Re: [patch] New test+use texinfo @click - @HAVE_MAKEINFO_CLICK@ [Re: doc build failure (Re: [patch 04/12] entryval#3: Virtual tail call frames)]

[Joel Brobecker <brobecker@adacore.com>]

Hi Jan,

> gdb/
> 2011-10-10  Jan Kratochvil  <jan.kratochvil@redhat.com>
> 
> 	Fix compatibility with texinfo versions older than 4.12.
> 	* Makefile.in (MAKEINFO): Set to @MAKEINFO@.
> 	(MAKEINFOFLAGS, MAKEINFO_EXTRA_FLAGS, MAKEINFO_CMD): New.
> 	(MAKEHTMLFLAGS): Use MAKEINFO_CMD.
> 	(FLAGS_TO_PASS): Add MAKEINFOFLAGS and MAKEINFO_EXTRA_FLAGS.
> 	* configure: Regenerate.
> 	* configure.ac (MAKEINFO): Find it, from libiberty/configure.ac.
> 	(MAKEINFOFLAGS): Pre-set it to --split-size=5000000.
> 	(MAKEINFO_EXTRA_FLAGS): New test for -DHAVE_MAKEINFO_CLICK.

There is one bit that I don't understand:

> --- a/gdb/configure.ac
> +++ b/gdb/configure.ac
> @@ -87,6 +87,38 @@ PACKAGE=gdb
>  AC_DEFINE_UNQUOTED(PACKAGE, "$PACKAGE", [Name of this package. ])
>  AC_SUBST(PACKAGE)
>  
> +# Do we have a single-tree copy of texinfo?  Even if we do, we can't
> +# rely on it - libiberty is built before texinfo.
> +AC_CHECK_PROG(MAKEINFO, makeinfo, makeinfo, )
> +if test "x$MAKEINFO" = "x"; then
> +  AC_ERROR([makeinfo is required for compilation])

Why do we error out if makeinfo is not available? I think we should do
the same as other projects such as binutils where we display a warning
("makeinfo is missing, documentation will not be built"), and then
continue. We used to be able to build without requiring makeinfo, and
installing makeinfo on all machines on our network would be a real pain.
For us, we know how to work around the problem (configure with
MAKEINFO=true), but less experience users might not.

Let me know what you think, and I'll see if I can whip up a patch.

-- 
Joel