David, as clarification: The poll(1) builtin is acting up for types. I think Roland's design anticipates that it can not create new variables in a compound variable which is embedded in a type. However ksh even acts up if some one wants to *read* a compound variable member which does not exist.
For example: ksh -c 'typeset -T y_t=( compound ex=(i=1)) ; y_t v ; print "${v.ex.nono}"' ./arch/linux.i386-64/bin/ksh: ex.nono: is not an element of v In poll(1) I get the same warning, despite the use of NV_NOFAIL|NV_NOADD flags in the nv_open() call. I think the nv_open() call should just fail silently, right? Olga On Sat, Aug 18, 2012 at 2:31 AM, ольга крыжановская <olga.kryzhanov...@gmail.com> wrote: > David, why do I get a 'x.sh: line 8: extensions.i: is not an element > of x' for the script below? > > typeset -T y_t=( > compound extensions > ) > y_t x; > x.extensions=( integer i=1 ) > print "$x" > > This is rather a *big* problem since I assumed that compound variables > are always extensible, and if I add an array of compound variables I > can both add array elements and a variable number of data in these > compound variables in the array. > > Olga > -- > , _ _ , > { \/`o;====- Olga Kryzhanovska -====;o`\/ } > .----'-/`-/ olga.kryzhanov...@gmail.com \-`\-'----. > `'-..-| / http://twitter.com/fleyta \ |-..-'` > /\/\ Solaris/BSD//C/C++ programmer /\/\ > `--` `--` -- , _ _ , { \/`o;====- Olga Kryzhanovska -====;o`\/ } .----'-/`-/ olga.kryzhanov...@gmail.com \-`\-'----. `'-..-| / http://twitter.com/fleyta \ |-..-'` /\/\ Solaris/BSD//C/C++ programmer /\/\ `--` `--` _______________________________________________ ast-developers mailing list ast-developers@research.att.com https://mailman.research.att.com/mailman/listinfo/ast-developers