#! /bin/ksh
#
# file: sym2num
#
# purpose: 1. to create a header file containing the identifiers
#             and replacement list consisting of the symbols used
#             in a message set as well as the messages themselves.
#
#          2. to provide output to standard out which can be used
#             by gencat.
#

(( $# != 2 )) && { print "usage: sym2num <symbol name> <source file>" ;\
                 exit 1 ; }

SYMBOL_NAME=$1
CAP_SYMBOL_NAME=`echo $1 | tr '[:lower:]' '[:upper:]'`  # capitalized symbol
SOURCE_FILE=$2
inc_file=${SYMBOL_NAME}_msg.h     # include file

#### Step 1: Create a header file ####
# write first five lines to "<symbolname>_msg.h"
cat <<! > ${inc_file}
#ifndef _H_${CAP_SYMBOL_NAME}_MSG
#define _H_${CAP_SYMBOL_NAME}_MSG
#include <limits.h>
#include <nl_types.h>
#define MF_${CAP_SYMBOL_NAME} "${SYMBOL_NAME}.cat"



/* The following was generated from ${SOURCE_FILE}. */
!

# locate all lines which start with "$set", and then increment 
# the set counter and reset the message counter to zero.
awk '/^\$set/ { SET_NAME=$2; SET_COUNT++; MES_COUNT=0;  
       printf( "\n/* definitions for set %s */\n", SET_NAME ); 
       printf( "#define %s %s\n\n", SET_NAME, SET_COUNT) }   #to be continued

# find all lines which begin with a combination of letters, numbers
# and underscores followed by spaces and then a quotation mark,
# and increment the message counter. 
/^[0-9A-Za-z_]+[ |\t]*\"/ { MES_NAME=$1; MES_COUNT++; 
      printf( "#define %s %s\n", MES_NAME, MES_COUNT) }
      END {print "#define LAST_MSG_NO "MES_COUNT"\n\n#endif"}' \
       ${SOURCE_FILE}  >> ${inc_file}

#### Step 2: Create input to gencat ####       
# cat the definition statements of the include file and cat them to
# the source message file.  The awk file only creates define statements
# and blank lines.  The output should be sent to the preprocessor.
# There are two differences between sym2num and mkcatdefs: sym2num
# does not create a $delset line, and sym2num converts all symbols
# to numbers, even those in comment statements. 
if [[ -f /usr/ccs/lib/cpp ]]   #IBM & SUN
then
   cpp_path=/usr/ccs/lib       
elif [[ -f /usr/libexec/cpp ]] #BSD
then
   cpp_path=/usr/libexec
elif [[ -f /usr/bin/cpp ]]     #BSD
then
   cpp_path=/usr/bin
elif [[ -f /lib/cpp ]]         #HP
then
   cpp_path=/lib
fi
   ( cat ${inc_file} | sed -n /define/p ; cat ${SOURCE_FILE} ) | \
   ${cpp_path}/cpp -P | sed -e '/^$/d' -e 's/^\$$/\$ /' \
    -e 's/\"\"$/\"/' -e 's/XDQUOTE/\"/' -e "s/XSQUOTE/\'/" \
    -e 's/^\([1-9][0-9]*\)[ ]*[\"]*\(\"\)\(.*\)/\1 \2\3/'