From 5b69bc7ca790156ee4422fa1611c05ad6d5ebc47 Mon Sep 17 00:00:00 2001
From: Iustin Pop <iustin@google.com>
Date: Wed, 13 Jan 2010 16:56:42 +0100
Subject: [PATCH] Add an UnescapeAndSplit function
In many cases, where we accept (usually from the command line) a list of
parameters, we remove the use of the separator as an component of any of
the elements.
This patch adds a new function that can split strings of the form
'a,b\,c,d' into ['a', 'b,c', 'd'], with proper un-escaping of
double-backslashes.
Signed-off-by: Iustin Pop <iustin@google.com>
Reviewed-by: Michael Hanselmann <hansmi@google.com>
---
lib/utils.py | 42 +++++++++++++++++++++++++++++++++++
test/ganeti.utils_unittest.py | 34 +++++++++++++++++++++++++++-
2 files changed, 75 insertions(+), 1 deletion(-)
diff --git a/lib/utils.py b/lib/utils.py
index 1abec0ef2..6e3340a77 100644
--- a/lib/utils.py
+++ b/lib/utils.py
@@ -1897,6 +1897,48 @@ def SafeEncode(text):
return resu
+def UnescapeAndSplit(text, sep=","):
+ """Split and unescape a string based on a given separator.
+
+ This function splits a string based on a separator where the
+ separator itself can be escape in order to be an element of the
+ elements. The escaping rules are (assuming coma being the
+ separator):
+ - a plain , separates the elements
+ - a sequence \\\\, (double backslash plus comma) is handled as a
+ backslash plus a separator comma
+ - a sequence \, (backslash plus comma) is handled as a
+ non-separator comma
+
+ @type text: string
+ @param text: the string to split
+ @type sep: string
+ @param text: the separator
+ @rtype: string
+ @return: a list of strings
+
+ """
+ # we split the list by sep (with no escaping at this stage)
+ slist = text.split(sep)
+ # next, we revisit the elements and if any of them ended with an odd
+ # number of backslashes, then we join it with the next
+ rlist = []
+ while slist:
+ e1 = slist.pop(0)
+ if e1.endswith("\\"):
+ num_b = len(e1) - len(e1.rstrip("\\"))
+ if num_b % 2 == 1:
+ e2 = slist.pop(0)
+ # here the backslashes remain (all), and will be reduced in
+ # the next step
+ rlist.append(e1 + sep + e2)
+ continue
+ rlist.append(e1)
+ # finally, replace backslash-something with something
+ rlist = [re.sub(r"\\(.)", r"\1", v) for v in rlist]
+ return rlist
+
+
def CommaJoin(names):
"""Nicely join a set of identifiers.
diff --git a/test/ganeti.utils_unittest.py b/test/ganeti.utils_unittest.py
index fe7464de5..f0763550d 100755
--- a/test/ganeti.utils_unittest.py
+++ b/test/ganeti.utils_unittest.py
@@ -45,7 +45,8 @@ from ganeti.utils import IsProcessAlive, RunCmd, \
ParseUnit, AddAuthorizedKey, RemoveAuthorizedKey, \
ShellQuote, ShellQuoteArgs, TcpPing, ListVisibleFiles, \
SetEtcHostsEntry, RemoveEtcHostsEntry, FirstFree, OwnIpAddress, \
- TailFile, ForceDictType, SafeEncode, IsNormAbsPath, FormatTime
+ TailFile, ForceDictType, SafeEncode, IsNormAbsPath, FormatTime, \
+ UnescapeAndSplit
from ganeti.errors import LockError, UnitParseError, GenericError, \
ProgrammerError
@@ -1053,5 +1054,36 @@ class TestFormatTime(unittest.TestCase):
FormatTime(int(time.time()))
+class TestUnescapeAndSplit(unittest.TestCase):
+ """Testing case for UnescapeAndSplit"""
+
+ def setUp(self):
+ # testing more that one separator for regexp safety
+ self._seps = [",", "+", "."]
+
+ def testSimple(self):
+ a = ["a", "b", "c", "d"]
+ for sep in self._seps:
+ self.failUnlessEqual(UnescapeAndSplit(sep.join(a), sep=sep), a)
+
+ def testEscape(self):
+ for sep in self._seps:
+ a = ["a", "b\\" + sep + "c", "d"]
+ b = ["a", "b" + sep + "c", "d"]
+ self.failUnlessEqual(UnescapeAndSplit(sep.join(a), sep=sep), b)
+
+ def testDoubleEscape(self):
+ for sep in self._seps:
+ a = ["a", "b\\\\", "c", "d"]
+ b = ["a", "b\\", "c", "d"]
+ self.failUnlessEqual(UnescapeAndSplit(sep.join(a), sep=sep), b)
+
+ def testThreeEscape(self):
+ for sep in self._seps:
+ a = ["a", "b\\\\\\" + sep + "c", "d"]
+ b = ["a", "b\\" + sep + "c", "d"]
+ self.failUnlessEqual(UnescapeAndSplit(sep.join(a), sep=sep), b)
+
+
if __name__ == '__main__':
testutils.GanetiTestProgram()
--
GitLab