Merge #2188

b10-loadzone tries to remove comments from RData strings

src/bin/loadzone/tests/correct/Makefile.am

@@ -12,6 +12,7 @@ EXTRA_DIST += ttl1.db
 EXTRA_DIST += ttl2.db
 EXTRA_DIST += ttlext.db
 EXTRA_DIST += example.db
+EXTRA_DIST += comment.db
 
 noinst_SCRIPTS = correct_test.sh
 

src/bin/loadzone/tests/correct/comment.db

+; Test removal of comments and not removal from strings.
+; We had a bug - see #2188.
+comment.example.com.		60	IN	SOA	ns1.example.com. hostmaster.example.com. 1 43200 900 1814400 7200
+comment.example.com.		60	IN	NS	ns1.example.com.
+comment.example.com.		60	IN	TXT	"Simple text"
+comment.example.com.		60	IN	TXT	"; No comment"
+comment.example.com.		60	IN	TXT	"Also no comment here" ; But here it is a comment
+comment.example.com.		60	IN	TXT	"A combination ; see?" ; This is a "comment

src/bin/loadzone/tests/correct/correct_test.sh.in

@@ -48,6 +48,9 @@ ${LOADZONE_PATH}/b10-loadzone -d ${TEST_OUTPUT_PATH}/zone.sqlite3 ttlext.db >> /
 echo "loadzone example.com. from example.db"
 ${LOADZONE_PATH}/b10-loadzone -d ${TEST_OUTPUT_PATH}/zone.sqlite3 example.db >> /dev/null
 
+echo "loadzone comment.example.com. from comment.db"
+${LOADZONE_PATH}/b10-loadzone -d ${TEST_OUTPUT_PATH}/zone.sqlite3 comment.db >> /dev/null
+
 echo "I:test master file \$INCLUDE semantics"
 echo "I:test master file BIND 8 compatibility TTL and \$TTL semantics"
 echo "I:test master file RFC1035 TTL and \$TTL semantics"
@@ -55,6 +58,7 @@ echo "I:test master file BIND8 compatibility and mixed \$INCLUDE with \$TTL sema
 echo "I:test master file RFC1035 TTL and mixed \$INCLUDE with \$TTL semantics"
 echo "I:test master file BIND9 extenstion of TTL"
 echo "I:test master file RFC1035 missing CLASS, TTL, NAME semantics"
+echo "I:test master file comments"
 
 ${PYTHON_EXEC} ${TEST_FILE_PATH}/get_zonedatas.py ${TEST_OUTPUT_PATH}/zone.sqlite3 > ${TEST_OUTPUT_PATH}/test.out
 echo "Compare test results."

src/bin/loadzone/tests/correct/get_zonedatas.py

 from isc.datasrc import sqlite3_ds
 import sys
 ZONE_FILE = sys.argv[1]
-zonename_set = ["include.", "ttl1.", "ttl2.", "mix1.", "mix2.", "ttlext.", "example.com."]
+zonename_set = ["include.", "ttl1.", "ttl2.", "mix1.", "mix2.", "ttlext.", "example.com.", "comment.example.com."]
 for zone_name in zonename_set:
     for rr_data in sqlite3_ds.get_zone_datas(zone_name, ZONE_FILE):
         data_len = len(rr_data[2])

src/bin/loadzone/tests/correct/known.test.out

@@ -77,3 +77,9 @@ ns2.example.com.		80	IN	A	1.1.1.1
 ns3.example.com.		60	IN	A	2.2.2.2
 ns4.example.com.		60	IN	A	3.3.3.3
 ns5.example.com.		90	IN	A	4.4.4.4
+comment.example.com.		60	IN	SOA	ns1.example.com. hostmaster.example.com. 1 43200 900 1814400 7200
+comment.example.com.		60	IN	NS	ns1.example.com.
+comment.example.com.		60	IN	TXT	"Simple text"
+comment.example.com.		60	IN	TXT	"; No comment"
+comment.example.com.		60	IN	TXT	"Also no comment here"
+comment.example.com.		60	IN	TXT	"A combination ; see?"

src/lib/python/isc/datasrc/master.py

@@ -46,11 +46,16 @@ def pop(line):
 #   whitespace removed, and all other whitespace compressed to
 #   single spaces
 #########################################################################
-decomment = re.compile('\s*(?:;.*)+')
+decomment = re.compile('^\s*((?:[^;"]|"[^"]*")*)\s*(?:|;.*)$')
+# Regular expression explained:
+# First, ignore any whitespace at the start. Then take the content,
+# each bit is either a harmless character (no ; nor ") or a string -
+# sequence between " " not containing double quotes. Then there may
+# be a comment at the end.
 def cleanup(s):
     global decomment
     s = s.strip().expandtabs()
-    s = decomment.sub('', s)
+    s = decomment.search(s).group(1)
     return ' '.join(s.split())
 
 #########################################################################