Re: Mainline: make[2]: ada/doctools/xgnatug: Command not found

[Kelley Cook <kelleycook at yahoo dot com>]

> Am I the only one seeing this (powerpc-unknown-linux-gnu) on
mainline:
>
> make[3]: Leaving directory `/home/toon/compilers/snapshots/obj/gcc'
> echo timestamp > stmp-multilib
> ada/doctools/xgnatug vms ../../gcc/gcc/ada/gnat_ug.texi \
>            ../../gcc/gcc/ada/ug_words \
>    ../../gcc/gcc/ada/gnat_ug_vms.texi
> make[2]: ada/doctools/xgnatug: Command not found
> make[2]: *** [../../gcc/gcc/ada/gnat_ug_vms.texi] Error 127
> make[2]: Leaving directory `/home/toon/compilers/snapshots/obj/gcc'
> make[1]: *** [stage1_build] Error 2
> make[1]: Leaving directory `/home/toon/compilers/snapshots/obj/gcc'
> make: *** [bootstrap] Error 2

Hi,

First of all I apologize for the delay.  I moved over the weekend and
apparently the person in charge at my cable ISP was very efficient and
turned off my email address.  So it appears that I'll be using this
temporary email account for a few weeks until my new ISP is up and
going.

Second of all I also apologize for causing build problems for people.

However, looking at the Makefile rules, I still don't exactly see how
anyone is seeing this error.  After many tries, I was only able to
recreate it if I touched $(srcdir)/stamp-xgnatug between builds.  
 
Otherwise since the four pre-generated .texi files still have
timestamps later than stamp-xgnatug they should not have required
themselves to be a rebuilt which means that xgnatug isn't required to
have been built at all (which the impetus for the patch).

That being said, I do see a potential problem for those that do not use
gcc_update to checkout their sources (or at least gcc_update --touch
after they are checked out).  The CVS timestamp for stamp-xgnatug is
later than the CVS timestamp for the four .texi files, which could
certainly explain the behaviour that people are seeing.

I would be interested in knowing if one were to run "contrib/gcc_update
--touch" from the source directory does that permantly fixes the
problem.  If not it might be useful to know the version of make you are
using as well as if you are attempting parallel builds.

The following patch would fix that aforementioned problem with the CVS
timestamps.  I am hopeful that will fix the problems that people are
seeing.  Otherwise reverting the original patch would be prudent.

Regardless, I would think that this patch should be applied since
gnat_ug.texi was updated in 2003.  I'll commit it as obvious tonight,
unless someone wishes to do it before then.  (hopefully yahoo mail
doesn't mung it)

2003-12-23  Kelley Cook  <kcook@gcc.gnu.org>

	* gnat_ug.texi: Force a CVS commit by updating copyright.
	* gnat_ug_vxw.texi: Regenerate.
	* gnat_ug_wnt.texi: Regenerate.
	* gnat_ug_vms.texi: Regenerate.
	* gnat_ug_unx.texi: Regenerate.
	
--- gnat_ug.texi.orig	2003-12-23 10:30:45.498318000 -0500
+++ gnat_ug.texi	2003-12-23 10:31:13.779266000 -0500
@@ -79,7 +79,7 @@
 @c %**end of header
 
 @copying
-Copyright @copyright{} 1995-2002, Free Software Foundation
+Copyright @copyright{} 1995-2003, Free Software Foundation
 
 Permission is granted to copy, distribute and/or modify this document
 under the terms of the GNU Free Documentation License, Version 1.2



__________________________________
Do you Yahoo!?
New Yahoo! Photos - easier uploading and sharing.
http://photos.yahoo.com/