mlandau@bbn.com (Matt Landau) (08/24/87)
The recent posting of Paul Placeway's tcsh version 5.4 was most welcome,
but Rich Salz and I have found a couple of annoying bugs that we haven't
been able to track down. These all happen on a Sun 3 under SunOS 3.2, using
the vanilla BSD 4.2 shell as the basis for tcsh (we didn't have 3.2 sources at
that time). They include:
- Tcsh goes into a strange mode where it duplicates the
first few characters of every line it prints, but only
on "slow" terminals connected to serial ports. ("Slow"
meaning the tty requires lots of padding.)
- Every once in a while, tcsh will arbitrarily change your
kill character to ^W and leave you without a werase character,
even if you've bound ^W to backward-delete-word.
- Using ~ to expand usenames will sometimes result in the
message "<cmd>: bad file number".
Has anyone else seen these? Any fixes, or are our problems caused by using
the BSD csh sources instead of the Sun sources? If there are fixes, please
send them to me, and I'll pass them along to Rich for posting ASAP in
comp.sources.unix.
--
Matt Landau A rock feels no pain...
mlandau@bbn.com ...and an island never criespaul%tut.cis.ohio-state.edu@osu-eddie.UUCP (Paul Placeway) (08/29/87)
In article <8053@slate.BBN.COM> mlandau@bbn.com (Matt Landau) writes: >The recent posting of Paul Placeway's tcsh version 5.4 was most welcome, >but Rich Salz and I have found a couple of annoying bugs that we haven't >been able to track down. These all happen on a Sun 3 under SunOS 3.2, using >the vanilla BSD 4.2 shell as the basis for tcsh (we didn't have 3.2 sources at >that time). They include: > > - Tcsh goes into a strange mode where it duplicates the > first few characters of every line it prints, but only > on "slow" terminals connected to serial ports. ("Slow" > meaning the tty requires lots of padding.) Boy would I **LOVE** to find out how to fix this one. The most annoying part of it is that if you push a second tcsh, the bug goes away, but when you exit the subshell, the bug comes back. Note that tcsh is NOT the only program that does this in this state; stty does it too (do a 'stty everything' and look at the first line). It sounds like a Sun kernal bug to me (based on the stty example). > - Every once in a while, tcsh will arbitrarily change your > kill character to ^W and leave you without a werase character, > even if you've bound ^W to backward-delete-word. Easy fix. Look in ed.init.c for the places that kill is set; one of them sets it to ^W instead of ^U. Change it. > - Using ~ to expand usenames will sometimes result in the > message "<cmd>: bad file number". I havn't ever seen this one. Could someone perhaps send a repeat-by (including the machine type, OS version & which csh source was used). >If there are fixes, please >send them to me, and I'll pass them along to Rich for posting ASAP in >comp.sources.unix. I will be sending the first official patch kit Real Soon Now. This weekend our machines have to move out of their room (so that some ceiling work can be done), and we are getting rid of our Vax (osu-eddie), but as soon as things are stable... Note that since osu-eddie is going away, UUCP to here is going to be touch and go for a while. I am told that ...!cbosgd!cis.ohio-state.edu!paul should get to me, but if not, try ...!cbosgd!osu-cis!paul. -=- -- Paul Placeway Department of Computer and Information Science SNail: The Ohio State University 2036 Neil Ave. Columbus OH USA 43210-1277 ARPA: paul@ohio-state.{arpa,csnet} (soon): paul@tut.cis.ohio-state.edu change> UUCP: ...!cbosgd!cis.ohio-state.edu!paul
gore@nucsrl.UUCP (Jacob Gore) (09/01/87)
/ nucsrl:comp.sources.d / paul%tut.cis.ohio-state.edu@osu-eddie.UUCP (Paul Placeway) / 11:26 pm Aug 28, 1987 / >In article <8053@slate.BBN.COM> mlandau@bbn.com (Matt Landau) writes: >> - Tcsh goes into a strange mode where it duplicates the >> first few characters of every line it prints, but only >> on "slow" terminals connected to serial ports. ("Slow" >> meaning the tty requires lots of padding.) > >Boy would I **LOVE** to find out how to fix this one. The most >annoying part of it is that if you push a second tcsh, the bug goes >away, but when you exit the subshell, the bug comes back. [...] I haven't got to compiling this tcsh yet, but I think I may have the answer to this, from my experience with a previous version of tcsh: Try putting "set term=$TERM" in your .login Jacob Gore gore@EECS.NWU.Edu Northwestern Univ., EECS Dept. {gargoyle,ihnp4,chinet}!nucsrl!gore