On Sat, May 17, 2014 at 08:41:30AM +0200, Christian Couder wrote: > Signed-off-by: Christian Couder <chriscool@xxxxxxxxxxxxx> > --- > t/t6050-replace.sh | 29 +++++++++++++++++++++++++++++ > 1 file changed, 29 insertions(+) Yay, tests. > +test_expect_success 'setup a fake editor' ' > + cat >fakeeditor <<-\EOF && > + #!/bin/sh > + sed -e "s/A U Thor/A fake Thor/" "$1" >"$1.new" > + mv "$1.new" "$1" > + EOF > + chmod +x fakeeditor > +' Please use write_script, like: test_expect_success 'setup a fake editor' ' write_script fakeeditor <<-\EOF sed -e "s/A U Thor/A fake Thor/" "$1" >"$1.new" mv "$1.new" "$1" EOF ' which will use the $(SHELL_PATH) shebang. It doesn't matter for such a simple script here (which even a broken #!/bin/sh could manage), but in general, I think we're trying to set it consistently. You could also &&-chain the commands in your fakeeditor script, though I find it unlikely that sed will fail. > +test_expect_success '--edit with and without already replaced object' ' > + GIT_EDITOR=./fakeeditor test_must_fail git replace --edit "$PARA3" && Unfortunately it's not portable to do a one-shot environment variable on a shell function like this; some shells leave the variable set after the function returns. We've been using: test_must_fail env GIT_EDITOR=./fakeeditor git ... The regular GIT_EDITOR=./fakeeditor git ... ones are OK. -Peff -- To unsubscribe from this list: send the line "unsubscribe git" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html