Merge pull request #34 from volkanunsal/fix/issue-25-rruleset-parsing-whitespace

Fix RRULESET parsing with indented/whitespace input

Author: volkanunsal

postgres-rrule.sql

@@ -1,6 +1,3 @@
--- postgres-rrule v2.1.0
--- Generated on 2026-02-04 17:21:30 UTC
-
 DROP SCHEMA IF EXISTS _rrule CASCADE;
 
 DROP CAST IF EXISTS (_rrule.RRULE AS TEXT);
@@ -149,7 +146,7 @@ $$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;-- Extracts a line value from mul
 CREATE OR REPLACE FUNCTION _rrule.extract_line (input TEXT, marker TEXT)
 RETURNS TEXT AS $$
   -- Clear spaces at the front of the lines
-  WITH trimmed_input as (SELECT regexp_replace(input, '^\s*',  '', 'ng') "r"),
+  WITH trimmed_input as (SELECT regexp_replace(input, '^[ \t]*',  '', 'ng') "r"),
   -- Clear all lines except the ones starting with marker
   filtered_lines as (SELECT regexp_replace(trimmed_input."r", '^(?!' || marker || ').*?$',  '', 'ng') "r" FROM trimmed_input),
   -- Replace carriage returns with blank space.
@@ -165,7 +162,7 @@ $$ LANGUAGE SQL IMMUTABLE STRICT PARALLEL SAFE;
 CREATE OR REPLACE FUNCTION _rrule.parse_line (input TEXT, marker TEXT)
 RETURNS SETOF TEXT AS $$
   -- Clear spaces at the front of the lines
-  WITH trimmed_input as (SELECT regexp_replace(input, '^\s*',  '', 'ng') "r"),
+  WITH trimmed_input as (SELECT regexp_replace(input, '^[ \t]*',  '', 'ng') "r"),
   -- Clear all lines except the ones starting with marker
   filtered_lines as (SELECT regexp_replace(trimmed_input."r", '^(?!' || marker || ').*?$',  '', 'ng') "r" FROM trimmed_input),
   -- Replace carriage returns with blank space.

src/functions/0004-extract_line.sql

@@ -9,7 +9,7 @@
 CREATE OR REPLACE FUNCTION _rrule.extract_line (input TEXT, marker TEXT)
 RETURNS TEXT AS $$
   -- Clear spaces at the front of the lines
-  WITH trimmed_input as (SELECT regexp_replace(input, '^\s*',  '', 'ng') "r"),
+  WITH trimmed_input as (SELECT regexp_replace(input, '^[ \t]*',  '', 'ng') "r"),
   -- Clear all lines except the ones starting with marker
   filtered_lines as (SELECT regexp_replace(trimmed_input."r", '^(?!' || marker || ').*?$',  '', 'ng') "r" FROM trimmed_input),
   -- Replace carriage returns with blank space.

src/functions/0005-parse_line.sql

@@ -1,7 +1,7 @@
 CREATE OR REPLACE FUNCTION _rrule.parse_line (input TEXT, marker TEXT)
 RETURNS SETOF TEXT AS $$
   -- Clear spaces at the front of the lines
-  WITH trimmed_input as (SELECT regexp_replace(input, '^\s*',  '', 'ng') "r"),
+  WITH trimmed_input as (SELECT regexp_replace(input, '^[ \t]*',  '', 'ng') "r"),
   -- Clear all lines except the ones starting with marker
   filtered_lines as (SELECT regexp_replace(trimmed_input."r", '^(?!' || marker || ').*?$',  '', 'ng') "r" FROM trimmed_input),
   -- Replace carriage returns with blank space.

tests/test_whitespace_parsing.sql

@@ -0,0 +1,73 @@
+BEGIN;
+
+SELECT plan(8);
+
+SET search_path TO _rrule, public;
+
+-- Test 1: RRULESET with space-indented lines (exact scenario from issue #25)
+SELECT is(
+    (rruleset('
+    DTSTART:19970902T090000
+    RRULE:FREQ=WEEKLY;UNTIL=19980902T090000
+    '))."dtstart"::TEXT,
+    '1997-09-02 09:00:00',
+    'RRULESET with space-indented lines parses DTSTART correctly'
+);
+
+-- Test 2: Verify the RRULE is also parsed correctly from indented input
+SELECT is(
+    ((rruleset('
+    DTSTART:19970902T090000
+    RRULE:FREQ=WEEKLY;UNTIL=19980902T090000
+    '))."rrule")[1]."freq"::TEXT,
+    'WEEKLY',
+    'RRULESET with space-indented lines parses RRULE FREQ correctly'
+);
+
+-- Test 3: RRULESET with tab-indented lines
+SELECT is(
+    (rruleset(E'\tDTSTART:19970902T090000\n\tRRULE:FREQ=DAILY;COUNT=10\n'))."dtstart"::TEXT,
+    '1997-09-02 09:00:00',
+    'RRULESET with tab-indented lines parses DTSTART correctly'
+);
+
+-- Test 4: Verify RRULE from tab-indented input
+SELECT is(
+    ((rruleset(E'\tDTSTART:19970902T090000\n\tRRULE:FREQ=DAILY;COUNT=10\n'))."rrule")[1]."freq"::TEXT,
+    'DAILY',
+    'RRULESET with tab-indented lines parses RRULE FREQ correctly'
+);
+
+-- Test 5: RRULESET with mixed whitespace (spaces and tabs)
+SELECT is(
+    (rruleset(E'  \tDTSTART:19970902T090000\n \t RRULE:FREQ=MONTHLY;COUNT=5\n'))."dtstart"::TEXT,
+    '1997-09-02 09:00:00',
+    'RRULESET with mixed spaces and tabs parses DTSTART correctly'
+);
+
+-- Test 6: Verify RRULE from mixed whitespace input
+SELECT is(
+    ((rruleset(E'  \tDTSTART:19970902T090000\n \t RRULE:FREQ=MONTHLY;COUNT=5\n'))."rrule")[1]."freq"::TEXT,
+    'MONTHLY',
+    'RRULESET with mixed spaces and tabs parses RRULE FREQ correctly'
+);
+
+-- Test 7: Regression test - RRULESET without any indentation still works
+SELECT is(
+    (rruleset('DTSTART:19970902T090000
+RRULE:FREQ=WEEKLY;UNTIL=19980902T090000'))."dtstart"::TEXT,
+    '1997-09-02 09:00:00',
+    'RRULESET without indentation still parses DTSTART correctly (regression)'
+);
+
+-- Test 8: Regression test - RRULE without indentation still works
+SELECT is(
+    ((rruleset('DTSTART:19970902T090000
+RRULE:FREQ=WEEKLY;UNTIL=19980902T090000'))."rrule")[1]."freq"::TEXT,
+    'WEEKLY',
+    'RRULESET without indentation still parses RRULE FREQ correctly (regression)'
+);
+
+SELECT * FROM finish();
+
+ROLLBACK;