new expected behavior? src/bin/rm/rm.c

[Matthew Dillon]

:I don't want to specify -ff every time I remove something from a default
:install.
:
:The value, I thought, of -I was that it doesn't ask about *every* file
:vs 'rm -ir'; if you use -I as a shell alias, what's the problem with
:overriding it with -f? Habitual "-rf" entries are not a dumb user trait,
:it's an experienced admin trait. If people learn with -I they'll never
:start accidently using -f, they may only use it in scripts.
:
:The real benefit of -I is that it means we don't need to specify -f all
:the time, so what's the problem of -f overriding -i _and_ -I, when it is
:used? -- in which case a background test is not really necessary (but
:may be a good idea anyway).
:
:// George
:
:-- 
:George Georgalis, systems architect, administrator Linux BSD IXOYE

    I was indeed thinking about that.  I always use 'rm -rf' myself, even
    when the 'f' is not necessary.  Whether -I should operate in that case
    is a real question.

    I think -I should clearly not operate in the backgrounded case.  Using
    the same 'dumb user' argument, no beginning user will ever background an
    rm :-).  In fact, most experienced sysads don't background rm's either
    unless they really, really mean it.  The idea is to be able to alias rm
    to rm -I and have the added safety feature without creating unexpected 
    operation.  I will make that change immediately.

    But I'm still not sure what we should do about interactive foreground
    operation.  This may seem a bit presumptuous, but I want -I to work
    for me!

							-Matt