On 5/6/19 10:40 AM, enh via Toybox wrote: > --- > tests/sed.test | 4 ++-- > 1 file changed, 2 insertions(+), 2 deletions(-) > > diff --git a/tests/sed.test b/tests/sed.test > index 6b27fff8..04603a21 100755 > --- a/tests/sed.test > +++ b/tests/sed.test > @@ -169,11 +169,11 @@ testing '-z' 'sed -z "s/\n/-/g"' "a-b-c" "" "a\nb\nc" > > # toybox handling of empty capturing groups broke minjail. Check that we > # correctly replace an empty capturing group with the empty string: > -testing '\n with empty capture' \ > +testing '\N with empty capture' \ > 'sed -E "s/(ARM_)?(NR_)([a-z]*) (.*)/\1\2\3/"' "NR_read" "" "NR_read foo" > # ...but also that we report an error for a backreference to a group that > # isn't in the pattern: > -testing '\n too high' \ > +testing '\N too high' \ > 'sed -E "s/(.*)/\2/p" 2>/dev/null || echo OK' "OK\n" "" "foo" > > # -i with $ last line test
Um, I'm not seeing these as newlines? Hmmm... scripts/runtest.sh function testing()... NAME="$CMDNAME $1" echo "$SHOWPASS: $NAME" So your echo is defaulting to -e (and debian's isn't), which is why you needed -E. Except we _just_ added -E to toybox and I'm not quite comfortable trusting it to be there on the host yet... Ah, printf is in aosp/prebuilts/build-tools/linux-x86/bin/toybox (which took 14 hours to download through my cable modem, but is now on my laptop). Does switching to that fix it? Rob _______________________________________________ Toybox mailing list Toybox@lists.landley.net http://lists.landley.net/listinfo.cgi/toybox-landley.net