, they retain their value during the entire session. To clear them, use the CLEAR script command (which also clears all local script variables as well). Global string variables can also be sent from function key macros. the escape sequence for this is !=(A-Z). For example, to send the contents of the variable H$ you would set a function key macro to: "!=H". The remainder of the macro will not be executed once this escape is processed. Added some new script commands, their syntaxes are: GETINT prompt, var [,min, max] Opens an integer input requester with the current value of (which can be created as part of calling this function). The optional and values declare the minimum and maximum entries that the requester will allow the user to enter. If not supplied, any entered value is valid. The OK gadget (or a in the string gadget) must be selected in order to update the value of . GETSTRING prompt, svar Opens a string input requester with the current string value for (which can be created as part of calling this function). The OK gadget (or a in the string gadget) must be selected in order to update the value of . TRACE = boolean When TRUE or ON is used this variable will enable script tracing. An inverse colored string formated as " @: NNN " will be printed to the screen where "NNN" is the linenumber of the script line that was executed. DATE$ Returns a string of the current system date formatted as per the setting of DATEFMT. TIME$ Returns a string of the current system time formatted as per the setting of HR24MODE. === 1.909 ================================================================== Ok, a lot of tweaks to file transfers, please check the whole works to make sure there's nothing left that I overlooked. One thing, those of you that were complaining of problems with where files were being written to during a download, make sure that the extended pathnames option is not selected. Also, note well that some menu items have changed, please review them and see if there are other, more logical changes that I should do to them. Lastly, the function key macros are now working as they did in JR-Comm. The character escapes are as follows: ~ Pause one half second. ^^ The '^' character itself. ^~ The '~' character itself. ^! The '!' character itself. ^char (A-Z) Send char as a control character. !%decimal Character pacing, in decimal tenths of seconds. !@filename Load a new macros file. !*filename Load a new phonebook file. !&filename Launch Terminus script. !*filename Launch ARexx macro. !$hex Send a hex character. !#decimal Chain to macro where decimal equals: 0-9 Function keys 1-10 alone. 10-19 " " " w/CTRL 20-29 " " " w/SHIFT 30-39 " " " w/ALT The pacing option is only active for the macro that it is declared in, it is reset whenever another macro is started or when you chain to a new macro. Launching a Terminus script will flush any currently running scripts first. All four macros that load or launch halt execution of the macro after running. No recursion is checked for when chaining to another macro, be careful! === 1.908 ================================================================== Quick post to remove that blasted requester, the flu left my brain kinda soft, sorry about overlooking that one... Anyways, also changed the review buffer in this one. It now has a scroll gadget in the right border. I still need to add arrow gadgets there for finer control, wanted to get this out first though. The review buffer now uses a better storage method for the lines in the buffer that results in faster stepping throught the buffer as well as being able to know how many lines are in there, which is what is needed to have a scroll gadget. It added four extra bytes of overhead per line, so there is not too much loss in storage. I've also fixed the problem with keyboard support, but there are a few things you have to remember. Anything typed while the review buffer window is active will be sent out of the serial port, so you get the best of what the integrated review buffer had, the only thing is that none of the cursor keys are sent, nor is or the / qualified sequences as they control the review buffer. You would have to close the review buffer or activate the main window if you need to send those keys. The cursor keys still control the review buffer as before, still closes the requester and starts a clip. The change was how you end a clip, which is what does now. This was done so that you could still send a out the serial port. I've also added the numeric keypad sequences back in as they were described in the user manual. This should give everyone enough of a choice to control the buffer now. Please hammer the review buffer and cliping operations, I want to make sure that all this surgery was successful. There's still a handful of bugs I need to zap, but I wanted to get this one out quick. I should have another one posted by Saturday evening that should be pretty much it. === 1.907 ================================================================== After a week-long bout with the flu I've tried to get caught up with some bugs that were reported. Killed quite a few of them, but there are a few left that still need to be located. It's starting to act pretty solid here, at least I hope so... ;-) Really worked on the file transfer code again, especially ZMODEM step down and XPR operations. The REMOTE function needs some work yet, but it should be working for the most part now. The major problem I found deals with the way a PD directory lister, "LS version 4.1ljr" operates. It makes the foolish assumption of the console.device being local, so it times out before the remote console has had a chance to respond with the window bounds report. I may have to build a parser to monitor output and fake console reports locally, not sure yet. Another thing that I'm going to do is redo the review buffer to add a scroll bar and get rid of the gadgets along the bottom of the window, too ugly the way it is now. Lastly, will change the function key macros to the old JR-Comm method, will add some escapes for launching Terminus and ARexx scripts. === 1.906 ================================================================== Ok, not too many fixes in this one, had to leave for a wedding this weekend. XPR is in there now, so please try it with any modules that you may have around. I've done some testing with the ZMODEM, Quick B and B+ modules, but not too much. I also don't have access to a KERMIT server, so if someone could help out there I'd appreciate it. === 1.905 ================================================================== Changed some logic for the default script pathname in preparation for use with the Installer utility. It will now look for "TERMINUS:Scripts" and assign it if so, otherwise "TERMINUS:" will be used. I'm still at a loss to explain the gurus that were reported, but this one is a complete recompile, I hope it is no longer around. === 1.904 ================================================================== Primarily added/changed things operationally as listed below, there are some bugfixes as well for this release, of course!. It should be rather solid for the most part now. We'll see though... ;-) Please edit your phonebook file(s) as follows: Change all PENUMBER entries by adding 1 to the current value, PENUMBER0$ becomes PENUMBER1$, and so on. This was changed to clean up some logic in the dialer. Please edit your console definition file(s) as follows: The PALETTE command now only has four entries, WORKBENCH, COLOR, MONO and CUSTOM. Please change all CUSTOM entries by deleting the number attached to it; "CUSTOM 1" becomes "CUSTOM". This is needed since there is logic internal that determines the correct palette values to choose from. Also remove any ANSIMOUSE commands, it was removed and changed to MOUSEPACE as per the changes outlined below. ZMODEM is now present and should be working correctly. Please exercise the logic that steps down from ZMODEM to YMODEM to XMODEM by choosing ZMODEM on your end but using the other two protocols on the remote side. Sadly, most Amiga BBS' will not work the other way around, but PeeCee systems using DSZ (almost all do) will probably work correctly. I still have to add support for XPR though. Probably will add a seperate requester so that an activation string can be entered. I still have to investigate the protocol a bit more before finalizing it. I also could use a lot of script testing, ARexx and internal if possible. I'm going to try to put together a simple host script that uses the remote CLI (which also needs to be checked). One last thing, one report of the insert line ANSI sequence under VT-220 emulation was posted but I can't duplicate it. Hope a few of you can stress that emulation and capture the offending sequence (with RAW mode) and send it up to me. Still haven't heard any suggestions from anyone on alternative menu key bindings. I am considering dropping some of the less used items prior to release so that I can get closer to the JR-Comm layout, I'd like some suggestions on this though. PHONEBOOK --------- Changed key operation, now selects/deselects an entry while will select an entry and then start the dialer immediately, it does not deselect an entry if it was selected. Use to select an entry for auto-recording as before. DIALER ------ A one second delay is now inserted prior to sending the next modem command. This should eliminate the problem where some modems were not able to respond to another command being sent so soon after the last modem response was received. The dialer now sounds an alarm tone when a connection is established. If a script has not been defined, or auto-record is enabled, the alarm will sound for several seconds, you can bypass it by pressing any key or the EXIT gadget. ANSI MOUSE ---------- It is now always enabled, but you have to use to do an ANSI mouse operation, this should make it coexist peacefully now. The gadget that used to activate the mouse feature has been changed to toggle the way the ANSI sequences are sent. You can now have them blasted out the serial port in one shot or paced as it was before. REVIEW BUFFER ------------- Major change here. It now uses a seperate window for showing the buffer contents. The increase/decrease menu items were removed since the window can be sized (only vertically, not horizontally). Tried using a scroll gadget to move around the buffer, but due to using a packed circular buffer, it refused to work properly. There are six gadgets along the bottom of the review window, but since the old JR-Comm method of using the cursor keys is also used again (had to for the benefit of the A600), I might pull them so that the window can show more lines. - Close window. / - Move one line. - Move one page. - Top/bottom of buffer. The reason for the change to a seperate window was to eliminate the hidden file transfer stats mode, butt-ugly it was... PUBLIC SCREENS -------------- Well, there was another reason too... Added a command line keyword PUBLIC that will open all custom screens to the same size as the workbench screen as 2.0 public screens. The terminal display will be windowed as if it was on the workbench screen. Screen changes are ignored if vistor windows are open on the screen, a requester will be posted informing you of same as well. You can also open the terminal display on another public screen by selecting WORKBENCH and making the public screen you want to open on the active screen. TEXT SCROLLING -------------- Rewrote most of the scroll code for two reasons, speed and support for interleaved bitmap support for a future release of AmigaDOS. This will help loads for public screens when it becomes available. Scroll rates should be much faster now overall, especially on non-public custom screens. ASCII SEND ---------- Added support for character pacing. This mode will pace each character by waiting for a character to be received prior to sending the next one. FILE TRANSFERS -------------- YMODEM and ZMODEM now supports extended pathnames. When enabled, the entire pathname will be sent and received. Subdirectories will be created as needed when received. The download path is not used when extended pathnames are in effect, so files will be written and read from the current directory. The UPLOAD script command supports AmigaDOS wildcards and also has a recursive file collection mode that allows you to send entire directories when extended pathnames are being used. The stats requester now shows the protocol in effect, and is updated accordingly if a "step down", ZMODEM to YMODEM, and so on, occurs. When a file transfer completes an alarm tone sounds for several seconds, pressing a key or the ABORT gadget is selected. If the transfer is aborted the alarm is not sounded. Added flag for ZMODEM, SKIPNOTLOCAL, that is used in conjunction with the ZMODE setting. When this flag is active it will skip over any files that are not present on the receivers system. AREXX ----- Added menu item to kill all active macros. It cannot selectively kill macros yet. FUNCTION KEY MACROS ------------------- Added support for the CHAIN script command from within a function key macro to start a script. Any stopped/executing scripts will be cancelled prior to CHAINing. SERIAL REQUESTER ---------------- Changed data len and stop bits integer gadgets to cycle gadgets. ABOUT REQUESTER --------------- Changed INFO menu item to ABOUT. Now opens a requester instead of blasting the info to the terminal display. ONCARRIER --------- The ONCARRIER command is now triggered when a change of DCD state occurs, not just from high to low as before. INTEGER GADGETS --------------- Selecting an integer gadget via its keyboard equivalent now opens a string requester with the name of the gadget in the title, the minimum and maximum allowable values and the string activated for numeric entry. A bell will sound and the gadget will reactivate if the value entered is outside the allowed range. All program multitasking is disabled while waiting for entry. SCRIPT CONTROL -------------- You can now stop a script from the menu when a script is in INPUT mode, no other menu items are able to be used during this time though. 2.0 SCREEN DRAGGING ------------------- Got rid of the two scan line area when using custom screens under 2.0, it wasn't needed due to the ability to drag a screen with . It's still there for 1.3 though. === 1.903 ================================================================== Small fix for TERMINUS: pathname problem. === 1.902 ================================================================== This release adds XMODEM and YMODEM file transfers, some changes and a few other additions. All bugs reported to the beta conference have been fixed (hopefully...) and are listed in the bugfixes.txt file. While you're testing this release I'll be adding ZMODEM, which should only take a week since there's not as much work to do as there was with the initial XMODEM and YMODEM which required the file and serial I/O to be debugged along with the protocol code. I'll then work on the XPR stuff with the release following the ZMODEM release. The final beta will be the one with the registration/installation utility which will allow easier distribution of updates than the current method of having to call the support BBS. ---- Almost everyone made a note about the border around the display. Just to be sure that you are all up to speed, I'll talk about this again here. There's a "problem" with the way the 2.0 release of AmigaDOS looks at colors and how a terminal does. Basically, 2.0 wants a lightly colored color 0 for the background which is used in requesters and in the main screen. This along with a few other colors gives a very good 3-D look to gadgets, window borders and other things. Unfortunately, most all terminals use a black background and colored text. This is a definate collison with the two when used together. What's nice is that the ECS chipset provides for "border blanking" which tells the Denise chip when to turn on and off video output to provide for black borders around the screen. When used with Terminus this gives you the 2.0 look to the requesters and gadgets and makes for a really nice display. But, when you use it with a non-ECS Amiga you get a black rectangle which is the screen bordered by the light color 0 background, which really makes it look ugly. The fix? The only thing you can do is make color 0 black, which is what the CUSTOM 1, 2 and 4 palettes have, colors 0 and 1 are swapped in those so that you get a solid terminal background color, but it destroys the look of the requesters and gadgets. A necessary compromise. Anyways, edit the palettes in the .default file to switch the custom palettes over to the normal palettes so that when you switch emulations you don't have to also update the palette. ---- One other thing was the speed at which the requesters are opened and closed. This is especially noticable if you're using 512k of chip. If you are, Terminus will enable {CONSOLE}{CHIP RAM SAVER}{I}, it really saves on chip use, but slows down the requesters a good bit. They're also slowed down a bit due to the dynamic allocation of all the gadgets, text and some other internal structures when each requester is opened. This can't be helped. BTW, as more 2.0 specific applications come out you'll see this more often as the GadTools and BOOPSI gadget generation will cause much the same delay. The savings is in code size by doing things this way. ---- Quite a few comments have been made about the way the menus and their hotkeys are setup. I've included a table here of the way they were in JR-Comm and the way they exist now in Terminus. I'm sure a few changes could be made, but I want to be able to have each menu item have a hotkey as a lot of requests for that were made. But, due to that restriction, some compromises are inevitable. Anyways, post some sugestions on changes in the beta conference on what should be changed, maybe editing and posting this list would help so that everyone can hash it out together. Key JR-Comm Terminus uppercase Terminus lowercase ----------------------------------------------------------------------- 0 Print registration 1 Phonebook directory Serial 2 Transfer parameters Transfer 3 Serial prefs Macros 4 Modem prefs Palette 5 Terminal prefs Paths 6 Macro prefs Modem 7 Palette Port 8 General General 9 Console A Append capture Append capture Append w/save B Send break Load buffer Send break C Close capture Clear buffer Close capture D Download Dial number Start download E Print screen Start exact search Redial F Open w/save Start search Repeat search G Display file ANSI mouse? H Hangup HEX display? Hangup modem I Append w/save Info ASCII send J Send macro Hide stats? K Kill buffer Iconify Doorway? L LoadDefaults Load defaults M IBM Doorway Reset timer Chat? N Screen title Decrease Increase O OpenCapture Open capture Open w/save P Printer Printer? Print screen Q Quit Quit R Redial Record Resume S ASCII send Quote clipboard Send clipboard T Timer reset Start Stop U Upload Add upload Start upload V View buffer Review from Review at end W WriteDefaults Write defaults Quiet serial? X Clear screen Reset console Clear screen Y Chat line Create entry Update entry Z Send Password Send password Directory One last comment on this stuff. The uppercase 'Q' for quitting Terminus will stay and the lowercase 'q' will remain unassigned. The reason for this was due to a ton of complaints over the past few years about the query requester. It seems that some people can't read the docs, nor do some browsing around the requesters before flaming the crap out of me. Since I'm not too fond of having a "simple" hotkey that can zap an application without asking the user first, and since a lot of people seem to be annoyed when they *are* asked, the only way I could work this out was to use a key sequence that requires a bit of thought in order to exit the program. It may take a bit of time to get used to it, even I had to go through this, but once you do it becomes quite painless and second nature to do. Please use it for a few weeks and see if I'm right. ---- It seems I took the ability for this program to multitask within itself a bit too far. I've had to add an awful lot of code to enable/disable menu items depending on what the program is doing. Most of this centers around dialing and file transfers which require that most menu items that can affect the enviroment, paths, serial device, console and such be disabled while doing that particular operation. Another is script recording, the other 3 items in the script menu are disabled while recording. If I somehow munged the logic and an item doesn't get reenabled afterward, please let me know. ---- The file stats requester presented a bit of problem. Even though you could open up the review buffer to browse through it while transfering a file, you couldn't get to see the entire buffer since you can't get rid of the requester on a custom screen, only on Workbench. I've added a menu item "Hide stats" that will close the requester and post some info to the status line while the transfer is in effect. It's nowhere's as detailed, all it shows is "XFERING: NNNNNN of NNNNNN" which is the number of bytes out of the filesize. Of course, if you are receiving an XMODEM file you won't know the exact size beforehand, so the second number will be zero. The logfile has the start of the file transfer recorded as well as the end, even if it was aborted, along with stats on the transfer itself. The general operation of file transfers is much cleaner now than it was with JR-Comm, especially when it comes to abort detection from the user or from the remote system. Please verify the integrity of this as well as the protocols themselves. ---- The review buffer got cleaned up, there were some real problems in there. Also, the clip function now supports reverse clipping. Please check to see if the placement of the review cursor and what you get agrees with what you think should be clipped. I'll explain this a bit more. When clipping forward the first character in the clip is, of course, the character you started the clip at, which is called the "anchor". The end of the clip is the character *previous* to the cursor, not the one under it. When you do a reverse clip however, the "anchor" becomes the last character in the clip and the first character is the one underneath the cursor, not the one after it. I'm pretty sure that this is the most intuitive way for it to clip, just want some feedback from you to be certain though. ---- Ok, onto the phonebook and dialer. Quite a bit changed here due to me not being satisfied with the way it worked before. For one, I didn't like the fact that you couldn't add to or take away from the selected list when dialing. Now you can. You can also add multiple numbers via the item [PHONEBOOK][DIAL NUMBER...][D], although the dialer is immediately started when the first one is entered. The dialer now logs the connected baud rate to the logfile, this is not the fixed rate in the serial settings requester, it's the rate returned from the CONNECT response. That is also sent to the screen as well. The phonebook file is also written after a connection is made so that the connection date can be updated for the entry that was just connected. No entries with the names "New Entry" or "Dial Number" will *not* be written to the file. When carrier is lost, or when you hangup, the time connected and cost (if {ENTRY}{CENTS/MIN.}{/} is non-zero) is logged. One thing that I couldn't fix due to the way the main and serial processes interact with one another is the logging of carrier loss while a file transfer is in effect. What happens is that the entry connection is logged first, then the aborted file transfer. Correcting it is something that's a low priority at the moment. The highlight bar now moves down one line when selecting an entry via the key. Also, scrolls the entry list up/down a page at a time with the and cursor keys while will move to the top or bottom of the list. ---- Another addition is a menu item for enabling/disabling the ANSI mouse. I couldn't figure out how to handle the logic that causes the screen to clear when it is toggled from within the console requester without resorting to a new way of handling things that would require some major surgery. Besides, this is more convienent anyway, as was pointed out in the beta conference. ---- Well, that's about it for now, enjoy! === 1.901 ================================================================== Ok, here's the first beta of Terminus 2.0. Yep, JR-Comm has had a name change! Anyways, this release does not have any file transfers in it, just ASCII send. I'm going to be testing out that portion of the program while you have at the emulations, scripts and everything else. It should be about two weeks before the file transfer stuff is debugged, so concentrate on the other stuff in the meantime. Terminus uses a completely different set of fonts, please install them in your FONTS: directory. Also, copy the other files to a directory and assign the logical name TERMINUS: to that directory. I've included a few scripts with this beta, a demo program that tests most of the script language and a modem initialization script that should allow anyone to setup just about any modem in existence. Please test it with your modem and see if it gets it setup without any additional assistence. One other script is host.scp. It's not a complete host yet, I just haven't had the time to work on it. What it currently does is emulate the login sequence and user check for password and security level. It doesn't answer the modem or do anything else yet, maybe one of you can get the script more fleshed out so that it can actually be used as a simple host. Remember to look at the REMOTECLI command and include it in the script if you do decide to work on it. The program should be fairly solid now as alot of problems have been eliminated during alpha testing, but I'm sure some other things still lurk around in there. The doc file is huge as you've no doubt noticed. It's also formatted so that it can be printed to a laser or pin printer. You'll need to set up each printer a bit differently in order to print the docs correctly, so follow these directions. Laser printers: --------------- Set up your printer preferences to 6lpi, 10cpi, 60 lpp and left column to 0. You also have to make sure that your printer does not have a skip over perforation option that is on, if it is, turn it off or you're going to get a blank sheet of paper on every other page. Pin printers: ------------- Set up your printer preferences to 6lpi, 10cpi, 66lpp and the left column to 5. Skip over perforation should also be turned off and it should advance the paper whenever a form feed is seen. Finally, set the top of page 3 lines down from where it normally is as there are only 60 lines of text for each page. Use the CLI command type >prt: term.txt and you should get a 159 page document. If you have any problems with this please let me know. Anyways, there's alot of work left to do, I need to write the installation program and to work on a few other things besides the file transfer code. This is a pretty rough beta, that's why only a few of you have access to it at the moment. I'll add people to the testing over the next few weeks and things should shape up rather fast now. Hope you enjoy this beast, it's quite a bit different than JR-Comm... -jack-