Re: problems using a pty

[buhrow%lothlorien.nfbcal.org@localhost (Brian Buhrow)]

        Hello.  More questions.  

1.  Are you using the functions in libcurses(3) to manipulate the terminal
or are you using terminal ioctls directly to do it?  If I do a man cbreak I
get a function in the libcurses library, but I know you can use ioctls or
the stty(1) command to do the same thing.  If you're using curses, and you
do it with ioctls directly, does the behavior change when you switch modes?

2.  If you use select(2) rather than poll(2) do you get the same behavior?

3.  If things work before you call cbreak() but break after you call it, do
they work again after you call nocbreak(), assuming you're using the
currses interface?

        I'm wondering if this is either a bug in poll(2) or a bug in
curses(3).

Hope this helps.
-Brian
On Feb 19, 11:26pm, Brett Lymn wrote:
} Subject: Re: problems using a pty
} On Fri, Feb 12, 2010 at 09:43:05AM -0800, Brian Buhrow wrote:
} >     Hello.  What wait channel is the slave process waiting on? 
} 
} ttyout
} 
} > Can you
} > try this test on a 4.x system? 
} 
} Well, after installing qemu from pkgsrc and doing an install, yes.  I
} observe the same problem on a 4.x system.
} 
} > .  If you test on a 4.x system and it works, then it's a bug in
} > the pty code, if not, then it's probably something in your test.
} > 
} 
} Well, the following things puzzle me.. I won't rule out me doing
} something wrong but:
} 
} 1) without a call to cbreak() things work fine
} 2) if I take my section of test code, wrap a main(){} around it, build
}    and run that then things work as the should.
} 3) why is an ioctl on the slave stdin affecting other file
}    descriptors?  If I do a ktruss on the slave using a wrapper script
}    then I can see that:
}         a) read returns immediately with 0 bytes read when it should
}            wait for a character.
}         b) the slave is writing to the pipe back to the director but a
}            poll() on the director side of that same pipe says there is
}            no data to read.... without cbreak() there is data to read
}            - I know this because the director polls the pipe fd before
}            it reads to avoid blocking when the slave either dies or
}            does not return output when it should.  Without cbreak()
}            the whole poll/read data thing happens fine.
}         c) writes to a slave debug dribble file stop as well, after every
}            write I fflush the file handle so data should be written
}            but the writes seem to stop after the cbreak() call.  In
}            the kdump output I can see write being called on the file
}            handle.
} 
} > Just my thoughts.
} > 
} 
} thanks they were appreciated :)
} 
} -- 
} Brett Lymn
} "Warning:
} The information contained in this email and any attached files is
} confidential to BAE Systems Australia. If you are not the intended
} recipient, any use, disclosure or copying of this email or any
} attachments is expressly prohibited.  If you have received this email
} in error, please notify us immediately. VIRUS: Every care has been
} taken to ensure this email and its attachments are virus free,
} however, any loss or damage incurred in using this email is not the
} sender's responsibility.  It is your responsibility to ensure virus
} checks are completed before installing any data sent in this email to
} your computer."
} 
} 
>-- End of excerpt from Brett Lymn