You are viewing a plain text version of this content. The canonical link for it is here.
Posted to commits@cassandra.apache.org by ca...@codespot.com on 2012/03/24 20:25:35 UTC
[cassandra-dbapi2] 2 new revisions pushed by pcannon@gmail.com on
2012-03-24 19:25 GMT
2 new revisions:
Revision: 649efebe5471
Author: paul cannon <pa...@datastax.com>
Date: Tue Feb 28 15:33:03 2012
Log: fix error message on unknown CqlResultType...
http://code.google.com/a/apache-extras.org/p/cassandra-dbapi2/source/detail?r=649efebe5471
Revision: f5c00574cb80
Author: paul cannon <pa...@datastax.com>
Date: Fri Mar 23 17:10:42 2012
Log: multistep param parsing...
http://code.google.com/a/apache-extras.org/p/cassandra-dbapi2/source/detail?r=f5c00574cb80
==============================================================================
Revision: 649efebe5471
Author: paul cannon <pa...@datastax.com>
Date: Tue Feb 28 15:33:03 2012
Log: fix error message on unknown CqlResultType
It's not at all clear how it ever happened, but a user had the
cursor.process_execution_results() method receive None as the
response.type one time (which generated "TypeError: cannot concatenate
'str' and 'NoneType' objects" instead of "unknown result type None",
because of the invalid expectation of stringliness herein fixed,
which wouldn't have been very much more helpful, but at least this is a
little bit more correct). User also saw a KsDef returned with None in
the 'name' attribute, suggesting something outside of python-cql being
badly broken.
so, not much of a problem, but woohoo.
http://code.google.com/a/apache-extras.org/p/cassandra-dbapi2/source/detail?r=649efebe5471
Modified:
/cql/cursor.py
=======================================
--- /cql/cursor.py Fri Feb 10 06:41:21 2012
+++ /cql/cursor.py Tue Feb 28 15:33:03 2012
@@ -155,7 +155,7 @@
self.rowcount = 0
self.description = _VOID_DESCRIPTION
else:
- raise Exception('unknown result type ' + response.type)
+ raise Exception('unknown result type %s' % response.type)
# 'Return values are not defined.'
return True
==============================================================================
Revision: f5c00574cb80
Author: paul cannon <pa...@datastax.com>
Date: Fri Mar 23 17:10:42 2012
Log: multistep param parsing
this should have linear complexity, unlike the old regex with all
the backtracking it required.
http://code.google.com/a/apache-extras.org/p/cassandra-dbapi2/source/detail?r=f5c00574cb80
Modified:
/cql/marshal.py
=======================================
--- /cql/marshal.py Fri Feb 10 06:59:55 2012
+++ /cql/marshal.py Fri Mar 23 17:10:42 2012
@@ -50,24 +50,6 @@
def __init__(self, bytes):
self.bytes = bytes
-_param_re = re.compile(r"""
- ( # stuff that is not substitution markers
- (?: ' [^']* ' # string literal; ignore colons in here
- | [^'] # eat anything else
- )*?
- )
- (?<! [a-zA-Z0-9_'] ) # no colons immediately preceded by an
ident or str literal
- :
- ( [a-zA-Z_][a-zA-Z0-9_]* ) # the param name
-""", re.S | re.X)
-
-_comment_re = re.compile(r"""
- // .*? $
- | -- .*? $
- | /\* .*? \*/
- | ' [^']* '
-""", re.S | re.M | re.X)
-
BYTES_TYPE = "org.apache.cassandra.db.marshal.BytesType"
ASCII_TYPE = "org.apache.cassandra.db.marshal.AsciiType"
BOOLEAN_TYPE = "org.apache.cassandra.db.marshal.BooleanType"
@@ -84,6 +66,34 @@
TIME_UUID_TYPE = "org.apache.cassandra.db.marshal.TimeUUIDType"
COUNTER_COLUMN_TYPE = "org.apache.cassandra.db.marshal.CounterColumnType"
+stringlit_re = re.compile(r"""('[^']*'|"[^"]*")""")
+comments_re = re.compile(r'(/\*(?:[^*]|\*[^/])*\*/|//.*$|--.*$)',
re.MULTILINE)
+param_re = re.compile(r'''
+ ( [^a-z0-9_] ) # don't match : at the beginning of the text
(meaning it
+ # immediately follows a comment or string literal) or
+ # right after an identifier character
+ : ( [a-z0-9_]+ )
+ ( [^a-z0-9_] ) # and don't match a param that is immediately
followed by
+ # a comment or string literal either
+''', re.IGNORECASE | re.VERBOSE)
+
+def replace_param_substitutions(query, replacer):
+ split_strings = stringlit_re.split(' ' + query + ' ')
+ split_str_and_cmt = []
+ for p in split_strings:
+ if p[:1] in '\'"':
+ split_str_and_cmt.append(p)
+ else:
+ split_str_and_cmt.extend(comments_re.split(p))
+ output = []
+ for p in split_str_and_cmt:
+ if p[:1] in '\'"' or p[:2] in ('--', '//', '/*'):
+ output.append(p)
+ else:
+ output.append(param_re.sub(replacer, p))
+ assert output[0][0] == ' ' and output[-1][-1] == ' '
+ return ''.join(output)[1:-1]
+
class PreparedQuery(object):
def __init__(self, querytext, itemid, vartypes, paramnames):
self.querytext = querytext
@@ -97,14 +107,6 @@
def encode_params(self, params):
return [cql_marshal(params[n], t) for (n, t) in
zip(self.paramnames, self.vartypes)]
-def blank_comments(query):
- def teh_blanker(match):
- m = match.group(0)
- if m.startswith("'"):
- return m
- return ' ' * len(m)
- return _comment_re.sub(teh_blanker, query)
-
def prepare_inline(query, params):
"""
For every match of the form ":param_name", call cql_quote
@@ -112,21 +114,19 @@
with the result
"""
- # kill comments first, so that we don't have to try to parse around
them.
- # but keep the character count the same, so that location-tagged error
- # messages still work
- query = blank_comments(query)
- return _param_re.sub(lambda m: m.group(1) +
cql_quote(params[m.group(2)]), query)
+ def param_replacer(match):
+ return match.group(1) + cql_quote(params[match.group(2)]) +
match.group(3)
+ return replace_param_substitutions(query, param_replacer)
def prepare_query(querytext):
- querytext = blank_comments(querytext)
paramnames = []
def found_param(match):
pre_param_text = match.group(1)
paramname = match.group(2)
+ post_param_text = match.group(3)
paramnames.append(paramname)
- return pre_param_text + '?'
- transformed_query = _param_re.sub(found_param, querytext)
+ return pre_param_text + '?' + post_param_text
+ transformed_query = replace_param_substitutions(querytext, found_param)
return transformed_query, paramnames