Re: Patch: missing --run makeinfo

[Tom Tromey]

Alexandre> What's wrong with `(makeinfo --version > /dev/null 2>&1)'?
Alexandre> I'm not sure running sh explicitly is a good idea.

I used to do this because, I think, of a misreading of the Standards
(look for the install-info code).  Now I suspect that the Standards
writer simply didn't consider using a subshell.

I agree with you and I am checking this in:

2001-07-04  Tom Tromey  <address@hidden>

        * missing: Updated.
        * lib/missing (makeinfo): Use subshell, not explicit sh
        invocation.

Tom

Index: lib/missing
===================================================================
RCS file: /cvs/automake/automake/lib/missing,v
retrieving revision 1.15
diff -u -r1.15 missing
--- lib/missing 2001/07/02 00:48:12 1.15
+++ lib/missing 2001/07/04 15:29:07
@@ -209,7 +209,7 @@
     ;;
 
   makeinfo)
-    if test -z "$run" && sh -c 'makeinfo --version > /dev/null 2>&1'; then
+    if test -z "$run" && (makeinfo --version > /dev/null 2>&1); then
        # We have makeinfo, but it failed.
        exit 1
     fi