Add ability to run .sql and .sh scripts at the container startup after initdb

[b-a-t]

Official image allows to run arbitary *.sql and *.sh scripts after initialization with initdb, but before full production start of the container. Having similar feature with your image would be great too. That would resolve, for example, #54 and #70 without adding extra logic to the image.

Library image uses /docker-entrypoint-initdb.d/ directory as a place where all such code snippets should be placed. That can be done, for example in the inherited image. Something like this can be added(possibly requires more sanity checks?):

diff --git a/runtime/functions b/runtime/functions
index 440c206..448ac1d 100755
--- a/runtime/functions
+++ b/runtime/functions
@@ -284,6 +284,18 @@ configure_recovery() {
   fi
 }

+initdb_scripts() {
+  echo "Running DB initialization scripts..."
+  for file in /docker-entrypoint-initdb.d/*; do
+    case "$file" in
+      *.sh)     echo "running $file"; . "$file";;
+      *.sql)    echo "running $file"; psql -U ${PG_USER} -f "$file" >/dev/null 2>&1; echo;;
+      *.sql.gz) echo "running $file"; gunzip -c "$file" | psql -U ${PG_USER} >/dev/null 2>&1; echo;;
+      *)        echo "ignoring $file";;
+    esac
+  done
+}
+
 create_user() {
   if [[ -n ${DB_USER} ]]; then
     case $REPLICATION_MODE in
@@ -379,6 +391,7 @@ configure_postgresql() {
   set_postgresql_param "listen_addresses" "127.0.0.1" quiet
   exec_as_postgres ${PG_BINDIR}/pg_ctl -D ${PG_DATADIR} -w start >/dev/null

+  initdb_scripts
   create_user
   create_database
   create_replication_user

[ashish235]

+1