Misc notes
----------

regexprops-generic.texi is generated via a utility in findutils.

How to update gnulib manual on www.gnu.org
------------------------------------------

1) You need a non-anonymous checkout of the web pages directory.

   $ cvs -d :ext:[email protected]:/web/gnulib \
         checkout gnulib

2) Get familiar with the instructions for web page maintainers.
   https://www.gnu.org/server/standards/gnu-website-guidelines.html
   especially the note about symlinks.
   Some of this page may also be relevant:
   https://www.gnu.org/server/standards/README.webmastering.html

3) Assuming GNULIB_CHECKOUT refers to a checkout of the gnulib dir,
   and      GNULIB_WWW_CHECKOUT refers to the other directory created above (1),
   do

   GNULIB_WWW_CHECKOUT=`cd $GNULIB_WWW_CHECKOUT && pwd`
   cd $GNULIB_WWW_CHECKOUT
   cvs update -dP
   tar -cf html-CVS.tar manual/html_node/CVS manual/html_node/.symlinks
   cd $GNULIB_CHECKOUT
   git checkout master
   git pull
   cd doc
   make undocumented-modules.texi updated-stamp
   ../build-aux/gendocs.sh \
     -I ../build-aux \
     --email [email protected] \
     -o "$GNULIB_WWW_CHECKOUT/manual" \
     gnulib "The GNU Portability Library"
   cd $GNULIB_WWW_CHECKOUT
   tar -xf html-CVS.tar
   rm html-CVS.tar
   newfiles=`cvs -n update -dP | sed -n -e 's|^[?] manual/html_node/|manual/html_node/|p'`
   if test -n "$newfiles"; then
     cvs add $newfiles
   fi

   Run 'cvs diff -u' and verify that the output looks sane.

4) Commit the modified and the new files.

5) Find the files which have not been overwritten (because they belonged
   to sections that have been removed or renamed):

   $ cd manual/html_node
   $ ls -lt

   Remove these files and commit their removal to CVS.
   For each of these files, add a line to the file .symlinks. This will
   ensure that hyperlinks to the removed files will redirect to the entire
   manual; this is better than a 404 error.

There is a problem with 'index.html' being written twice (once for POSIX
function 'index', once for the table of contents); you can ignore this issue.

Note: For updating the www.gnu.org pages of *other* GNU packages, package
specific scripts can be used, or build-aux/gnu-web-doc-update may possibly
work.