On 7:01:33 pm 2005-07-12 "Paul D. Smith" <[EMAIL PROTECTED]> wrote: > %% Eli Zaretskii <[EMAIL PROTECTED]> writes: > > ez> Yes, I think testing unixy_shell would be okay. > > Good stuff; so if !unixy_shell we want to go back to the old model, > where make would eat the backslash-newline pair before invoking the > shell, right? > > What about the TAB on the next line? In POSIX if the first character > after a backslash-newline is a TAB, make will remove it from the > command script to be invoked. >
Would it remove the tab or change it to a space? > The other change between the old behavior and new behavior was that > make used to replace the backslash-newline with a space. So, if you > had this: > > echo foo\ > bar > > would print "foo bar". In POSIX, no whitespace is added so this > would print "foobar". > > I like the POSIX way better but it's not backward-compatible so it's > up to you guys. > As long as tab is converted to space then not converting \\n to a space would be ok. So: echo foo\ bar prints "foo bar" while echo foo\ bar prints "foobar" Earnie _______________________________________________ Make-w32 mailing list Make-w32@gnu.org http://lists.gnu.org/mailman/listinfo/make-w32