Make more difficult sanitize of the expression string before eval

ANNOUNCE.rst

@@ -4,7 +4,10 @@ Announcing NumExpr 2.8.5
 
 Hi everyone, 
 
-**Under development.**
+In 2.8.5 we have added a new function, `validate` which checks an expression `ex`
+for validity, for usage where the program is parsing a user input. There are also 
+consequences for this sort of usage, since `eval(ex)` is called, and as such we 
+do some string sanitization as described below.
 
 Project documentation is available at:
 
@@ -13,7 +16,23 @@ http://numexpr.readthedocs.io/
 Changes from 2.8.4 to 2.8.5
 ---------------------------
 
-**Under development.**
+* A `validate` function has been added. This function checks the inputs, returning 
+  `None` on success or raising an exception on invalid inputs. This function was 
+  added as numerous projects seem to be using NumExpr for parsing user inputs.
+  `re_evaluate` may be called directly following `validate`.
+* As an addendum to the use of NumExpr for parsing user inputs, is that NumExpr
+  calls `eval` on the inputs. A regular expression is now applied to help sanitize 
+  the input expression string, forbidding '__', ':', and ';'. Attribute access 
+  is also banned except for '.r' for real and '.i'  for imag.
+* Thanks to timbrist for a fix to behavior of NumExpr with integers to negative 
+  powers. NumExpr was pre-checking integer powers for negative values, which 
+  was both inefficient and causing parsing errors in some situations. Now NumExpr
+  will simply return 0 as a result for such cases. While NumExpr generally tries 
+  to follow NumPy behavior, performance is also critical. 
+* Thanks to peadar for some fixes to how NumExpr launches threads for embedded 
+  applications.
+* Thanks to de11n for making parsing of the `site.cfg` for MKL consistent among 
+  all shared platforms.
 
 
 What's Numexpr?

RELEASE_NOTES.rst

@@ -5,7 +5,24 @@ Release notes for NumExpr 2.8 series
 Changes from 2.8.4 to 2.8.5
 ---------------------------
 
-**Under development.**
+* A `validate` function has been added. This function checks the inputs, returning 
+  `None` on success or raising an exception on invalid inputs. This function was 
+  added as numerous projects seem to be using NumExpr for parsing user inputs.
+  `re_evaluate` may be called directly following `validate`.
+* As an addendum to the use of NumExpr for parsing user inputs, is that NumExpr
+  calls `eval` on the inputs. A regular expression is now applied to help sanitize 
+  the input expression string, forbidding '__', ':', and ';'. Attribute access 
+  is also banned except for '.r' for real and '.i'  for imag.
+* Thanks to timbrist for a fix to behavior of NumExpr with integers to negative 
+  powers. NumExpr was pre-checking integer powers for negative values, which 
+  was both inefficient and causing parsing errors in some situations. Now NumExpr
+  will simply return 0 as a result for such cases. While NumExpr generally tries 
+  to follow NumPy behavior, performance is also critical. 
+* Thanks to peadar for some fixes to how NumExpr launches threads for embedded 
+  applications.
+* Thanks to de11n for making parsing of the `site.cfg` for MKL consistent among 
+  all shared platforms.
+
 
 Changes from 2.8.3 to 2.8.4
 ---------------------------

doc/user_guide.rst

@@ -1,7 +1,7 @@
-NumExpr 2.0 User Guide
+NumExpr 2.8 User Guide
 ======================
 
-The :code:`numexpr` package supplies routines for the fast evaluation of
+The NumExpr package supplies routines for the fast evaluation of
 array expressions elementwise by using a vector-based virtual
 machine.
 
@@ -11,23 +11,33 @@ Using it is simple::
     >>> import numexpr as ne
     >>> a = np.arange(10)
     >>> b = np.arange(0, 20, 2)
-    >>> c = ne.evaluate("2*a+3*b")
+    >>> c = ne.evaluate('2*a + 3*b')
     >>> c
     array([ 0,  8, 16, 24, 32, 40, 48, 56, 64, 72])
 
 
+It is also possible to use NumExpr to validate an expression::
+
+    >>> ne.validate('2*a + 3*b')
+
+which returns `None` on success or raises an exception on invalid inputs.
+
+and it can also re_evaluate an expression::
+
+    >>> b = np.arange(0, 40, 4)
+    >>> ne.re_evaluate()
+
 Building
 --------
 
-*NumExpr* requires Python_ 2.6 or greater, and NumPy_ 1.7 or greater.  It is 
+*NumExpr* requires Python_ 3.7 or greater, and NumPy_ 1.13 or greater.  It is 
 built in the standard Python way:
 
 .. code-block:: bash
 
-    $ python setup.py build
-    $ python setup.py install
+    $ pip install .
 
-You must have a C-compiler (i.e. MSVC on Windows and GCC on Linux) installed.
+You must have a C-compiler (i.e. MSVC Build tools on Windows and GCC on Linux) installed.
 
 Then change to a directory that is not the repository directory (e.g. `/tmp`) and 
 test :code:`numexpr` with:
@@ -268,9 +278,6 @@ General routines
   * :code:`detect_number_of_cores()`: Detects the number of cores on a system.
 
 
-
-
-
 Intel's VML specific support routines
 -------------------------------------
 

numexpr/necompiler.py

@@ -260,15 +260,17 @@ def __init__(self, astnode):
     def __str__(self):
         return 'Immediate(%d)' % (self.node.value,)
 
-_forbidden_re = re.compile('[\;[\:]|__')
+
+_forbidden_re = re.compile('[\;[\:]|__|\.[abcdefghjklmnopqstuvwxyzA-Z_]')
 def stringToExpression(s, types, context):
     """Given a string, convert it to a tree of ExpressionNode's.
     """
     # sanitize the string for obvious attack vectors that NumExpr cannot 
     # parse into its homebrew AST. This is to protect the call to `eval` below.
-    # We forbid `;`, `:`. `[` and `__`
-    # We would like to forbid `.` but it is both a reference and decimal point.
-    if _forbidden_re.search(s) is not None:
+    # We forbid `;`, `:`. `[` and `__`, and attribute access via '.'.
+    # We cannot ban `.real` or `.imag` however...
+    no_whitespace = re.sub(r'\s+', '', s)
+    if _forbidden_re.search(no_whitespace) is not None:
         raise ValueError(f'Expression {s} has forbidden control characters.')
 
     old_ctx = expressions._context.get_current_context()
@@ -766,7 +768,6 @@ def getArguments(names, local_dict=None, global_dict=None, _frame_depth: int=2):
 _names_cache = CacheDict(256)
 _numexpr_cache = CacheDict(256)
 _numexpr_last = {}
-_numexpr_sanity = set()
 evaluate_lock = threading.Lock()
 
 # MAYBE: decorate this function to add attributes instead of having the 
@@ -828,6 +829,13 @@ def validate(ex: str,
     _frame_depth: int
         The calling frame depth. Unless you are a NumExpr developer you should 
         not set this value.
+
+    Note
+    ----
+    Both `validate` and by extension `evaluate` call `eval(ex)`, which is 
+    potentially dangerous on unsanitized inputs. As such, NumExpr does some 
+    sanitization, banning the character ':;[', the dunder '__', and attribute
+    access to all but '.r' for real and '.i' for imag access to complex numbers.
     """
     global _numexpr_last
 
@@ -857,8 +865,6 @@ def validate(ex: str,
         kwargs = {'out': out, 'order': order, 'casting': casting,
                 'ex_uses_vml': ex_uses_vml}
         _numexpr_last = dict(ex=compiled_ex, argnames=names, kwargs=kwargs)
-        # with evaluate_lock:
-        #     return compiled_ex(*arguments, **kwargs)
     except Exception as e:
         return e
     return None
@@ -918,7 +924,12 @@ def evaluate(ex: str,
         The calling frame depth. Unless you are a NumExpr developer you should 
         not set this value.
 
-    
+    Note
+    ----
+    Both `validate` and by extension `evaluate` call `eval(ex)`, which is 
+    potentially dangerous on unsanitized inputs. As such, NumExpr does some 
+    sanitization, banning the character ':;[', the dunder '__', and attribute
+    access to all but '.r' for real and '.i' for imag access to complex numbers.
     """
     # We could avoid code duplication if we called validate and then re_evaluate 
     # here, but they we have difficulties with the `sys.getframe(2)` call in

numexpr/tests/test_numexpr.py

@@ -536,13 +536,27 @@ def test_forbidden_tokens(self):
 
         # Forbid semicolon
         try:
-            evaluate('import os; os.cpu_count()')
+            evaluate('import os;')
         except ValueError:
             pass
         else:
             self.fail()
 
-        # I struggle to come up with cases for our ban on `'` and `"`
+        # Attribute access
+        try:
+            evaluate('os.cpucount()')
+        except ValueError:
+            pass
+        else:
+            self.fail()
+
+        # But decimal point must pass
+        a = 3.0
+        evaluate('a*2.')
+        evaluate('2.+a')
+
+
+