further fixes to maintain_901

Now correctly handles &, <, or > appearing in the tcn_value
or tcn_source.  Also introduces a trivial utility stored
function, evergreen.xml_escape():

evergreen.xml_escape(TEXT) returns TEXT: converts '&', '<',
and '>' to XML entities

Signed-off-by: Galen Charlton <gmc@esilibrary.com>

git-svn-id: svn://svn.open-ils.org/ILS/trunk@20014 dcc99617-32d9-48b4-a31d-7c20da2025e4

diff --git a/Open-ILS/src/sql/Pg/000.functions.general.sql b/Open-ILS/src/sql/Pg/000.functions.general.sql
index f960341..cb6f82b 100644 (file)
--- a/Open-ILS/src/sql/Pg/000.functions.general.sql
+++ b/Open-ILS/src/sql/Pg/000.functions.general.sql
@@ -11,6 +11,13 @@ CREATE OR REPLACE FUNCTION evergreen.lowercase( TEXT ) RETURNS TEXT AS $$
     return lc(shift);
 $$ LANGUAGE PLPERLU STRICT IMMUTABLE;
 
+CREATE OR REPLACE FUNCTION evergreen.xml_escape(str TEXT) RETURNS text AS $$
+    SELECT REPLACE(REPLACE(REPLACE($1,
+       '&', '&'),
+       '<', '&lt;'),
+       '>', '&gt;');
+$$ LANGUAGE SQL IMMUTABLE;
+
 CREATE OR REPLACE FUNCTION evergreen.change_db_setting(setting_name TEXT, settings TEXT[]) RETURNS VOID AS $$
 BEGIN
 EXECUTE 'ALTER DATABASE ' || quote_ident(current_database()) || ' SET ' || quote_ident(setting_name) || ' = ' || array_to_string(settings, ',');

diff --git a/Open-ILS/src/sql/Pg/002.functions.config.sql b/Open-ILS/src/sql/Pg/002.functions.config.sql
index 99f98d8..cc594d0 100644 (file)
--- a/Open-ILS/src/sql/Pg/002.functions.config.sql
+++ b/Open-ILS/src/sql/Pg/002.functions.config.sql
@@ -422,8 +422,6 @@ $f$ LANGUAGE PLPERLU;
 CREATE OR REPLACE FUNCTION maintain_901 () RETURNS TRIGGER AS $func$
 DECLARE
     use_id_for_tcn BOOLEAN;
-    norm_tcn_value TEXT;
-    norm_tcn_source TEXT;
 BEGIN
     -- Remove any existing 901 fields before we insert the authoritative one
     NEW.marc := REGEXP_REPLACE(NEW.marc, E'<datafield[^>]*?tag="901".+?</datafield>', '', 'g');
@@ -435,21 +433,14 @@ BEGIN
 
         IF use_id_for_tcn = 't' THEN
             NEW.tcn_value := NEW.id;
-            norm_tcn_value := NEW.tcn_value;
-        ELSE
-            -- yes, ampersands can show up in tcn_values ...
-            norm_tcn_value := REGEXP_REPLACE(NEW.tcn_value, E'&(?!\\S+;)', '&amp;', 'g');
         END IF;
-        -- ... and TCN sources
-        -- FIXME we have here yet another (stub) version of entityize
-        norm_tcn_source := REGEXP_REPLACE(NEW.tcn_source, E'&(?!\\S+;)', '&amp;', 'g');
 
         NEW.marc := REGEXP_REPLACE(
             NEW.marc,
             E'(</(?:[^:]*?:)?record>)',
             E'<datafield tag="901" ind1=" " ind2=" ">' ||
-                '<subfield code="a">' || norm_tcn_value || E'</subfield>' ||
-                '<subfield code="b">' || norm_tcn_source || E'</subfield>' ||
+                '<subfield code="a">' || evergreen.xml_escape(NEW.tcn_value) || E'</subfield>' ||
+                '<subfield code="b">' || evergreen.xml_escape(NEW.tcn_source) || E'</subfield>' ||
                 '<subfield code="c">' || NEW.id || E'</subfield>' ||
                 '<subfield code="t">' || TG_TABLE_SCHEMA || E'</subfield>' ||
                 CASE WHEN NEW.owner IS NOT NULL THEN '<subfield code="o">' || NEW.owner || E'</subfield>' ELSE '' END ||

diff --git a/Open-ILS/src/sql/Pg/002.schema.config.sql b/Open-ILS/src/sql/Pg/002.schema.config.sql
index 2f4138c..cf73e7b 100644 (file)
--- a/Open-ILS/src/sql/Pg/002.schema.config.sql
+++ b/Open-ILS/src/sql/Pg/002.schema.config.sql
@@ -70,7 +70,7 @@ CREATE TABLE config.upgrade_log (
     install_date    TIMESTAMP WITH TIME ZONE NOT NULL DEFAULT NOW()
 );
 
-INSERT INTO config.upgrade_log (version) VALUES ('0508'); -- gmc
+INSERT INTO config.upgrade_log (version) VALUES ('0509'); -- gmc
 
 CREATE TABLE config.bib_source (
        id              SERIAL  PRIMARY KEY,