fix: Multiline quote parsing of httpd conf files (#3392)

* The parsr logic didn't handle line continuations properly, and didn't
  handle multiline quoted directives either. I fixed the logic and added
  in tests.

Signed-off-by: Ryan Blakley <[email protected]>

insights/combiners/tests/test_httpd_conf_tree.py

@@ -509,6 +509,8 @@
     SetEnv GIT_HTTP_EXPORT_ALL
     DocumentRoot /var/www
 #    ScriptAlias /git/ /usr/libexec/git-core/git-http-backend/
+    ProxyPass "http://backend.example.com/a-long-path-to-demonstrate" \
+        connectiontimeout=1 retry=0
     ScriptAliasMatch \
         "(?x)^/git/(.*/(HEAD | \
         info/refs | \
@@ -521,6 +523,10 @@
 </VirtualHost>
 """
 
+INVALID_CONF = """
+<<<>>>><>KDLKJLDSF><SDNF<KNSD><FN
+""".strip()
+
 
 def test_mixed_case_tags():
     httpd = httpd_conf.HttpdConf(context_wrap(HTTPD_CONF_MIXED, path='/etc/httpd/conf/httpd.conf'))
@@ -857,14 +863,32 @@ def test_empty_attr():
 
 def test_multiline_quote():
     with pytest.raises(ParseException):
-        httpd_conf.HttpdConf(context_wrap(HTTPD_MULTILINE_QUOTE, path='/etc/httpd/conf/error.conf'))
-
-    httpd = httpd_conf.HttpdConf(context_wrap(HTTPD_REGEX_AND_OP_ATTRS, path='/etc/httpd/conf/httpd.conf'))
-    result = HttpdConfTree([httpd])
-
-    rewrite_cond = result["RewriteCond"]
-    assert len(rewrite_cond) == 3
-
-    if_version = result["IfVersion"]
-    assert len(if_version) == 1
-    assert if_version.value == "< 2.4"
+        httpd_conf.HttpdConf(context_wrap(INVALID_CONF, path='/etc/httpd/conf/error.conf'))
+
+    ml_quote = httpd_conf.HttpdConf(context_wrap(HTTPD_MULTILINE_QUOTE, path='/etc/httpd/conf.d/multiline-quote.conf'))
+    httpd = httpd_conf.HttpdConf(context_wrap(HTTPD_CONF_1, path='/etc/httpd/conf/httpd.conf'))
+    result = HttpdConfTree([httpd, ml_quote])
+
+    assert result['JustFotTest_NoSec'].value == "/var/www/cgi"
+    assert result['IfModule']['ServerLimit'].value == 256
+    assert result['Listen'].value == 81
+
+    # For the below tests, check the string value, then check the actual
+    # attrs since quotes from quoted text are removed. The attrs store the
+    # actual parsed attrs in a list form.
+    assert result['VirtualHost']['SetEnv'].values == ['GIT_PROJECT_ROOT /var/www/git', 'GIT_HTTP_EXPORT_ALL']
+    assert result['VirtualHost']['SetEnv'].children[0].attrs == ['GIT_PROJECT_ROOT', '/var/www/git']
+    assert result['VirtualHost']['SetEnv'].children[1].attrs == ['GIT_HTTP_EXPORT_ALL']
+
+    assert result['VirtualHost']['ProxyPass'].value == "http://backend.example.com/a-long-path-to-demonstrate connectiontimeout=1 retry=0"  # noqa
+    assert result['VirtualHost']['ProxyPass'].children[-1].attrs == [
+        'http://backend.example.com/a-long-path-to-demonstrate',
+        'connectiontimeout=1',
+        'retry=0'
+    ]
+
+    assert result['VirtualHost']['ScriptAliasMatch'].value == r"(?x)^/git/(.*/(HEAD | info/refs | objects/(info/[^/]+ | [0-9a-f]{2}/[0-9a-f]{38} | pack/pack-[0-9a-f]{40}\.(pack|idx)) | git-(upload|receive)-pack))$ /usr/libexec/git-core/git-http-backend/$1"  # noqa
+    assert result['VirtualHost']['ScriptAliasMatch'].children[-1].attrs == [
+        r'(?x)^/git/(.*/(HEAD | info/refs | objects/(info/[^/]+ | [0-9a-f]{2}/[0-9a-f]{38} | pack/pack-[0-9a-f]{40}\.(pack|idx)) | git-(upload|receive)-pack))$',  # noqa
+        '/usr/libexec/git-core/git-http-backend/$1'
+    ]

insights/parsers/httpd_conf.py

@@ -22,23 +22,24 @@ def __init__(self, ctx):
         Complex = Forward()
         Comment = (WS >> OneLineComment("#")).map(lambda x: None)
 
+        Cont = Char("\\") + EOL
         First = InSet(string.ascii_letters + "_/")
         Rest = String(string.ascii_letters + "_/" + string.digits)
-        Name = (First + Rest).map("".join)
+        FirstRest = (First + Rest).map("".join)
+        Name = (FirstRest << (Many(WSChar) + Cont)) | FirstRest
 
         Num = Number & (WSChar | LineEnd)
 
         StartName = WS >> PosMarker(StartTagName(Letters)) << WS
         EndName = WS >> EndTagName(Letters, ignore_case=True) << WS
 
-        Cont = Char("\\") + EOL
         AttrStart = Many(WSChar)
         AttrEnd = (Many(WSChar) + Cont) | Many(WSChar)
 
         BareAttr = String(set(string.printable) - (set(string.whitespace) | set("<>'\"")))
         OpAttr = (Literal("!=") | Literal("<=") | Literal(">=") | InSet("<>")) & WSChar + BareAttr
         EmptyAttr = String('"\'', min_length=2)
-        Attr = AttrStart >> (Num | QuotedString | OpAttr | BareAttr | EmptyAttr) << AttrEnd
+        Attr = AttrStart >> (Num | QuotedString.map(self.remove_cont) | OpAttr | BareAttr | EmptyAttr) << AttrEnd
         Attrs = Many(Attr)
 
         StartTag = (WS + LT) >> (StartName + Attrs) << (GT + WS)
@@ -51,6 +52,9 @@ def __init__(self, ctx):
 
         self.Top = Doc + EOF
 
+    def remove_cont(self, val):
+        return "".join([x.strip().strip("\\") for x in val.split("\n")])
+
     def typed(self, val):
         try:
             v = val.lower()
@@ -77,7 +81,7 @@ def __call__(self, content):
         try:
             return self.Top(content)
         except Exception:
-            raise ParseException("There was an exception when parsing the config file.")
+            raise ParseException("There was an exception when parsing one of the httpd config files.")
 
 
 class HttpdConfBase(ConfigParser):