jlhaferman@l_cae07.icaen.uiowa.edu (Jeffrey Lawrence Haferman) (07/31/90)
SR 10.2, /bin/csh is my shell, I can't get history | tail -15 to work. 'history' works fine, tail <filename> works fine, I can't get the pipe to work. Ideas? (I don't get an error back, just seems to wait a little while, and then I get a prompt back.) Jeff Haferman internet: jlhaferman@icaen.uiowa.edu Department of Mechanical Engineering University of Iowa Iowa City IA 52240
jimr@metro (Jim Richardson) (08/01/90)
In article <2011@ns-mx.uiowa.edu>, jlhaferman@l_cae07.icaen.uiowa.edu (Jeffrey Lawrence Haferman) writes: > > SR 10.2, /bin/csh is my shell, I can't get > > history | tail -15 > > to work. 'history' works fine, tail <filename> works > fine, I can't get the pipe to work. Ideas? > > (I don't get an error back, just seems to wait a little > while, and then I get a prompt back.) I haven't observed this problem with history: perhaps it depends on how many lines the history contains. However, I have similar problems with commands like tail -300 syslogd.log where syslogd.log is a logfile for syslogd, and therefore constantly open for write, if it's important (that's why tail is needed instead of cv). If I ask tail for only say 250 lines, all is well. But over about 300, the command sometimes just sits there until interrupted, and sometimes as Jeffrey reports it pauses then comes back to the prompt with no output. There seems to be some bad interaction between tail and the DM going on here. If after the tail command has failed, you push and pop the pad in which it was executed, you sometimes see very odd changes: some of the lines that tail should have produced appear and/or disappear. In fact, the pad seems to be in a very unstable and inconsistent state. Redirection of output via e.g. "tail -300 syslogd.log" > /tmp/slog succeeds correctly. Is that also true for Jeffrey's history example? -- Jim Richardson Department of Pure Mathematics, University of Sydney, NSW 2006, Australia Internet: jimr@maths.su.oz.au ACSNET: jimr@maths.su.oz FAX: +61 2 692 4534
petersn@physics.utoronto.ca (Mike Peterson) (08/02/90)
In article <2011@ns-mx.uiowa.edu> jlhaferman@l_cae07.icaen.uiowa.edu (Jeffrey Lawrence Haferman) writes: > >SR 10.2, /bin/csh is my shell, I can't get > > history | tail -15 > >to work. 'history' works fine, tail <filename> works >fine, I can't get the pipe to work. Ideas? > >(I don't get an error back, just seems to wait a little > while, and then I get a prompt back.) Pipe commands will fail at SR10.2 / SR10.2.p if the first component of the pipeline manages to complete before the pipeline is set up (it is not just history that does this, but it is one of the easiest ones since it is built-in to csh). I have called it in and APR'ed it -- it will not be fixed in SR10.2, but is supposedly fixed in SR10.3. Mike Peterson, U/Toronto Chemistry. system@alchemy.chem.utoronto.ca
rees@pisa.ifs.umich.edu (Jim Rees) (08/03/90)
In article <1990Aug1.022258.7408@metro.ucc.su.OZ.AU>, jimr@metro (Jim Richardson) writes: In article <2011@ns-mx.uiowa.edu>, jlhaferman@l_cae07.icaen.uiowa.edu (Jeffrey Lawrence Haferman) writes: > > SR 10.2, /bin/csh is my shell, I can't get > > history | tail -15 > > to work. 'history' works fine, tail <filename> works > fine, I can't get the pipe to work. Ideas? This works for me. It may depend on whether you're in a pad, xterm, or what. "history" is a builtin and may have redirection strangeness (the redirection is done in an entirely different way from how it's done for normal commands). I haven't observed this problem with history: perhaps it depends on how many lines the history contains. However, I have similar problems with commands like tail -300 syslogd.log "tail" has a limit on how many lines it can store up, although it should fail gracefully. The "tail" man page seems to be out of date. In particular, the '-f' option works but is undocumented. This is terribly useful for files like syslogd.log. It monitors the file and regularly prints out any extra lines that came in recently.