Balazs Scheidler <[EMAIL PROTECTED]> writes:
> On Sat, 6 Feb 1999, Keresztfalvi Gabor wrote:
>
> > Hi all!
> > The call to tty_setctty in server.c is too late in the code...
> > In make_pty there is in[0]=pty->slave, so if we close(in[0]) prior to
> > calling tty_setctty(pty->slave) then we'll get a errno=9 (bad file
> > handler).
Hmm, I think that some earlier version used tty_setctty(STDIN_FILENO),
but I changed it in an attempt to make the code a little clearer.
Undoing that change should solve the problem, I think.
> > BTW is somebody working on tty_encode_term_mode out there? Or may I start it?
> > (First for POSIX systems (with termios)).
>
> None that I know of. However Niels doesn't seem to respond to messages (at
> least for mine ones) I don't know what happened to him.
I'm still alive ;) although I was too busy last week. The latest
things I have been doing on lsh, before vanishing, is a mechanism for
allocating strings whose lengths are not known in advance. This is
useful for the parts of the "advanced format" s-exp-parser which are
not written yet, for allocation for zlib output strings, and perhaps
also for constructing term-mode strings.
I've not done anything about tty_encode_term_mode yet, so it would be
great if Keresztg could start on that. For allocation, keep it simple,
just allocating the worst-case size should do fine for now. If you
think that should be improved, let me know and we can fix that later.
Happy hacking,
/Niels
