Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!swrinde!sdd.hp.com!news.cs.indiana.edu!noose.ecn.purdue.edu!mentor.cc.purdue.edu!mace.cc.purdue.edu!abe From: abe@mace.cc.purdue.edu (Vic Abell) Newsgroups: comp.sys.next Subject: Re: Bash 1.05 as login shell on NeXT Mach 2.0 Message-ID: <7352@mace.cc.purdue.edu> Date: 29 Apr 91 19:13:37 GMT References: Distribution: comp Organization: Purdue University Computing Center Lines: 24 In article rearl@gnu.ai.mit.edu (Robert Earl) writes: >I recently compiled and installed bash 1.05 (including some bug fixes) >and then tried to make it my login shell. I tried to test it out with >an rlogin from a remote host and never got a prompt; I checked the >process from another terminal and it was using 95% CPU, and wasn't >connected to a controlling tty! This sounds similar to the error reported by pmy@vivaldi.acc.virginia.edu (Pete Yadlowsky) about using more under ksh via rlogin to a NeXT work station. The error is in rlogind, which fails to set a process group number for the shell it invokes. Apparently this doesn't bother csh, which must supply a process group number if none is set. Ksh doesn't, and probably the same is true for bash. I submitted my tentative fix to ksh-i-88f to David Korn and he answered with a better suggestion -- he had already come across the problem on an Alliant and had an #ifdef'd fix. That fix also works for NeXTStep 2.0. The fix involves reading the process group ID with getpgrp(0). If it is zero, it should be set to the process group ID that is obtained with the TIOCGPRGRP ioctl. Vic Abell