fixed bug preventing "pair_coeff" commands from being understood corr…

…ectly if they appear outside of the "system.in.settings" file (ie. outside the "In Settings" section). (This is the place where those commands usually appear.) Now "pair_coeff" commands can appear in any file created by the user, and they will be post-processed correctly (even if the @atom-variable-names contain wildcards).

moltemplate/scripts/moltemplate.sh

@@ -6,8 +6,8 @@
 # Copyright (c) 2013
 
 G_PROGRAM_NAME="moltemplate.sh"
-G_VERSION="2.18.5"
-G_DATE="2020-8-06"
+G_VERSION="2.18.6"
+G_DATE="2020-8-13"
 
 echo "${G_PROGRAM_NAME} v${G_VERSION} ${G_DATE}" >&2
 echo "" >&2
@@ -1355,6 +1355,7 @@ IFS=$OIFS
 # "pair_coeff", "bond_coeff", "angle_coeff", "dihedral_coeff", "improper_coeff".
 # These files need to be processed further.
 
+
 OUT_FILES_WITH_COEFF_COMMANDS=""
 for f in *.template; do
     HAS_COEFF_COMMANDS=`awk 'BEGIN{has_coeffs=0} {if ((NF>0)&&(($1=="pair_coeff")||($1=="bond_coeff")||($1=="angle_coeff")||($1=="dihedral_coeff")||($1=="improper_coeff"))) has_coeffs=1} END{print has_coeffs}' < "$f"`
@@ -1368,7 +1369,6 @@ for f in *.template; do
 done
 
 
-
 # Deal with wildcard characters ('*', '?') in "_coeff" commands
 # appearing in any LAMMPS input scripts generated by moltemplate.
 # Replace them with explicit variable names.  Do this before rendering
@@ -2248,6 +2248,37 @@ else
 fi
 
 
+# Later on, it will be useful to keep track of the files that the user created.
+# This must be inferred from the files ending in ".template".
+# (Do this now before we discard all of the ".template" files.)
+
+RENDERED_FILES=""
+for f in *.template; do
+
+    # Moltemplate (ttree.py) generates two copies of each file that the
+    # user creates using the "write" and "write_once" commands:
+    # 1) The original file requested by the user.  This file is a "rendered"
+    #    file.  All of the $-style and @-style variables have been replaced
+    #    with other text (usually integers).
+    # 2) A version of that file that contains the original variable names
+    #    instead of integers.  (These file names end in ".template".)
+    # We no longer care about the files ending in ".template", but we can loop
+    # over files ending in ".template" created by moltemplate.  We can infer
+    # the name of the other file name by removing the ".template" suffix.
+
+    bn=`basename "$f" .template`  #file name without the ".template" suffix
+    if [ ! -f "$bn" ]; then
+        continue
+    fi
+    if [ -n "$RENDERED_FILES" ]; then
+        printf -v RENDERED_FILES "${RENDERED_FILES}\n$bn"
+    else
+        RENDERED_FILES="$bn"
+    fi
+done
+
+
+
 # ############## CLEAN UP ################
 
 # A lot of files have been created along the way.
@@ -2359,17 +2390,72 @@ if [ -e "$in_prefix_no_space" ]; then
     mv -f "$in_prefix_no_space" output_ttree/
 fi
 
-
+# ------ Postprocess coeff commands after rendering. ----------
 # Swap the order of atom types I, J in all "pair_coeff I J ..." commands
 # whenever I > J.  Do this for every input script file generated by moltemplate.
 # (Perhaps later I'll check to make sure the user did not specify contradictory
 #  "pair_coeff I J ..." and "pair_coeff J I ..." commands, but I won't do it
 #  here, because at this point we've thrown away the original atom type names,
 #  so there's no easy way to explain the problem to the user if there is one.)
 
-echo "" > input_scripts_so_far.tmp
+# -- First we must figure out which files we need to consider. --
+# Ugly details.  We already created a variable ($RENDERED_FILES) storing a
+# list of files which are created by the user.  However some of those files no
+# longer exist. We also need do add several files to this list and remove
+# duplicates.  Lacking data structures, SH and BASH are really bad at this kind
+# thing, so the code is ugly.  Must create temporary files and sort them.
+
+echo "$RENDERED_FILES" > RENDERED_FILENAMES.tmp
+echo "$OUT_FILE_INIT" >> RENDERED_FILENAMES.tmp
+echo "$OUT_FILE_INPUT_SCRIPT " >> RENDERED_FILENAMES.tmp
+echo "$OUT_FILE_SETTINGS " >> RENDERED_FILENAMES.tmp
+echo "$RENDERED_FILES_WITH_COEFF_COMMANDS" >> RENDERED_FILENAMES.tmp
+
+# Now remove duplicate file names
+sort RENDERED_FILENAMES.tmp > RENDERED_FILENAMES.tmp2
+awk '{if ((NF>0)&&(NR>1)&&($0!=prev)) {print $0} prev=$0}' \
+    < RENDERED_FILENAMES.tmp2 > RENDERED_FILENAMES.tmp3
+
+# Now update the list of files stored in $RENDERED_FILES
+RENDERED_FILES=""
+while read file_name; do
+    if [ ! -f "$file_name" ]; then
+        continue  # make sure the file exists
+    fi
+    if [ -n "$RENDERED_FILES" ]; then
+        printf -v RENDERED_FILES "${RENDERED_FILES}\n$file_name"
+    else
+        RENDERED_FILES="$file_name"
+    fi
+done < RENDERED_FILENAMES.tmp3
+
+rm -f RENDERED_FILENAMES.tmp*
+
+
+# Now filter out files which do not contain "_coeff" commands
+RENDERED_FILES_WITH_COEFF_COMMANDS=""
+IFS=$CR
+for file_name in $RENDERED_FILES; do
+    if [ ! -f "$file_name" ]; then
+        continue  # make sure the file exists
+    fi
+    HAS_COEFF_COMMANDS=`awk 'BEGIN{has_coeffs=0} {if ((NF>0)&&(($1=="pair_coeff")||($1=="bond_coeff")||($1=="angle_coeff")||($1=="dihedral_coeff")||($1=="improper_coeff"))) has_coeffs=1} END{print has_coeffs}' < "$file_name"`
+    if [ "$HAS_COEFF_COMMANDS" -eq "1" ]; then
+        if [ -n "$RENDERED_FILES_WITH_COEFF_COMMANDS" ]; then
+            printf -v RENDERED_FILES_WITH_COEFF_COMMANDS "${RENDERED_FILES_WITH_COEFF_COMMANDS}\n$file_name"
+        else
+            RENDERED_FILES_WITH_COEFF_COMMANDS="$file_name"
+        fi
+    fi
+done
+IFS=$OIFS
+
 
-for file_name in "$OUT_FILE_INIT" "$OUT_FILE_INPUT_SCRIPT" "$OUT_FILE_SETTINGS"; do
+
+# ---- Now loop over all user-created files containing "coeff" commands. ----
+echo "" > input_scripts_so_far.tmp
+IFS=$CR
+for file_name in $RENDERED_FILES_WITH_COEFF_COMMANDS; do
     if [ -s "$file_name" ]; then
         echo "postprocessing file \"$file_name\"" >&2
         if ! $PYTHON_COMMAND "${PY_SCR_DIR}/postprocess_input_script.py" input_scripts_so_far.tmp < "$file_name" > "$file_name.tmp"; then
@@ -2404,6 +2490,7 @@ for file_name in "$OUT_FILE_INIT" "$OUT_FILE_INPUT_SCRIPT" "$OUT_FILE_SETTINGS";
         fi
     fi
 done
+IFS=$OIFS
 
 
 # Process any custom input script files created by the user afterwards

setup.py

@@ -45,9 +45,9 @@
 
   url='https://github.com/jewettaij/moltemplate',
 
-  download_url='https://github.com/jewettaij/moltemplate/archive/v2.18.5.zip',
+  download_url='https://github.com/jewettaij/moltemplate/archive/v2.18.6.zip',
 
-  version='2.18.5',
+  version='2.18.6',
 
   keywords=['simulation', 'LAMMPS', 'molecule editor', 'molecule builder',
             'ESPResSo'],