On Tue, Feb 12, 2013 at 09:29:26AM -0800, Junio C Hamano wrote:
> John Keeping <john@xxxxxxxxxxxxx> writes:
>
> > ... the following fixup is also needed to avoid relying on the shell
> > emitting a literal backslash when a backslash isn't followed by a known
> > escape character.
> >
> > -- >8 --
> >
> > diff --git a/t/t3404-rebase-interactive.sh b/t/t3404-rebase-interactive.sh
> > index cbe36bf..84bd525 100755
> > --- a/t/t3404-rebase-interactive.sh
> > +++ b/t/t3404-rebase-interactive.sh
> > @@ -947,7 +947,7 @@ test_expect_success 'rebase -i respects core.commentchar' '
> > test_when_finished "git config --unset core.commentchar" &&
> > cat >comment-lines.sh <<EOF &&
> > #!$SHELL_PATH
> > -sed -e "2,\$ s/^/\\\\\\/" "\$1" >"\$1".tmp
> > +sed -e "2,\$ s/^/\\\\\\\\/" "\$1" >"\$1".tmp
> > mv "\$1".tmp "\$1"
> > EOF
> > chmod a+x comment-lines.sh &&
>
> Yeek. If you used write_script with here-text that does not
> interpolate,
>
> write_script remove-all-but-the-first.sh <<\EOF
> sed -e '2,$s/^/\\/' <"$1" >"$1.tmp" &&
> mv "$1.tmp" "$1"
> EOF
>
> the above would be much more readable.
Yet another thing for me to learn about ;-)
Do you mean to use that outside the test case, so that the single quotes
work? Or do I still need some level of escaping?
> I am not sure if I understand what you meant by "literal backslash
> blah blah", though.
It turns out that having this in the script works (in bash and dash
although I haven't checked what Posix has to say about it):
sed -e "2,$ s/^/\\\/"
and is equivalent to:
sed -e '2,$ s/^/\\/'
because backslashes that aren't recognised as part of an escape sequence
are not treated specially.
John
--
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