build: doc: Add a man page post-processor to build_man.sh

- If there is a "Modules" section, delete it
- If "Detailed Description" is empty, delete "Detailed Description" line
- Reposition SYNOPSIS (with headers that we inserted) to start of page,
  integrating with defined functions to look like other man pages
- Delete all "Definition at line nnn" lines
- Delete lines that make older versions of man o/p an unwanted blank line
For better readability, shell function definitions are separated by blank
lines, and there is a bit of annotation.

Signed-off-by: Duncan Roe <duncan_roe@optusnet.com.au>
Signed-off-by: Pablo Neira Ayuso <pablo@netfilter.org>

1 files changed, 195 insertions, 5 deletions

diff --git a/doxygen/build_man.sh b/doxygen/build_man.sh
index 304a305..e0cda71 100755
--- a/doxygen/build_man.sh
+++ b/doxygen/build_man.sh
@@ -1,20 +1,37 @@
 #!/bin/bash -p
-# We need to use bash for its associative array facility
+
+# Script to process man pages output by doxygen.
+# We need to use bash for its associative array facility.
 # (`bash -p` prevents import of functions from the environment).
+
 declare -A renamed_page
-main(){ set -e; cd man/man3; rm -f _*
+
+main(){
+  set -e
+  cd man/man3; rm -f _*
   count_real_pages
   rename_real_pages
   make_symlinks
+  post_process
 }
-count_real_pages(){ page_count=0
+
+count_real_pages(){
+  page_count=0
+  #
+  # Count "real" man pages (i.e. not generated by MAN_LINKS)
+  # MAN_LINKS pages are 1-liners starting .so
+  # Method: list files in descending order of size,
+  # looking for the first 1-liner
+  #
   for i in $(ls -S)
   do head -n1 $i | grep -E -q '^\.so' && break
     page_count=$(($page_count + 1))
   done
   first_link=$(($page_count + 1))
 }
-rename_real_pages(){ for i in $(ls -S | head -n$page_count)
+
+rename_real_pages(){
+  for i in $(ls -S | head -n$page_count)
   do for j in $(ls -S | tail -n+$first_link)
     do grep -E -q $i$ $j && break
     done
@@ -22,8 +39,181 @@ rename_real_pages(){ for i in $(ls -S | head -n$page_count)
     renamed_page[$i]=$j
   done
 }
-make_symlinks(){ for j in $(ls -S | tail -n+$first_link)
+
+make_symlinks(){
+  for j in $(ls -S | tail -n+$first_link)
   do ln -sf ${renamed_page[$(cat $j | cut -f2 -d/)]} $j
   done
 }
+
+post_process(){
+  make_temp_files
+  #
+  # DIAGNOSTIC / DEVELOPMENT CODE
+  # set -x and restrict processing to keep_me: un-comment to activate
+  # Change keep_me as required
+  #
+  #keep_me=nfq_icmp_get_hdr.3;\
+  #do_diagnostics;\
+  #
+  # Work through the "real" man pages
+  for target in $(ls -S | head -n$page_count)
+  do mygrep "^\\.SH \"Function Documentation" $target
+    # Next file if this isn't a function page
+    [ $linnum -ne 0 ] || continue
+
+    del_modules
+    del_bogus_synopsis
+    fix_name_line
+    move_synopsis
+    del_empty_det_desc
+    del_def_at_lines
+    fix_double_blanks
+  done
+
+  remove_temp_files
+}
+
+fix_double_blanks(){
+  linnum=1
+  #
+  # Older versions of man display a blank line on encountering "\fB\fP";
+  # newer versions of man do not.
+  # doxygen emits "\fB\fP" on seeing "\par" on a line by itself.
+  # "\par" gives us double-spacing in the web doc, which we want, but double-
+  # spacing looks odd in a man page so remove "\fB\fP".
+  #
+  while [ $linnum -ne 0 ]
+  do mygrep \\\\fB\\\\fP $target
+    [ $linnum -eq 0 ] || delete_lines $linnum $linnum
+  done
+}
+
+del_def_at_lines(){
+  linnum=1
+  while [ $linnum -ne 0 ]
+  do mygrep "^Definition at line [[:digit:]]* of file" $target
+    [ $linnum -eq 0 ] || delete_lines $(($linnum - 1)) $linnum
+  done
+}
+
+# Only invoked if you un-comment the 2 diagnostic / development lines above
+do_diagnostics(){
+  mv $keep_me xxx
+  rm *.3
+  mv xxx $keep_me
+  page_count=1
+  set -x
+}
+
+del_empty_det_desc(){
+  mygrep "^\\.SH \"Function Documentation" $target
+  i=$linnum
+  mygrep "^\\.SH \"Detailed Description" $target
+  [ $linnum -ne 0  ] || return 0
+  [ $(($i - $linnum)) -eq 3 ] || return 0
+  delete_lines $linnum $(($i -1))
+}
+
+move_synopsis(){
+  mygrep "SH SYNOPSIS" $target
+  [ $linnum -ne 0  ] || return 0
+  i=$linnum
+  # If this is a doxygen-created synopsis, leave it.
+  # (We haven't inserted our own one in the source yet)
+  mygrep "^\\.SS \"Functions" $target
+  [ $i -gt $linnum ] || return 0
+
+  mygrep "^\\.SH \"Function Documentation" $target
+  j=$(($linnum - 1))
+  head -n$(($j - 1)) $target | tail -n$(($linnum - $i - 1)) >$fileC
+  delete_lines $i $j
+  mygrep "^\\.SS \"Functions" $target
+  head -n$(($linnum - 1)) $target >$fileA
+  tail -n+$(($linnum + 1)) $target >$fileB
+  cat $fileA $fileC $fileB >$target
+}
+
+fix_name_line(){
+  all_funcs=""
+
+  # Search a shortened version of the page in case there are .RI lines later
+  mygrep "^\\.SH \"Function Documentation" $target
+  head -n$linnum $target >$fileC
+
+  while :
+  do mygrep ^\\.RI $fileC
+    [ $linnum -ne 0 ] || break
+    # Discard this entry
+    tail -n+$(($linnum + 1)) $fileC >$fileB
+    cp $fileB $fileC
+
+    func=$(cat $fileG | cut -f2 -d\\ | cut -c3-)
+    [ -z "$all_funcs" ] && all_funcs=$func ||\
+      all_funcs="$all_funcs, $func"
+  done
+  # For now, assume name is at line 5
+  head -n4 $target >$fileA
+  desc=$(head -n5 $target | tail -n1 | cut -f3- -d" ")
+  tail -n+6 $target >$fileB
+  cat $fileA >$target
+  echo "$all_funcs \\- $desc" >>$target
+  cat $fileB >>$target
+}
+
+del_modules(){
+  mygrep "^\.SS \"Modules" $target
+  [ $linnum -ne 0  ] || return 0
+  i=$linnum
+  mygrep "^\\.SS \"Functions" $target
+  delete_lines $i $(($linnum - 1))
+}
+
+del_bogus_synopsis(){
+  mygrep "SH SYNOPSIS" $target
+  #
+  # doxygen 1.8.20 inserts its own SYNOPSIS line but there is no mention
+  # in the documentation or git log what to do with it.
+  # So get rid of it
+  #
+  [ $linnum -ne 0  ] || return 0
+  i=$linnum
+  # Look for the next one
+  tail -n+$(($i + 1)) $target >$fileC;\
+  mygrep "SH SYNOPSIS" $fileC
+  [ $linnum -ne 0  ] || return 0
+
+  mygrep "^\\.SS \"Functions" $target
+  delete_lines $i $(($linnum - 1))
+}
+
+# Delete lines $1 through $2 from $target
+delete_lines(){
+  head -n$(($1 - 1)) $target >$fileA
+  tail -n+$(($2 +1)) $target >$fileB
+  cat $fileA $fileB >$target
+}
+
+mygrep(){
+  set +e
+  grep -En "$1" $2 2>/dev/null >$fileH
+  [ $? -ne 0 ] && linnum=0 ||\
+    { head -n1 $fileH >$fileG; linnum=$(cat $fileG | cut -f1 -d:); }
+  set -e
+}
+
+make_temp_files(){
+  temps="A B C G H"
+  for i in $temps
+  do declare -g file$i=$(mktemp)
+  done
+}
+
+remove_temp_files(){
+  for i in $temps
+  do j=file$i
+    rm ${!j}
+  done
+}
+
 main