cc: [email protected]
Subject: Re: [ast-developers] ksh: problem with arrays
--------
> Hi,
>
> one our user reported change in ksh's behavior. I've checked it and prepared
> thi
> s simplified test case. Could you look at it? Thanks, Michal
>
> Simplified test case:
>
> typeset xx[7];
> echo ${#xx[@]};
> typeset -p xx;
> xx[6]=hello; echo ${xx[6]};
> typeset -p xx;
> xx[5]=world; echo ${xx[5]};
> typeset -p xx;
> xx[8]=foo; typeset -p xx;
> unset xx; typeset -p xx;
> echo ${#xx[@]}
>
> ksh >=2010-10-10 prints:
> 7 #why is that 7? bash,mksh, ksh<=2008-10-01 prints 0, ksh 2008-10-09 ...
> 2010-0
> 9-24 prints 1
> typeset -a xx[7];xx=()
> <nothing> #should print hello
> typeset -a xx[7];xx=('' '' '' '' '' '' '') #definition unexpectedly changed
> and
> 'hello' is missing
> <nothing> #should print world
> typeset -a xx[7];xx=('' '' '' '' '' '' '')
> xx[5]: subscript out of range #why out of range? why xx[5] , this was
> xx[8]=foo
> assignment
> <nothing>
> 0
>
>
> ksh 2008-10-09 ... 2010-09-24 prints:
> 1
> typeset -a xx=([7]=)
> hello
> typeset -a xx=([6]=hello [7]=)
> world
> typeset -a xx=([5]=world [6]=hello [7]= [8]=foo)
> 0
>
>
> ksh <=2008-10-01 prints:
> 0
> typeset -a xx=([0]=)
> hello
> typeset -a xx=([6]=hello)
> world
> typeset -a xx=([5]=world [6]=hello [8]=foo)
> 0
>
>
> bash output for comparison:
> 0
> declare -a xx='()'
> hello
> declare -a xx='([6]="hello")'
> world
> declare -a xx='([5]="world" [6]="hello" [8]="foo")'
> bash: line 0: typeset: xx: not found
> 0
>
This bug was already reported and is fixed for the first bug fix release of
ksh93u which will be distributed this quarter.
The problem is that the declaration
typeset xx[7];
is incorrectly treating xx as a fixed size array of size 7 and currently
fixed sized arrays only for for fixed sized objects. The declaration for
fixed size arrays should be
typeset -a xx[7];
and the declaration
typeset xx[7];
is not even necessary, but should not have been treated as a fixed array.
Therefore the workaround is to delete this declaration
or use the declaration
typeset -a xx
instead. This is also the cause of the core dump you reported with
typeset -p.
Here is a patch that should fix this problem.
========================cut here==========================
--- old/bltins/typeset.c Thu Dec 30 19:15:57 2010
+++ new/bltins/typeset.c Thu Mar 3 11:02:49 2011
@@ -568,7 +568,7 @@
path_alias(np,path_absolute(shp,nv_name(np),NIL(Pathcomp_t*)));
continue;
}
- np =
nv_open(name,troot,nvflags|((nvflags&NV_ASSIGN)?0:NV_ARRAY)|NV_FARRAY);
+ np =
nv_open(name,troot,nvflags|((nvflags&NV_ASSIGN)?0:NV_ARRAY)|(iarray?NV_FARRAY:0));
if(nv_isnull(np) && !nv_isarray(np) &&
nv_isattr(np,NV_NOFREE))
nv_offattr(np,NV_NOFREE);
if(tp->pflag)
========================cut here==========================
David Korn
[email protected]
_______________________________________________
ast-developers mailing list
[email protected]
https://mailman.research.att.com/mailman/listinfo/ast-developers
