sarima@gryphon.COM (Stan Friesen) (03/01/89)
Hello again. Tanks to all of you who responded to my question about pseudo-ttys. It was quite helpful. I will try to find some time to assemble a summary of the responses to post, since I got several requests for that. Now, on to my real current problem. I am trying to attach cpio as a background/slave process while retaining the ability to process multiple tapes. The problem: the moment it tries to *read* the response from the master program telling it that the tape has been switched, it dumps core! At first I assumed that this was because it opened /dev/tty to read the response. That is why I wanted pseudo-ttys, so I could make the slave end of one the control terminal for the cpio. Well, all my test programs worked perfectly, the pseudo-ttys were opened and attached as control terminal, my test master properly did I/O on them and the slave's open() of /dev/tty went to the pseudo-tty as I wished. So I assumed everything was peachy and installed the whole mess in my actual application. Result: no change in the behavior of cpio, it still dumps core on the read of /dev/tty!! I have verified, from another terminal, that the control terminal for the cpio is in fact the slave end to the pseudo-tty, so I *assume* it must be proeprly attached. Could anyone help me? Why is the cpio dieing?? What can I do to get it to properly read the response from the master? -- Sarima Cardolandion sarima@gryphon.CTS.COM aka Stanley Friesen rutgers!marque!gryphon!sarima Sherman Oaks, CA
gregg@ihlpb.ATT.COM (Wonderly) (03/02/89)
From article <12715@gryphon.COM>, by sarima@gryphon.COM (Stan Friesen): > > Could anyone help me? Why is the cpio dieing?? What can I do to > get it to properly read the response from the master? Is it possible for you to run sdb on the core image and look at what is happening? Use a 't' and 'x' command to get all the gorry details... -- Gregg Wonderly DOMAIN: gregg@ihlpb.att.com AT&T Bell Laboratories UUCP: att!ihlpb!gregg