Roland Mainz wrote: > Glenn Fowler wrote: > > On Wed, 28 Jun 2006 00:59:24 +0200 Roland Mainz wrote: > > > I just did some experiments with the ksh93 test suite (on Solaris > > > Nevade/B37/i386) ... > > > > > ## ./src/cmd/ksh93/tests/options.sh > > > ./src/cmd/ksh93/tests/options.sh[118]: ./-ksh ignores .profile > > > > were the tests run with uid!=euid || gid!=egid || uid==0? > > AFAIK no. /usr/bin/id reports this: > -- snip -- > % id > uid=1000(test001) gid=100(users) > -- snip -- > ... and I just launched the tests via "ksh name_of_test.sh", e.g. > -- snip -- > % (cd /home/test001/ksh93/ksh93rplus_alpha2/build/ ; for i in > ./src/cmd/ksh93/tests/*.sh ; do echo "## $i" ; ./arch/sol11.i386/bin/ksh > "$i" ; done) > -- snip --
It seems the ksh93r+_alpha_20060630 fixed this together with my adjustment of the way how I call the test. I now use "env -" explicitly to clear the environment: -- snip -- % for i in $PWD/src/cmd/ksh93/tests/*.sh ; do rm -f core ; echo "## $i" ; export SHELL=$PWD/arch/sol10.sun4/bin/ok/ksh ; echo "## $SHELL $i" ; env - SHELL=$SHELL HOME=$HOME PATH=$PATH LANG=C $SHELL "$i" ; done -- snip -- Now all tests run without problems in the "C" locale (there is a glitch in the "en_US.UTF-8" locale vs. "options.sh", but this is IMO an unrelated issue and I'll file a seperate email about that... ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) roland.mainz at nrubsig.org \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 7950090 (;O/ \/ \O;)
