This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: Documentation generation patch [Take 2]
- To: Theodore Papadopoulo <Theodore dot Papadopoulo at sophia dot inria dot fr>
- Subject: Re: Documentation generation patch [Take 2]
- From: "Joseph S. Myers" <jsm28 at cam dot ac dot uk>
- Date: Wed, 9 May 2001 19:01:33 +0100 (BST)
- cc: <gcc-patches at gcc dot gnu dot org>
On Wed, 9 May 2001, Theodore Papadopoulo wrote:
> + c-tree.dvi: $(srcdir)/c-tree.texi
c-tree.dvi now needs to depend on fdl.texi as well.
> + TEXINPUTS=${texidir}:$(srcdir):$$TEXINPUTS tex c-tree.texi
> + texindex c-tree.??
> + TEXINPUTS=${texidir}:$(srcdir):$$TEXINPUTS tex c-tree.texi
Could you investigate whether, in all places where GCC currently uses this
idiom, texi2dvi could be used instead (with -I $(srcdir) to find the
files)? The following comment appears in f/Make-lang.in, but I wonder
whether with current (4.0 or greater) texinfo, and the -I option used as
the official way to find input files, it is necessary, or whether texi2dvi
could be used throughout.
# `tex2dvi' was used below, but the Texinfo 3.12 one won't work properly
# with the include files from $(srcdir). This use of TEXINPUTS may not
# be universally valid. `$(TEX)' should be used if it gets defined in
# gcc/Makefile.in.
I don't see any problem with this patch once the missing dependency is
added, though I'm not qualified to approve it. But I think that changing
to use texi2dvi would be a useful followup patch, if it works. (Of
course, use $(TEXI2DVI) rather than hardcoded texi2dvi.)
--
Joseph S. Myers
jsm28@cam.ac.uk