Patching QTERM for your system. This explains the patches in QTERM, and can be used to patch QTERM directly (it is written as if being used in that manner), however it also provides an explanation of the subroutines that would be needed if a QT-?????.Z patch source were to be written, based on the template QT-PATCH.Z provided. The first thing to do is to back QTERM up, and then invoke DDT, SID, ZSID, Z8E, or whatever your local patch utility is, in the following way: A>DDT QTERM.COM DDT (etc.) will read in QTERM, and then prompt. The following is a list of patch areas where QTERM should be changed to reflect your system. Some of these are mandatory (i.e. QTERM won't work without them), whereas others can be changed to null subroutines or empty data without preventing QTERM from working, it just won't have all the features available. 1. Modem input status: 0110 - 011F QTERM calls here to check modem input status. Return with the zero flag set if no character is available, or with the zero flag clear if a char is available. Generally this can be an input from the usart / sio / dart status port followed by an 'and'. 2. Read modem character: 0120 - 012F This gets a character from the modem input port once the input status has decided it's there. Return the character in the a register. Generally this can be an input from the usart / sio / dart data port. 3. Modem output status: 0130 - 013F Check if the modem output port can accept another character. Return with the zero flag set if the output port can't receive a character, or with the zero flag clear if the output port is ready. Generally this can be an input from the usart / sio / dart status port followed by an 'and'. 4. Write modem character: 0140 - 014F Send the character in the a register to the modem output port. This will only be called after the output status routine has returned a non-zero status. Generally this can be an output to the usart / sio / dart data port. These first four patches are all necessary for QTERM to work. The next few are not necessary, but they will be useful. 5. Start break: 0150 - 015F End break: 0160 - 016F The start break subroutine at 0150 should initiate a break condition on the modem output line, and 0160 should clear the break condition. If these are to be omitted, then just put return (C9) instructions at 0150 and 0160. Note that the Start Break routine need not check that the transmit buffer is empty, since there will always be a 1/10th. second delay after the last character is sent, before calling this subroutine. 6. Drop DTR: 0170 - 017F Restore DTR: 0180 - 018F The drop DTR subroutine causes DTR to be made inactive, and restore DTR returns DTR to an active state. If your modem does not respond to DTR, but can be made to hang up by sending a string, then put a return (C9) at 0170. Use the space from 0171 to 018F to contain the string, with the following notes: at 0171 should be the length of the string, to transmit a break, use an 0FFH byte, to cause a two second delay use an 0FEH byte. Hence the following could be used to hang up a Hayes compatible: 0C FE FE 2B 2B 2B FE FE 41 54 48 30 0D 0C - length: 12 bytes follow FE - delay (twice) 2B - '+' sent three times FE - delay (twice) 41 54 48 30 0D - ATH0 If neither DTR nor a string is to be used, then place a return (C9) at 0180 and 0171, and a nop (00) at 0170. The string is used only if a C9 is found at 0170, so by placing the C9 at 0171 the string print is inhibited. 7. Baud rate setting: 0190 - 019F Baud rate table: 01A0 - 01AF These two patch areas work together to allow QTERM to change the baud rate of the modem port. The baud rate table holds pairs of bytes for setting the baud rate to eight different values: 38400, 19200, 9600, 4800, 2400, 1200, 600 and 300, in that order. In these pairs, the first byte will be passed to the subroutine at 0190, and the second byte is used to enable that baud rate: an 0FFH in the second byte enables the rate, and a zero disables. So if your system only went up to 9600, (using a value of 1 to get 9600) the first six bytes in the table would be: 00 00 no value for 38400: disable by the 00 00 00 no value for 19200: disable by the 00 01 FF 01 is the value for 9600: enable by the FF In all cases of enabled baud rates, the subroutine at 0190 gets the appropriate value in the a register and should use it to set the baud rate. If this is to be omitted, then just put a return (C9) instruction at 0190, and fill the table from 01A0 to 01AF with 00's. 8. Communication mode setting: 01B0 - 01BF Communication mode table: 01C0 - 01CB These two patch areas work together to allow QTERM to change the communications format of the modem port. The mode table holds bytes for setting 12 different formats, selecting number of data bits (7 or 8) parity (odd, even, or none) and number of stop bits (1 or 2). In order the 12 values are for 7n1, 8n1, 7n2, 8n2, 7e1, 8e1, 7e2, 8e2, 7o1, 8o1, 7o2, and 8o2. The subroutine at 01B0 gets one of these values in the a register and should use it to set the communications mode. If this is to be omitted, then just put a return (C9) instruction at 01B0. 9. Reserved for later use: 01CC This byte is reserved for later expansion, and should not be used. 10. Protocol transfer size: 01CD During protocol transfers, disk reads and writes take place every 8K. This is normally possible without causing a timeout, and reduces disk access to a minimum. However if your disk is slow, you can drop this to 4, 2 or even 1 to reduce the size of transfer, and hence prevent timeouts. 11. Processor speed: 01CE This is the speed in Mhz that your Z80 runs at: 4, 6 or whatever. For a 2.5Mhz cpu, use 3. 12. Escape character: 01CF All special functions of QTERM are activated by the use of escape sequences. At 01CF is the byte used for the escape character (the default is ^\). Any byte can be used, but a little used value is best selected, also using a printable character (' ' thru '~') may have undesirable results. Note that to transmit the escape value itself, just type it twice. These previous three are necessary. 13. Signon message: 01D0 - 01EF This must be a string that identifies your system / terminal. It must be present, and is printed when QTERM first starts. It should be composed of printable characters, and terminated by a zero byte. 14. Clear screen: 01F0 - 01FF This must be a string that clears the terminal screen, and leaves the cursor in the top left hand corner. This should also be terminated by a zero byte. 15. Moveto: 0200 - 022E QTERM requires the ability to move the cursor around the screen. It calls this subroutine with the required coordinates in hl: where h is the row, and l the column to move to. The top left hand corner of the screen is 0,0; and the bottom right corner is 23,79. This subroutine will have to do terminal output: at 0109H is a routine that prints a character in the c register, and at 010CH is a routine to print a decimal number in hl (mainly for the use of vt100 and vt220 compatibles). Note that the above two subroutines may destroy all registers, so appropriate action should be taken if needed. 16. Teminal capability bit map: 022F This byte contains one bit set for each of the following terminal capabilities: bit 0: (01H) end highlight mode bit 1: (02H) start highlight mode bit 2: (04H) delete line bit 3: (08H) insert line bit 4: (10H) delete character bit 5: (20H) insert character bit 6: (40H) clear to end of line bit 7: (80H) clear to end of screen 17. Terminal capability strings: 0230 - 026F In this area are eight strings, each of which can be at most seven characters long. They are the strings to be printed to perform the terminal capabilities mentioned above. Each one of them should be terminated by a zero byte. Hence at 0230 is the string for end highlight, at 0238 is the string for start highlight, etc., with 0268 being the string for clear to end of screen. Programs that use these will check the terminal capability bitmap at 022F before using them, to determine if they are available. 18. Entry subroutine: 0270 - 0272 Upon entry to QTERM, this subroutine will be called. If it is not needed then a return instruction (0C9H) should be placed at 0270, otherwise there is enough space to put in a jump to code that is to be executed when QTERM starts. This can be used for several purposes: if custom initialisation is needed to enable communications, or select a particular baud rate, or whatever, this can be done here. In addition, if all chat scripts and disk access is to be done on a specific drive, then by using the CP/M BDOS functions to set drive (and set user if desired), QTERM can be made to automatically be in the correct place to find scripts. This is explained in QTCHAT.DOC 19. Exit subroutine: 0273 - 0275 After an Q has been issued to exit QTERM, this subroutine will be called immediately before exiting back to CP/M. As with the entry subroutine, if not needed, a return instruction (0C9H) should be placed at 0273H, otherwise any termination code can be added. 20. User subroutine: 0276 - 0278 The U command from terminal mode, and !U in chat scripts cause a call to this location. This can be used to do whatever is wanted, enabling special features, selecting different ports for communication whatever. Note that at 027C is a jump to ilprmt: an inline prompt subroutine. If the user subroutine is invoked from terminal mode, then calling this subroutine will prompt, and read a line of text into the buffer at 0080, it is terminated with a zero byte. If invoked with a !u from a chat script, then the remaining text on the line will be moved to the buffer, creating the impression it had just come from the keyboard. Following the call to ilprmt should be a prompt message, terminated by a null byte. NOTE: if no prompt is required, then two zero bytes are needed. call ilprmt db 'Prompt message\0' call ilprmt db 0,0 are examples. This subroutine should only be called once per invocation of the user subroutine, since a second call when used in a chat script may have unpredictable results. 21. Keyboard map: 0279 - 027B All keystrokes read from the keyboard are passed through the keyboard map subroutine, so that actions like mapping arrow keys to VT100 escape sequences can be performed. When this is called, the value of the key just pressed is in the a register, and the b register is zero. On exit the value in b determines what action is to be taken. If b is zero, then the value passed on to QTERM is whatever vaule is in the a register, so that placing a 'RET' instruction at 0279H causes no effect at all. If b contains 1, then QTERM will assume that the keyboard map routine "swallowed" the character, and instead of passing it on, QTERM immediately polls the keyboard for another character. If b contains 2, then QTERM takes this to mean that the keyboard map routine wishes to output another character without further input from the keyboard. In this case, QTERM passes the current value in a along, then calls straight into the keyboard map routine again, without polling the keyboard. To provide some examples: A. Assume that your system has some function keys that send the following strings: ^A 1, ^A 2, ^A 3, ^A 4, and you wish to map those keys to ^H ^J ^K and ^L, with ^A followed by any other character being mapped to just the second character. The keyboard map would start by looking for ^A, if it saw any other character, it would return it unchanged with b equal to zero. On getting a ^A, it wants to see the next character from the keyboard without sending anything on, so it sets b to 1, and is at liberty to return any value in a. QTERM immediately gets the next key value, and passes it to the keyboard map. If it's one of 1, 2, 3, or 4, then the keyboard map sets a to ^H, ^J, ^K, or ^L as appropriate, and returns with zero in b, otherwise it simply returns the value in a, again with b holding zero. B. Assume you want to do the reverse mapping: ^H ^J ^K and ^L to ^A 1, ^A 2 etc. Here, the keyboard map is looking for ^H ^J etc., passing all other characters unchanged, with b zero. Assume it sees a ^H, which is to be mapped to ^A 1. It sets b to 2 (to say that there is more to come) and returns a ^A in the a register. QTERM will pass the ^A on, and then call te map again, at which point it would return 1, with b set to zero this time: this is because there are no more characters to be sent. C. In the most complex case, assume that ^E followed by any other character is to be mapped to two copies of the character, followed by ^A. In this case, all characters save ^E are passed unchanged, with zero in b. When a ^E is detected, b is returned with 1, to say that the ^E was swallowed, and when the next character is passed to the map, it should be saved, but also returned, however b should be 2. QTERM will process the character, then since b was 2, it will call the map subroutine. The map routine returns the character again, with b set to 2 a second time. On the third call to the map routine, it should return the terminating ^A, with b equal to zero to say all the work is done. 22. ILPRMT subroutine jump: 027C - 027E These three bytes are reserved to hold a jump to the in line prompt subroutine, and should not be overwritten by the patch. 23. Patch area: 0280 - 04FF Since the area provided for the above patches is limited, it may be necessary to use more space. The block of memory from 0280 to 04FF is set aside for custom patches, this can be used if the individual spaces are not big enough. Once all the patches have been made, exit the patch program (usually by typing ^C), and finish up by saving a new copy of QTERM: A>SAVE 69 QTERMNEW.COM In addition, the patch area only can be saved as follows: A>SAVE 4 QTERMPAT.XXX Which will create a 1K file containing all the patches needed to make this particular version of QTERM work. By doing this, when a new release of QTERM needs to be patched, all that is necessary is to read in the new unpatched version with DDT or whatever, then overlay the patch area. This is typically done by typing: IQTERMPAT.XXX to DDT, SID, ZSID etc. to set up the command line to read QTERMPAT.XXX, then follow this with a: R to read it. This should overlay the saved patch area on the new version, hence doing all the patches at once. Then exit DDT with ^C, and do the first save shown above to save the new working version. NOTE: With V4.2 and later, the patch area has grown yet again, so again the overlaying of earlier patches will not work. By and large, overlaying patches in this manner is not recommended, it is far easier to work with the patch sources available, applying them with ZSM and ZPATCH as needed. However, the V4.3 patch area is the same as the V4.2 patch area, so no changes are needed to convert from V4.2 to V4.3