peter@sugar.UUCP (Peter DaSilva) (07/12/87)
What is the mapping between odd/even/mark/space/none and the parity bits in the serial driver on the Amiga. NONE -> 0 ODD -> SERF_PARITY_ON | SERF_PARITY_ODD EVEN -> SERF_PARITY_ON MARK -> ??? SPACE -> ??? Or do you have to map stuff yourself? -- -- Peter da Silva `-_-' ...!seismo!soma!uhnix1!sugar!peter (I said, NO PHOTOS!)
wtm@neoucom.UUCP (Bill Mayhew) (07/25/87)
NONE = no parity bits added ODD = data bit appended so tthat check sum of bits = 1 EVEN = data bit appended so that check sum of bits = 0 MARK = data bit = 0 is appended SPACE = data bit = 1 is appended Note that for most practical purposes, 8N parity is the equivalent of 7S or 8S, since the the appended bit is at the same logic level as the idle state of the output. In some systems, space partiy is used to replace the option of using 1 or 2 stop bits. For instance, 8N = one stop bit 8S = two stop bits. All this parity nonsense has often resulted in hours of frustration in getting two devices to talk to each other, especially when the manufacturers notions of wheter or not the partiy bit is included in the "word length" differ! --Bill
4526P@NAVPGS.BITNET (LT Scott A. Norton, USN) (07/28/87)
In <638@neoucom.UUCP>, wtm@neoucom.UUCP (Bill Mayhew) writes, > NONE = no parity bits added > ODD = data bit appended so tthat check sum of bits = 1 > EVEN = data bit appended so that check sum of bits = 0 > MARK = data bit = 0 is appended > SPACE = data bit = 1 is appended > Note that for most practical purposes, 8N parity is the equivalent > of 7S or 8S, since the the appended bit is at the same logic level > as the idle state of the output. Robot sez, "DANGER ROBINSON FAMILY!!" DBW's VT100 uses the parity setting to influence whether Kermit transfers will use 8 bits or just seven. I found if I set parity to "None", the IBM mainframe, which sends even parity, wouldn't send Kermit packets to my Amiga, since it's Ctrl-A begin-packet characters with parity were hex 81. With "None" parity, VT100 won't strip the parity bit. >All this parity >nonsense has often resulted in hours of frustration in getting two >devices to talk to each other, especially when the manufacturers >notions of wheter or not the partiy bit is included in the "word >length" differ! >--Bill I think the problem exists because there are two ends to a duplex serial line. As a transmitter, you can do five things with the parity bit: send odd parity, even parity, mark, space, or use the parity bit as a data bit and send 8 bit characters. But as a receiver, you must know what the transmitter is doing, and then interpret the parity bit. When I see a BBS advertise 8/N/1, I read the N = No parity to mean that the board will ignore parity bits it receives, and will not be sending parity information. It will probably send spaces for parity, but I will ignore it. For Kermit 8-bit and XMODEM protocols, you don't want the serial port to mess with the parity, but you want to give it an 8-bit character and have it sent as 8 bits. So, to avoid ambiguity, the transmitter needs to specify if the parity bit is being used for error checking ( even or odd ), as an extra stop bit ( mark or space ), or as data. Then, the receiver can take the appropriate action, detecting errors, ignoring marks or spaces, or using all 8 bits as data. LT Scott A. Norton, USN Naval Postgraduate School Monterey, CA 93943-5018 4526P@NavPGS.BITNET