> A reservation on usage of "/dev/ptmx" named object...
>
> I recall having heard Alan Cox to say NAY on this before on
> basis of this easily breaking iBCS -- Running SCO binaries
> on top of iBCS detects /dev/ptmx NAME, and then the libc stuff
> statically linked into those binaries ASSUMES existence of
> STREAMS in the system. (Need I continue ? Naeh, ask Alan..)
It seems that iBCS needs to be fixed. Linux shouldn't be broken because
of SCO binaries. /dev/ptmx is standard. Should we break Unix98
compatibility because SCO binaries running on iBCS make unwarranted
assumptions?
> If you could make provisions to have different name for the
> PTY-MUX device, all will be well. (But that is in fact user-
> space libc issue, not about this facility per se.)
I don't think that's a reasonable solution. I'd suggest fixing iBCS.
--Scott
@ @
=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-oOO-(_)-OOo-=-=-=-=-=
C. Scott Ananian: cananian@lcs.mit.edu / Declare the Truth boldly and
Laboratory for Computer Science/Crypto / without hindrance.
Massachusetts Institute of Technology /META-PARRESIAS AKOLUTOS:Acts 28:31
-.-. .-.. .. ..-. ..-. --- .-. -.. ... -.-. --- - - .- -. .- -. .. .- -.
PGP key available via finger and from http://www.pdos.lcs.mit.edu/~cananian