KERMIT FOR U-MAN 1000 UNDER CP/M-68k ==================================== This is a very untested version, but has been used to xfer files! Any improvements would be welcome. Know faults include Very poor intergration of parity with 8 bit quoting. The program will do quoting if requested, but should do so automatically if parity is selected. No documentation. The program is a ripoff of the C kermit in the protocol manual, with a few changes to make it a local only kermit. The program sends filesles "as-is" as it should have CR-LF on the end of every line already. This means that like MS-KERMIT it does not need to be told if the file is ASCII or binary. Some option should be added to allow it to detect the end of file correctly. With the ned of file problem etc some work should be done to check if creata creatb or creat should be used, and also the fopen. The simple terminal program enclosed is the guts of the connect command. This may be enough to type in to allow a series of commands to be sent to a mainframe to download the file. I found that pip would work at 600 baud to download the file. To compile kermit run the C compiler on the main program and assemble the small bdos.s file. Then link them together. There is no floating point so clink can be used. ------------------------------------------------------------------------- From: ORA021 @ LANCS.VAX1 6-JAN-1986 17:41 To: KERMIT Subj: To Icarus Sparry University of Bath December 26th. 1985 U-MAN 1000 KERMIT Thanks for this, which I moved over to my U-MAN via an IBM PC during the Christmas holiday. I am returning a version with some bug fixes. My changes are all flagged with comments including the string jgc and a date e.g /* jgc jan 1st 1986 */ The changes are:- The replacement of fopen with (*kfopen), kfopen is declared as pointer to function returning pointer to FILE. There is a new function en8quote() which sets the qflag to match the parameter supplied and selects either fopen or fopenb for kfopen accordingly. Calls to en8quote replace your assignments to qflag. In the command-line parse switch the parity select now drops through to select quoting if parity is other than null. In CP/M it is not possible to detect end of file at the byte level with binary files so the whole final sector is transmitted if quoting is selected, including the case where a text file is opened as binary, in order to send it with parity checking. It all works OK but it would be better to add an extra parameter to distinguish between ascii and binary for file opening purposes (as distinct from quoting or CRLF <-> LF translation). Then the text end of file character would be recognised, restoring single byte granularity to these files. How do you feel about this ? Your first version and this one both go between the U-MAN and IBM very smoothly at 9600 baud. John G. Crookes -------------------------------------------------------------------------- From: ORA021 @ LANCS.VAX1 6-JAN-1986 18:01 To: KERMIT Subj: To Icarus Sparry University of Bath January 6th. 1986 U-MAN 1000 KERMIT There is a further bug which I just noticed but haven't fixed. The kermit doesn't like to be started first and hangs if it is sending. John G. Crookes ---------------------------------------------------------------------------