[RFA] toplevel: Always use in tree makeinfo in texinfo subdir

[RFA] toplevel: Always use in tree makeinfo in texinfo subdir

[Corinna Vinschen]

Hi,

the MAKEINFO stuff in Makefile.tpl first tests, if an in-tree makeinfo
binary already exists.  If yes, it uses it, otherwise it tests the
installed makeinfo on the system, if the version is sufficent and
depending on the result, it uses the installed makeinfo or it calls
"missing".

However, this check has a leak when the texinfo directory itself gets
built.  It creates a makeinfo binary but it doesn't use it since at
the time the MAKEINFO test is done, the binary doesn't exist yet.
That doesn't make much sense since the makeinfo binary is build before
it gets used in texinfo itself.

So the idea of the below patch is to always use the self-built makeinfo
binary when building in the texinfo subdir.


Would that be ok with everyone?  If so, I'll check it into sourceware.

Corinna


	* Makefile.tpl (USUAL_MAKEINFO): Always use in-tree makeinfo when
	building in the texinfo directory.
	* Makefile.in: Regenerate.


Index: Makefile.tpl
===================================================================
RCS file: /cvs/src/src/Makefile.tpl,v
retrieving revision 1.105
diff -u -p -r1.105 Makefile.tpl
--- Makefile.tpl	9 Jun 2004 08:32:33 -0000	1.105
+++ Makefile.tpl	18 Jun 2004 11:19:17 -0000
@@ -273,7 +273,7 @@ M4 = `if [ -f $$r/m4/m4 ] ; \
 # For an installed makeinfo, we require it to be from texinfo 4.2 or
 # higher, else we use the "missing" dummy.
 MAKEINFO=@MAKEINFO@
-USUAL_MAKEINFO = `if [ -f $$r/texinfo/makeinfo/makeinfo ] ; \
+USUAL_MAKEINFO = `if [ -f $$r/texinfo/makeinfo/makeinfo ] || pwd | grep /texinfo$ > /dev/null ; \
 	then echo $$r/texinfo/makeinfo/makeinfo ; \
 	else if (makeinfo --version \
 	  | egrep 'texinfo[^0-9]*([1-3][0-9]|4\.[2-9]|[5-9])') >/dev/null 2>&1; \
Index: Makefile.in
===================================================================
RCS file: /cvs/src/src/Makefile.in,v
retrieving revision 1.175
diff -u -p -r1.175 Makefile.in
--- Makefile.in	9 Jun 2004 08:32:32 -0000	1.175
+++ Makefile.in	18 Jun 2004 11:19:18 -0000
@@ -270,7 +270,7 @@ M4 = `if [ -f $$r/m4/m4 ] ; \
 # For an installed makeinfo, we require it to be from texinfo 4.2 or
 # higher, else we use the "missing" dummy.
 MAKEINFO=@MAKEINFO@
-USUAL_MAKEINFO = `if [ -f $$r/texinfo/makeinfo/makeinfo ] ; \
+USUAL_MAKEINFO = `if [ -f $$r/texinfo/makeinfo/makeinfo ] || pwd | grep /texinfo$ > /dev/null ; \
 	then echo $$r/texinfo/makeinfo/makeinfo ; \
 	else if (makeinfo --version \
 	  | egrep 'texinfo[^0-9]*([1-3][0-9]|4\.[2-9]|[5-9])') >/dev/null 2>&1; \

-- 
Corinna Vinschen
Cygwin Co-Project Leader
Red Hat, Inc.

Re: [RFA] toplevel: Always use in tree makeinfo in texinfo subdir

[Paolo Bonzini]

> +USUAL_MAKEINFO = `if [ -f $$r/texinfo/makeinfo/makeinfo ] || pwd | grep /texinfo$ > /dev/null ; \

You should at least quote '/texinfo$'.

You may also look into supporting extra_configure_flags and 
extra_make_flags Autogen variables for modules.  Then just define them 
to USUAL_MAKEINFO="$$r/texinfo/makeinfo/makeinfo".

Maybe this will allow major cleanups WRT the way gcc is handled in the 
toplevel Makefile, so it is worth considering it.

Paolo

Re: [RFA] toplevel: Always use in tree makeinfo in texinfo subdir

[Alexandre Oliva]

On Jun 22, 2004, Paolo Bonzini <bonzini@gnu.org> wrote:

> +host_modules= { module= texinfo; no_install= true; bootstrap=true;
> +		extra_make_flags='USUAL_MAKEINFO="$$r/texinfo/makeinfo/makeinfo"'; };

This doesn't look right to me.  We need to pass MAKEINFO, not
USUAL_MAKEINFO down, and we can only do so when build and host are the
same.

[thinks some more]

Oh, I get it!  When build and host are the same, MAKEINFO expands to
USUAL_MAKEINFO, so we get what we want, without additional ugliness.
Neat!

Patch is ok.  Please check it in.

-- 
Alexandre Oliva             http://www.ic.unicamp.br/~oliva/
Red Hat Compiler Engineer   aoliva@{redhat.com, gcc.gnu.org}
Free Software Evangelist  oliva@{lsd.ic.unicamp.br, gnu.org}

Re: [RFA] toplevel: Always use in tree makeinfo in texinfo subdir

[Paolo Bonzini]

No, I don't like it anymore.  Looks unmaintainable.

I've checked in the other bits (Makefile.tpl and a regenerated 
Makefile.in without the Makefile.def patch), I'll see for a better solution.

Paolo