Re: bash vs ksh problem

From: Dan Mercer (dmercer_at_mn.rr.com)
Date: 06/01/04

  • Next message: gregg: "Re: bash vs ksh problem" 
    Date: Tue, 01 Jun 2004 15:50:49 GMT

    "Aquarius2431" <aquarius2431@yahoo.com> wrote in message news:c2d830c4.0406010355.2a6fd130@posting.google.com...
    : Using the ksh this simple test program
    : works as expected:
    :
    : #!/bin/ksh
    :
    : x=1
    : while [ x -lt 10 ]

    You should have specified $x. However, in ksh, the context with
    which a token is used can coerce it to that context. In this case
    the token "x" is used in a numeric context, so instead of treating
    it as a literal "x" it treats it as the variable x. This coercion
    is recursive, up to a limit:

        $ z=1
        $ y=z
        $ x=y
        $ [ x -lt 2 -a x -gt 0 ] && echo ok
        ok

    This is very useful at times but also can create some pretty hairy
    problems to debug, as you've seen.

    BTW, you ought to write ksh syntax if you go to all the trouble
    to use "#!/bin/ksh" in your shebang:

        typeset -i x=1
        while (( x < 10 ))
            do
            ...
            (( x = x + 1))
            done

    also, because of the coercion feature and our typesetting
    the variable as an integer, you can simply do:

        x=x+1

    ksh sees that as three tokens: "x", "=" and "x+1". Using the coercion
    feature, it knows, that "x" can't be set to a string value, so it
    recursively treats the string as an arithmetic expression. This will
    not work correctly without the typeset. Instead, x will be set
    to the string "x+1". When it gets to the test (( x < 10 )), it will
    substitute "x+1" for x, then again substitute "x+1" for that x until
    the recursion limit is reached (I think recursion is set to 200 levels).

    It can be a powerful feature, but is also a very dangerous one.

    Dan Mercer

    : do
    : echo $x
    : x=`expr $x + 1`
    : done
    :
    :
    : And here is the output:
    :
    : $ ./loop2
    : 1
    : 2
    : 3
    : 4
    : 5
    : 6
    : 7
    : 8
    : 9
    :
    : However on RH Linux using bash (/bin/sh on Linux)
    : it produces this error:
    :
    : $ ./loop
    : ./loop: line 4: [: x: integer expression expected
    :
    :
    : It appears to be failing evaulating this line:
    :
    : while [ x -lt 10 ]
    :
    : Is this some sort of known problem perhaps?
    : This is on Redhat Linux 8.0.
    :
    :
    : lk
    : www.theNewAgeSite.com

  • Next message: gregg: "Re: bash vs ksh problem"

    Relevant Pages

    • Re: [PATCH 2/2] tracing/events/lockdep: move tracepoints within recursive protection
      ... it allows one event to trace within another event. ... But surely not in the same context. ... You could do a 4 level recursion ... Why not allow a nested interrupt to trace? ...
      (Linux-Kernel)
    • Re: Why do lisps have stack limits?
      ... have as its implication that stack limits were bad, ... reasons for wanting deep recursion, and for this one can just write a ... pre-allocate lots of stack (for other implementations). ... context of resource limitations. ...
      (comp.lang.lisp)
    • Re: [PATCH 2/2] tracing/events/lockdep: move tracepoints within recursive protection
      ... it allows one event to trace within another event. ... But surely not in the same context. ... You could do a 4 level recursion ... protection like I did in perf-counter, ...
      (Linux-Kernel)
    • Re: Scope Question
      ... The number one solution is to split the recursion into a non-recursive ... ;; can freely refer to context ... set up the binding is to split the function into the interface wrapper ... (defun foo-recursion-helper (input) ...
      (comp.lang.lisp)
    • Re: Grammatical Recursion
      ... > iteration in the context of specific computation models. ... "formalism" in instead of "model" to avoid the confusion. ... Recursion and iteration are different types of ...
      (sci.lang)