пт, 5 мар. 2021 г. в 18:14, Theo Buehler <t...@theobuehler.org>: > > On Sun, Feb 21, 2021 at 10:04:07PM +0300, Vadim Zhukov wrote: > > Hello all. > > > > This continues the 'Another potential ksh bug?' thread on misc@: > > https://marc.info/?l=openbsd-misc&m=160736700220621&w=2 > > This present is a bit too late for Christmas, but at least the Day of > > Red Army is coming soon. I'm sure noone is against this patch then. > > > > The code in typeset() function, which is responsible for almost all > > shell variable tweaking, contains a bug: in case of array it passes > > "foo=var" instead of only variable name ("foo"), if the value was > > ever given. This results in, e.g., a bug reported by Jordan Geoghegan. > > Generally speaking, we had creating a (temporary) variable via global() > > in vpbase, and of course it didn't get the read-only marker. > > > > This fix is to use 'tvar' variable, which always contains shell > > variable name, without value. > > ok tb > > > As a bonus, this fixes the ifs.t:IFS-null-1 test (previously failing). > > I'm too lazy to check why. :-) > > I'm confused. I have 11 failing tests before and after your patch. > Also IFS-null-1 has been passing for a long time.
Hm-m-m. Probably something in my environment... I'll experiment separately then. Thanks!
