bagchi@eecs.umich.edu (Ranjan Bagchi) (01/06/91)
anyone care to explain?? (I`m in csh)
$ set p = "w | tail +3 "
$ echo $p
w | tail +3
$ $p
Usage: w [ -hlsuw ] [ user ]
$ w | tail +3
benjo ttyp3 10:13pm 41 4 -sh
benjo ttyp4 10:17pm 3:22 4 -csh
$ exit
$ exit
Process shell finished
shouldn't "$p" have the same effect as "w | tail +3" ?
-rj
--
--------------------------------------------------------------------------------
Ranjan Bagchi - At Large. Well Kinda. | what kind of person
bagchi@[eecs | would want to count syllables
caen, | just to write haiku?
math.lsa].umich.edu |
--------------------------------------------------------------------------------boehme@unvax.union.edu (Eric M. Boehm) (01/07/91)
bagchi@eecs.umich.edu (Ranjan Bagchi) writes: >$ set p = "w | tail +3 " >$ echo $p >w | tail +3 >$ $p >Usage: w [ -hlsuw ] [ user ] >$ w | tail +3 >benjo ttyp3 10:13pm 41 4 -sh >benjo ttyp4 10:17pm 3:22 4 -csh >$ exit >$ exit >shouldn't "$p" have the same effect as "w | tail +3" ? No, it should not. It has to do with the way the C Shell parses the command line. The following is taken from "The UNIX C Shell Field Guide, Gail Anderson, Paul Anderson, Prentice Hall, 1986, ISBN 0-13-937468-X", pages 262-266. (I highly recommend it). 1. History substitution - i.e.,interpret ! and ^. 2. Finding words - character strings delimited by blanks or tabs. Treat special characters &, |, ;, >, <, (, ), &&, ||, >>, << as words unless quoted. 3. Update the history list - after breaking command into words, put on history list. 4. Parse sequence of words in the following order: a. Quoting with ' and " b. Alias substitution c. I/O redirection, background execution, and pipes (recognized but not processed) d. Variable substitution e. Command substitution f. Filename expansion 5. Execute each command Relevant to your example: % set out = '>> ~/vault/words' % cat dict1 >> ~/vault/words # try to replace with the following % cat dict1 $out [contents of dict1 appear here] cat: cannot open >> "What went wrong? We wanted the C shell to expand $out to specify redirection. Unfortunately, since it has already handled redirection, it simply passes >> to the cat command as a filename argument. Note the importance of knowing precedence here." "The C shell executes a subshell to process command substitution; that is, the instance of the C shell that parses the command substitution is distinct from the one that evaluates the text inside command substitution marks." In your case, since the shell has already handled redirection, it passes | to w as an argument. -- Eric M. Boehm boehme@unvax.Union.EDU BOEHME@UNION.BITNET
mju@mudos.ann-arbor.mi.us (Marc Unangst) (01/08/91)
bagchi@eecs.umich.edu (Ranjan Bagchi) writes: > $ set p = "w | tail +3 " [...] > shouldn't "$p" have the same effect as "w | tail +3" ? No. The pipe character is not interpreted specially by the shell when you run "$p"; therefore, you are calling the "w" command with the arguments "|", "tail", and "+3". Try "eval $p". -- Marc Unangst | mju@mudos.ann-arbor.mi.us | "Bus error: passengers dumped" ...!umich!leebai!mudos!mju |
fpb@ittc.wec.com (Frank P. Bresz) (01/08/91)
In article <BAGCHI.91Jan6000722@snarf.eecs.umich.edu> bagchi@eecs.umich.edu (Ranjan Bagchi) writes: >anyone care to explain?? (I`m in csh) >$ set p = "w | tail +3 " >$ echo $p >w | tail +3 >$ $p >Usage: w [ -hlsuw ] [ user ] >$ w | tail +3 >benjo ttyp3 10:13pm 41 4 -sh >benjo ttyp4 10:17pm 3:22 4 -csh >$ exit >$ exit >Process shell finished >shouldn't "$p" have the same effect as "w | tail +3" ? > -rj Well, I'll take a shot at an explanation. By saying: set p = "w | tail +3 " You have effectively escaped the normal behavior of | which is to run a pipe to the next command in line. Thereby passing the | to the 'w' command as the first argument. 'w' does not understand what the '|' symbol means so it spits out it's usage command. This command shows what I mean: ittc% w \| tail +3 Usage: w [ -hlsuw ] [ user ] Had you instead said : eval $p You would have received the 'w' portion only the 'tail' would not execute, (at least it didn't under my csh SunOS4.1, why I don't know). This scenario worked for both bash and sh. $ p="w | tail +3" $ $p Usage: w [ -hlsuw ] [ user ] $ eval $p fpb console 12:44pm 6:15 9:32 9:27 sv_xv_sel_svc young ttyh5 12:40pm 5:54 telnet wssi3 young ttyha 11:29am 6:46 5 5 telnet wssi3 miller ttyhc 2:33pm 3:33 3 -csh fpb ttyp0 12:44pm 28 /emacs/sun4/etc/bash fpb ttyp3 6:26pm 9 1 w mance ttyp9 3:10pm 3:49 - In CSH however I tried and received : ittc% set p = " w | tail +3 " ittc% $p Usage: w [ -hlsuw ] [ user ] ittc% echo $p w | tail +3 ittc% eval $p 7:03pm up 2 days, 12:04, 6 users, load average: 1.60, 2.24, 2.05 User tty login@ idle JCPU PCPU what fpb console 12:44pm 6:19 9:45 9:40 sv_xv_sel_svc young ttyh5 12:40pm 5:58 telnet wssi3 young ttyha 11:29am 6:50 5 5 telnet wssi3 miller ttyhc 2:33pm 3:37 3 -csh fpb ttyp0 12:44pm 32 /emacs/sun4/etc/bash fpb ttyp3 6:26pm 10 1 w mance ttyp9 3:10pm 3:54 - ittc% Clearly the tail failed. Any CSH wizards know why. I received bash about 4 months ago and haven't touched csh since. Out of pure inquisitiveness what is wrong with : alias p "w | tail +3" p -- | () () () | Frank P. Bresz | Westinghouse Electric Corporation | \ /\ / | fpb@ittc.wec.com | ITTC Simulators Department | \/ \/ | uunet!ittc!fpb | Those who can, do. Those who can't, simulate. | ---------- | +1 412 733 6749 | My opinions are mine, WEC don't want 'em.
allbery@NCoast.ORG (Brandon S. Allbery KB8JRR) (01/08/91)
As quoted from <BAGCHI.91Jan6000722@snarf.eecs.umich.edu> by bagchi@eecs.umich.edu (Ranjan Bagchi): +--------------- | anyone care to explain?? (I`m in csh) | $ echo $p | w | tail +3 | $ $p | Usage: w [ -hlsuw ] [ user ] | $ w | tail +3 | benjo ttyp3 10:13pm 41 4 -sh | benjo ttyp4 10:17pm 3:22 4 -csh +--------------- You'll find that sh does the same thing. The reason is that variables are handled *after* the line is broken into commands. Since | is a command separator, it is parsed relatively early. After sh or csh does command parsing, it expands variables... and since it's too late to handle |, it's passed literally. Similarly, < > is (I think) parsed *but not necessarily executed* before variable substitution. I also seem to remember redirecting into the output of a command working under sh but not under csh.... Shells are *weird* beasties. ++Brandon -- Me: Brandon S. Allbery VHF/UHF: KB8JRR on 220, 2m, 440 Internet: allbery@NCoast.ORG Packet: KB8JRR @ WA8BXN America OnLine: KB8JRR AMPR: KB8JRR.AmPR.ORG [44.70.4.88] uunet!usenet.ins.cwru.edu!ncoast!allbery Delphi: ALLBERY
rickert@mp.cs.niu.edu (Neil Rickert) (01/08/91)
In article <BAGCHI.91Jan6000722@snarf.eecs.umich.edu> bagchi@eecs.umich.edu (Ranjan Bagchi) writes: >anyone care to explain?? (I`m in csh) > >$ set p = "w | tail +3 " > >shouldn't "$p" have the same effect as "w | tail +3" ? > I suppose someone should point out that, taken literally, "$p" does have the same effect as "w | tail +3 ". % set p="w | tail +3 " % "w | tail +3 " w | tail +3 : Command not found. % "$p" w | tail +3 : Command not found. Of course this wasn't what you intended, but it also demonstrates that quoting does effect the meaning of shell commands. -- =*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*=*= Neil W. Rickert, Computer Science <rickert@cs.niu.edu> Northern Illinois Univ. DeKalb, IL 60115 +1-815-753-6940
christos@batcomputer.tn.cornell.edu (Christos S. Zoulas) (01/08/91)
In article <1991Jan7.142732.28884@unvax.union.edu> boehme@unvax.union.edu (Eric M. Boehm) writes: >bagchi@eecs.umich.edu (Ranjan Bagchi) writes: > >>$ set p = "w | tail +3 " >>$ echo $p >>w | tail +3 >>$ $p >>Usage: w [ -hlsuw ] [ user ] >>$ w | tail +3 >>benjo ttyp3 10:13pm 41 4 -sh >>benjo ttyp4 10:17pm 3:22 4 -csh >>$ exit >>$ exit > >>shouldn't "$p" have the same effect as "w | tail +3" ? > >No, it should not. It has to do with the way the C Shell parses the >command line. The following is taken from "The UNIX C Shell Field >Guide, Gail Anderson, Paul Anderson, Prentice Hall, 1986, ISBN >0-13-937468-X", pages 262-266. (I highly recommend it). > ... lines deleted... The above information is correct. On the other hand: $ eval $p should work, but it does not because eval in csh is broken. christos -- +------------------------------------------------------------------------+ | Christos Zoulas | 389 Theory Center, Electrical Engineering, | | christos@ee.cornell.edu | Cornell University, Ithaca NY 14853. | | christos@crnlee.bitnet | Phone: Disconnected | Fax: (607) 254 4565 |
greywolf@unisoft.UUCP (The Grey Wolf) (01/09/91)
In article <BAGCHI.91Jan6000722@snarf.eecs.umich.edu> bagchi@eecs.umich.edu (Ranjan Bagchi) writes: >anyone care to explain?? (I`m in csh) > > >$ set p = "w | tail +3 " >$ echo $p >w | tail +3 >$ $p >Usage: w [ -hlsuw ] [ user ] >$ w | tail +3 >benjo ttyp3 10:13pm 41 4 -sh >benjo ttyp4 10:17pm 3:22 4 -csh >$ exit >$ exit > >Process shell finished > > >shouldn't "$p" have the same effect as "w | tail +3" ? > > -rj Try: % eval $p Reasoning: In csh, while the arguments remain individual in the case of variable substitution, it is as though each argument has been quoted; specifically, it is though you have requested the following command line to be executed: % "w" "|" "tail" "+3" The pipe is never seen by the shell. To understand this one, try the following: % set p = "echo | & < >" % $p The same result occurs when one tries: % set p = ( echo | & < > ) % $p eval seems to be the only way to work around this problem. -- On the 'Net: Why are more and more fourth-level wizard(-wannabe)s trying to invoke ninth-level magic, instead of taking the time to climb the other (quite essential) thirteen levels so they can do this properly? ...!{ucbvax,acad,uunet,amdahl,pyramid}!unisoft!greywolf
speelmo@bronze.ucs.indiana.edu (Lance Speelmon -- UCS) (01/11/91)
In article <BAGCHI.91Jan6000722@snarf.eecs.umich.edu> bagchi@eecs.umich.edu (Ranjan Bagchi) writes: >anyone care to explain?? (I`m in csh) > Try: set p = `w | tail +3` i think that should fix it... Lance > >$ set p = "w | tail +3 " >$ echo $p >w | tail +3 >$ $p >Usage: w [ -hlsuw ] [ user ] >$ w | tail +3 >benjo ttyp3 10:13pm 41 4 -sh >benjo ttyp4 10:17pm 3:22 4 -csh >$ exit >$ exit > -- ============================================================================== | Lance Speelmon | University Computing Services | | speelmo@bronze.ucs.indiana.edu | Network Operations Center | ==============================================================================
allbery@NCoast.ORG (Brandon S. Allbery KB8JRR) (01/12/91)
As quoted from <1991Jan10.182958.21295@bronze.ucs.indiana.edu> by speelmo@bronze.ucs.indiana.edu (Lance Speelmon -- UCS): +--------------- | In article <BAGCHI.91Jan6000722@snarf.eecs.umich.edu> bagchi@eecs.umich.edu (Ranjan Bagchi) writes: | >anyone care to explain?? (I`m in csh) | Try: | set p = `w | tail +3` | i think that should fix it... +--------------- Ah, no. This will set "p" to the output of "w | tail +3"... once and only once. Ranjan is trying to use a variable as an alias. ++Brandon -- Me: Brandon S. Allbery VHF/UHF: KB8JRR on 220, 2m, 440 Internet: allbery@NCoast.ORG Packet: KB8JRR @ WA8BXN America OnLine: KB8JRR AMPR: KB8JRR.AmPR.ORG [44.70.4.88] uunet!usenet.ins.cwru.edu!ncoast!allbery Delphi: ALLBERY