Add a release date to the Changelog

fixed #86 and #87

.travis.yml

@@ -6,7 +6,7 @@ python:
   - 3.5
   - 3.6
   - 3.7
-  - 3.8-dev
+  - 3.8
   - pypy2.7-6.0
   - pypy3.5-6.0
 matrix:

CHANGELOG.rst

@@ -3,6 +3,19 @@
 Changelog
 ---------
 
+0.5.2 (2019-12-15)
+++++++++++++++++++
+
+- Add include_setitem to get_definition/is_definition and get_defined_names (#66)
+- Fix named expression error listing (#89, #90)
+- Fix some f-string tokenizer issues (#93)
+
+0.5.1 (2019-07-13)
+++++++++++++++++++
+
+- Fix: Some unicode identifiers were not correctly tokenized
+- Fix: Line continuations in f-strings are now working
+
 0.5.0 (2019-06-20)
 ++++++++++++++++++
 
@@ -17,19 +30,19 @@ Changelog
 - Python 3.8 support
 - FileIO support, it's now possible to use abstract file IO, support is alpha
 
-0.3.4 (2018-02-13)
+0.3.4 (2019-02-13)
 +++++++++++++++++++
 
 - Fix an f-string tokenizer error
 
-0.3.3 (2018-02-06)
+0.3.3 (2019-02-06)
 +++++++++++++++++++
 
 - Fix async errors in the diff parser
 - A fix in iter_errors
 - This is a very small bugfix release
 
-0.3.2 (2018-01-24)
+0.3.2 (2019-01-24)
 +++++++++++++++++++
 
 - 20+ bugfixes in the diff parser and 3 in the tokenizer

conftest.py

@@ -58,7 +58,9 @@ def pytest_generate_tests(metafunc):
     elif 'each_py3_version' in metafunc.fixturenames:
         metafunc.parametrize('each_py3_version', VERSIONS_3)
     elif 'version_ge_py36' in metafunc.fixturenames:
-        metafunc.parametrize('version_ge_py36', ['3.6', '3.7'])
+        metafunc.parametrize('version_ge_py36', ['3.6', '3.7', '3.8'])
+    elif 'version_ge_py38' in metafunc.fixturenames:
+        metafunc.parametrize('version_ge_py38', ['3.8'])
 
 
 class NormalizerIssueCase(object):

parso/__init__.py

@@ -43,7 +43,7 @@ from parso.grammar import Grammar, load_grammar
 from parso.utils import split_lines, python_bytes_to_unicode
 
 
-__version__ = '0.5.0'
+__version__ = '0.5.2'
 
 
 def parse(code=None, **kwargs):

parso/grammar.py

@@ -57,7 +57,8 @@ class Grammar(object):
         :param str path: The path to the file you want to open. Only needed for caching.
         :param bool cache: Keeps a copy of the parser tree in RAM and on disk
             if a path is given. Returns the cached trees if the corresponding
-            files on disk have not changed.
+            files on disk have not changed. Note that this stores pickle files
+            on your file system (e.g. for Linux in ``~/.cache/parso/``).
         :param bool diff_cache: Diffs the cached python module against the new
             code and tries to parse only the parts that have changed. Returns
             the same (changed) module that is found in cache. Using this option

parso/pgen2/generator.py

@@ -309,13 +309,39 @@ def _calculate_tree_traversal(nonterminal_to_dfas):
             _calculate_first_plans(nonterminal_to_dfas, first_plans, nonterminal)
 
     # Now that we have calculated the first terminals, we are sure that
-    # there is no left recursion or ambiguities.
+    # there is no left recursion.
 
     for dfas in nonterminal_to_dfas.values():
         for dfa_state in dfas:
+            transitions = dfa_state.transitions
             for nonterminal, next_dfa in dfa_state.nonterminal_arcs.items():
                 for transition, pushes in first_plans[nonterminal].items():
-                    dfa_state.transitions[transition] = DFAPlan(next_dfa, pushes)
+                    if transition in transitions:
+                        prev_plan = transitions[transition]
+                        # Make sure these are sorted so that error messages are
+                        # at least deterministic
+                        choices = sorted([
+                            (
+                                prev_plan.dfa_pushes[0].from_rule
+                                if prev_plan.dfa_pushes
+                                else prev_plan.next_dfa.from_rule
+                            ),
+                            (
+                                pushes[0].from_rule
+                                if pushes else next_dfa.from_rule
+                            ),
+                        ])
+                        raise ValueError(
+                            "Rule %s is ambiguous; given a %s token, we "
+                            "can't determine if we should evaluate %s or %s."
+                            % (
+                                (
+                                    dfa_state.from_rule,
+                                    transition,
+                                ) + tuple(choices)
+                            )
+                        )
+                    transitions[transition] = DFAPlan(next_dfa, pushes)
 
 
 def _calculate_first_plans(nonterminal_to_dfas, first_plans, nonterminal):
@@ -345,13 +371,6 @@ def _calculate_first_plans(nonterminal_to_dfas, first_plans, nonterminal):
                 raise ValueError("left recursion for rule %r" % nonterminal)
 
         for t, pushes in first_plans2.items():
-            check = new_first_plans.get(t)
-            if check is not None:
-                raise ValueError(
-                    "Rule %s is ambiguous; %s is the"
-                    " start of the rule %s as well as %s."
-                    % (nonterminal, t, nonterminal2, check[-1].from_rule)
-                )
             new_first_plans[t] = [next_] + pushes
 
     first_plans[nonterminal] = new_first_plans

parso/pgen2/grammar_parser.py

@@ -141,6 +141,9 @@ class NFAArc(object):
         self.next = next_
         self.nonterminal_or_string = nonterminal_or_string
 
+    def __repr__(self):
+        return '<%s: %s>' % (self.__class__.__name__, self.nonterminal_or_string)
+
 
 class NFAState(object):
     def __init__(self, from_rule):

parso/python/errors.py

@@ -52,7 +52,7 @@ def _is_future_import(import_from):
     # It looks like a __future__ import that is relative is still a future
     # import. That feels kind of odd, but whatever.
     # if import_from.level != 0:
-        # return False
+    #     return False
     from_names = import_from.get_from_names()
     return [n.value for n in from_names] == ['__future__']
 
@@ -94,19 +94,32 @@ def _is_future_import_first(import_from):
 
 
 def _iter_definition_exprs_from_lists(exprlist):
-    for child in exprlist.children[::2]:
-        if child.type == 'atom' and child.children[0] in ('(', '['):
-            testlist_comp = child.children[0]
-            if testlist_comp.type == 'testlist_comp':
-                for expr in _iter_definition_exprs_from_lists(testlist_comp):
-                    yield expr
-                continue
+    def check_expr(child):
+        if child.type == 'atom':
+            if child.children[0] == '(':
+                testlist_comp = child.children[1]
+                if testlist_comp.type == 'testlist_comp':
+                    for expr in _iter_definition_exprs_from_lists(testlist_comp):
+                        yield expr
+                    return
+                else:
+                    # It's a paren that doesn't do anything, like 1 + (1)
+                    for c in check_expr(testlist_comp):
+                        yield c
+                    return
             elif child.children[0] == '[':
                 yield testlist_comp
-                continue
-
+                return
         yield child
 
+    if exprlist.type in _STAR_EXPR_PARENTS:
+        for child in exprlist.children[::2]:
+            for c in check_expr(child):  # Python 2 sucks
+                yield c
+    else:
+        for c in check_expr(exprlist):  # Python 2 sucks
+            yield c
+
 
 def _get_expr_stmt_definition_exprs(expr_stmt):
     exprs = []
@@ -120,8 +133,6 @@ def _get_expr_stmt_definition_exprs(expr_stmt):
 
 def _get_for_stmt_definition_exprs(for_stmt):
     exprlist = for_stmt.children[1]
-    if exprlist.type != 'exprlist':
-        return [exprlist]
     return list(_iter_definition_exprs_from_lists(exprlist))
 
 
@@ -478,38 +489,38 @@ class _StringChecks(SyntaxRule):
     message = "bytes can only contain ASCII literal characters."
 
     def is_issue(self, leaf):
-            string_prefix = leaf.string_prefix.lower()
-            if 'b' in string_prefix \
-                    and self._normalizer.version >= (3, 0) \
-                    and any(c for c in leaf.value if ord(c) > 127):
-                # b'ä'
-                return True
+        string_prefix = leaf.string_prefix.lower()
+        if 'b' in string_prefix \
+                and self._normalizer.version >= (3, 0) \
+                and any(c for c in leaf.value if ord(c) > 127):
+            # b'ä'
+            return True
 
-            if 'r' not in string_prefix:
-                # Raw strings don't need to be checked if they have proper
-                # escaping.
-                is_bytes = self._normalizer.version < (3, 0)
-                if 'b' in string_prefix:
-                    is_bytes = True
-                if 'u' in string_prefix:
-                    is_bytes = False
+        if 'r' not in string_prefix:
+            # Raw strings don't need to be checked if they have proper
+            # escaping.
+            is_bytes = self._normalizer.version < (3, 0)
+            if 'b' in string_prefix:
+                is_bytes = True
+            if 'u' in string_prefix:
+                is_bytes = False
 
-                payload = leaf._get_payload()
-                if is_bytes:
-                    payload = payload.encode('utf-8')
-                    func = codecs.escape_decode
-                else:
-                    func = codecs.unicode_escape_decode
+            payload = leaf._get_payload()
+            if is_bytes:
+                payload = payload.encode('utf-8')
+                func = codecs.escape_decode
+            else:
+                func = codecs.unicode_escape_decode
 
-                try:
-                    with warnings.catch_warnings():
-                        # The warnings from parsing strings are not relevant.
-                        warnings.filterwarnings('ignore')
-                        func(payload)
-                except UnicodeDecodeError as e:
-                    self.add_issue(leaf, message='(unicode error) ' + str(e))
-                except ValueError as e:
-                    self.add_issue(leaf, message='(value error) ' + str(e))
+            try:
+                with warnings.catch_warnings():
+                    # The warnings from parsing strings are not relevant.
+                    warnings.filterwarnings('ignore')
+                    func(payload)
+            except UnicodeDecodeError as e:
+                self.add_issue(leaf, message='(unicode error) ' + str(e))
+            except ValueError as e:
+                self.add_issue(leaf, message='(value error) ' + str(e))
 
 
 @ErrorFinder.register_rule(value='*')
@@ -586,7 +597,7 @@ class _TrailingImportComma(SyntaxRule):
     message = "trailing comma not allowed without surrounding parentheses"
 
     def is_issue(self, node):
-        if node.children[-1] == ',':
+        if node.children[-1] == ',' and node.parent.children[-1] != ')':
             return True
 
 
@@ -883,7 +894,7 @@ class _FStringRule(SyntaxRule):
 
 
 class _CheckAssignmentRule(SyntaxRule):
-    def _check_assignment(self, node, is_deletion=False):
+    def _check_assignment(self, node, is_deletion=False, is_namedexpr=False):
         error = None
         type_ = node.type
         if type_ == 'lambdef':
@@ -907,9 +918,9 @@ class _CheckAssignmentRule(SyntaxRule):
                         # This is not a comprehension, they were handled
                         # further above.
                         for child in second.children[::2]:
-                            self._check_assignment(child, is_deletion)
+                            self._check_assignment(child, is_deletion, is_namedexpr)
                     else:  # Everything handled, must be useless brackets.
-                        self._check_assignment(second, is_deletion)
+                        self._check_assignment(second, is_deletion, is_namedexpr)
         elif type_ == 'keyword':
             if self._normalizer.version < (3, 8):
                 error = 'keyword'
@@ -939,17 +950,24 @@ class _CheckAssignmentRule(SyntaxRule):
                 assert trailer.type == 'trailer'
                 if trailer.children[0] == '(':
                     error = 'function call'
+                elif is_namedexpr and trailer.children[0] == '[':
+                    error = 'subscript'
+                elif is_namedexpr and trailer.children[0] == '.':
+                    error = 'attribute'
         elif type_ in ('testlist_star_expr', 'exprlist', 'testlist'):
             for child in node.children[::2]:
-                self._check_assignment(child, is_deletion)
+                self._check_assignment(child, is_deletion, is_namedexpr)
         elif ('expr' in type_ and type_ != 'star_expr'  # is a substring
               or '_test' in type_
               or type_ in ('term', 'factor')):
             error = 'operator'
 
         if error is not None:
-            cannot = "can't" if self._normalizer.version < (3, 8) else "cannot"
-            message = ' '.join([cannot, "delete" if is_deletion else "assign to", error])
+            if is_namedexpr:
+                message = 'cannot use named assignment with %s' % error
+            else:
+                cannot = "can't" if self._normalizer.version < (3, 8) else "cannot"
+                message = ' '.join([cannot, "delete" if is_deletion else "assign to", error])
             self.add_issue(node, message=message)
 
 
@@ -959,7 +977,6 @@ class _CompForRule(_CheckAssignmentRule):
 
     def is_issue(self, node):
         expr_list = node.children[1]
-        print(expr_list)
         if expr_list.type != 'expr_list':  # Already handled.
             self._check_assignment(expr_list)
 
@@ -1009,3 +1026,71 @@ class _ForStmtRule(_CheckAssignmentRule):
         expr_list = for_stmt.children[1]
         if expr_list.type != 'expr_list':  # Already handled.
             self._check_assignment(expr_list)
+
+
+@ErrorFinder.register_rule(type='namedexpr_test')
+class _NamedExprRule(_CheckAssignmentRule):
+    # namedexpr_test: test [':=' test]
+
+    def is_issue(self, namedexpr_test):
+        # assigned name
+        first = namedexpr_test.children[0]
+
+        def search_namedexpr_in_comp_for(node):
+            while True:
+                parent = node.parent
+                if parent is None:
+                    return parent
+                if parent.type == 'sync_comp_for' and parent.children[3] == node:
+                    return parent
+                node = parent
+
+        if search_namedexpr_in_comp_for(namedexpr_test):
+            # [i+1 for i in (i := range(5))]
+            # [i+1 for i in (j := range(5))]
+            # [i+1 for i in (lambda: (j := range(5)))()]
+            message = 'assignment expression cannot be used in a comprehension iterable expression'
+            self.add_issue(namedexpr_test, message=message)
+
+        # defined names
+        exprlist = list()
+
+        def process_comp_for(comp_for):
+            if comp_for.type == 'sync_comp_for':
+                comp = comp_for
+            elif comp_for.type == 'comp_for':
+                comp = comp_for.children[1]
+            exprlist.extend(_get_for_stmt_definition_exprs(comp))
+
+        def search_all_comp_ancestors(node):
+            has_ancestors = False
+            while True:
+                node = search_ancestor(node, 'testlist_comp', 'dictorsetmaker')
+                if node is None:
+                    break
+                for child in node.children:
+                    if child.type in _COMP_FOR_TYPES:
+                        process_comp_for(child)
+                        has_ancestors = True
+                        break
+            return has_ancestors
+
+        # check assignment expressions in comprehensions
+        search_all = search_all_comp_ancestors(namedexpr_test)
+        if search_all:
+            if self._normalizer.context.node.type == 'classdef':
+                message = 'assignment expression within a comprehension ' \
+                          'cannot be used in a class body'
+                self.add_issue(namedexpr_test, message=message)
+
+            namelist = [expr.value for expr in exprlist if expr.type == 'name']
+            if first.type == 'name' and first.value in namelist:
+                # [i := 0 for i, j in range(5)]
+                # [[(i := i) for j in range(5)] for i in range(5)]
+                # [i for i, j in range(5) if True or (i := 1)]
+                # [False and (i := 0) for i, j in range(5)]
+                message = 'assignment expression cannot rebind ' \
+                          'comprehension iteration variable %r' % first.value
+                self.add_issue(namedexpr_test, message=message)
+
+        self._check_assignment(first, is_namedexpr=True)

parso/python/grammar39.txt

@@ -0,0 +1,171 @@
+# Grammar for Python
+
+# NOTE WELL: You should also follow all the steps listed at
+# https://devguide.python.org/grammar/
+
+# Start symbols for the grammar:
+#       single_input is a single interactive statement;
+#       file_input is a module or sequence of commands read from an input file;
+#       eval_input is the input for the eval() functions.
+# NB: compound_stmt in single_input is followed by extra NEWLINE!
+single_input: NEWLINE | simple_stmt | compound_stmt NEWLINE
+file_input: (NEWLINE | stmt)* ENDMARKER
+eval_input: testlist NEWLINE* ENDMARKER
+
+decorator: '@' dotted_name [ '(' [arglist] ')' ] NEWLINE
+decorators: decorator+
+decorated: decorators (classdef | funcdef | async_funcdef)
+
+async_funcdef: 'async' funcdef
+funcdef: 'def' NAME parameters ['->' test] ':' suite
+
+parameters: '(' [typedargslist] ')'
+typedargslist: (
+  (tfpdef ['=' test] (',' tfpdef ['=' test])* ',' '/' [',' [ tfpdef ['=' test] (
+        ',' tfpdef ['=' test])* ([',' [
+        '*' [tfpdef] (',' tfpdef ['=' test])* [',' ['**' tfpdef [',']]]
+      | '**' tfpdef [',']]])
+  | '*' [tfpdef] (',' tfpdef ['=' test])* ([',' ['**' tfpdef [',']]])
+  | '**' tfpdef [',']]] )
+|  (tfpdef ['=' test] (',' tfpdef ['=' test])* [',' [
+        '*' [tfpdef] (',' tfpdef ['=' test])* [',' ['**' tfpdef [',']]]
+      | '**' tfpdef [',']]]
+  | '*' [tfpdef] (',' tfpdef ['=' test])* [',' ['**' tfpdef [',']]]
+  | '**' tfpdef [','])
+)
+tfpdef: NAME [':' test]
+varargslist: vfpdef ['=' test ](',' vfpdef ['=' test])* ',' '/' [',' [ (vfpdef ['=' test] (',' vfpdef ['=' test])* [',' [
+        '*' [vfpdef] (',' vfpdef ['=' test])* [',' ['**' vfpdef [',']]]
+      | '**' vfpdef [',']]]
+  | '*' [vfpdef] (',' vfpdef ['=' test])* [',' ['**' vfpdef [',']]]
+  | '**' vfpdef [',']) ]] | (vfpdef ['=' test] (',' vfpdef ['=' test])* [',' [
+        '*' [vfpdef] (',' vfpdef ['=' test])* [',' ['**' vfpdef [',']]]
+      | '**' vfpdef [',']]]
+  | '*' [vfpdef] (',' vfpdef ['=' test])* [',' ['**' vfpdef [',']]]
+  | '**' vfpdef [',']
+)
+vfpdef: NAME
+
+stmt: simple_stmt | compound_stmt
+simple_stmt: small_stmt (';' small_stmt)* [';'] NEWLINE
+small_stmt: (expr_stmt | del_stmt | pass_stmt | flow_stmt |
+             import_stmt | global_stmt | nonlocal_stmt | assert_stmt)
+expr_stmt: testlist_star_expr (annassign | augassign (yield_expr|testlist) |
+                     ('=' (yield_expr|testlist_star_expr))*)
+annassign: ':' test ['=' test]
+testlist_star_expr: (test|star_expr) (',' (test|star_expr))* [',']
+augassign: ('+=' | '-=' | '*=' | '@=' | '/=' | '%=' | '&=' | '|=' | '^=' |
+            '<<=' | '>>=' | '**=' | '//=')
+# For normal and annotated assignments, additional restrictions enforced by the interpreter
+del_stmt: 'del' exprlist
+pass_stmt: 'pass'
+flow_stmt: break_stmt | continue_stmt | return_stmt | raise_stmt | yield_stmt
+break_stmt: 'break'
+continue_stmt: 'continue'
+return_stmt: 'return' [testlist_star_expr]
+yield_stmt: yield_expr
+raise_stmt: 'raise' [test ['from' test]]
+import_stmt: import_name | import_from
+import_name: 'import' dotted_as_names
+# note below: the ('.' | '...') is necessary because '...' is tokenized as ELLIPSIS
+import_from: ('from' (('.' | '...')* dotted_name | ('.' | '...')+)
+              'import' ('*' | '(' import_as_names ')' | import_as_names))
+import_as_name: NAME ['as' NAME]
+dotted_as_name: dotted_name ['as' NAME]
+import_as_names: import_as_name (',' import_as_name)* [',']
+dotted_as_names: dotted_as_name (',' dotted_as_name)*
+dotted_name: NAME ('.' NAME)*
+global_stmt: 'global' NAME (',' NAME)*
+nonlocal_stmt: 'nonlocal' NAME (',' NAME)*
+assert_stmt: 'assert' test [',' test]
+
+compound_stmt: if_stmt | while_stmt | for_stmt | try_stmt | with_stmt | funcdef | classdef | decorated | async_stmt
+async_stmt: 'async' (funcdef | with_stmt | for_stmt)
+if_stmt: 'if' namedexpr_test ':' suite ('elif' namedexpr_test ':' suite)* ['else' ':' suite]
+while_stmt: 'while' namedexpr_test ':' suite ['else' ':' suite]
+for_stmt: 'for' exprlist 'in' testlist ':' suite ['else' ':' suite]
+try_stmt: ('try' ':' suite
+           ((except_clause ':' suite)+
+            ['else' ':' suite]
+            ['finally' ':' suite] |
+           'finally' ':' suite))
+with_stmt: 'with' with_item (',' with_item)*  ':' suite
+with_item: test ['as' expr]
+# NB compile.c makes sure that the default except clause is last
+except_clause: 'except' [test ['as' NAME]]
+suite: simple_stmt | NEWLINE INDENT stmt+ DEDENT
+
+namedexpr_test: test [':=' test]
+test: or_test ['if' or_test 'else' test] | lambdef
+test_nocond: or_test | lambdef_nocond
+lambdef: 'lambda' [varargslist] ':' test
+lambdef_nocond: 'lambda' [varargslist] ':' test_nocond
+or_test: and_test ('or' and_test)*
+and_test: not_test ('and' not_test)*
+not_test: 'not' not_test | comparison
+comparison: expr (comp_op expr)*
+# <> isn't actually a valid comparison operator in Python. It's here for the
+# sake of a __future__ import described in PEP 401 (which really works :-)
+comp_op: '<'|'>'|'=='|'>='|'<='|'<>'|'!='|'in'|'not' 'in'|'is'|'is' 'not'
+star_expr: '*' expr
+expr: xor_expr ('|' xor_expr)*
+xor_expr: and_expr ('^' and_expr)*
+and_expr: shift_expr ('&' shift_expr)*
+shift_expr: arith_expr (('<<'|'>>') arith_expr)*
+arith_expr: term (('+'|'-') term)*
+term: factor (('*'|'@'|'/'|'%'|'//') factor)*
+factor: ('+'|'-'|'~') factor | power
+power: atom_expr ['**' factor]
+atom_expr: ['await'] atom trailer*
+atom: ('(' [yield_expr|testlist_comp] ')' |
+       '[' [testlist_comp] ']' |
+       '{' [dictorsetmaker] '}' |
+       NAME | NUMBER | strings | '...' | 'None' | 'True' | 'False')
+testlist_comp: (namedexpr_test|star_expr) ( comp_for | (',' (namedexpr_test|star_expr))* [','] )
+trailer: '(' [arglist] ')' | '[' subscriptlist ']' | '.' NAME
+subscriptlist: subscript (',' subscript)* [',']
+subscript: test | [test] ':' [test] [sliceop]
+sliceop: ':' [test]
+exprlist: (expr|star_expr) (',' (expr|star_expr))* [',']
+testlist: test (',' test)* [',']
+dictorsetmaker: ( ((test ':' test | '**' expr)
+                   (comp_for | (',' (test ':' test | '**' expr))* [','])) |
+                  ((test | star_expr)
+                   (comp_for | (',' (test | star_expr))* [','])) )
+
+classdef: 'class' NAME ['(' [arglist] ')'] ':' suite
+
+arglist: argument (',' argument)*  [',']
+
+# The reason that keywords are test nodes instead of NAME is that using NAME
+# results in an ambiguity. ast.c makes sure it's a NAME.
+# "test '=' test" is really "keyword '=' test", but we have no such token.
+# These need to be in a single rule to avoid grammar that is ambiguous
+# to our LL(1) parser. Even though 'test' includes '*expr' in star_expr,
+# we explicitly match '*' here, too, to give it proper precedence.
+# Illegal combinations and orderings are blocked in ast.c:
+# multiple (test comp_for) arguments are blocked; keyword unpackings
+# that precede iterable unpackings are blocked; etc.
+argument: ( test [comp_for] |
+            test ':=' test |
+            test '=' test |
+            '**' test |
+            '*' test )
+
+comp_iter: comp_for | comp_if
+sync_comp_for: 'for' exprlist 'in' or_test [comp_iter]
+comp_for: ['async'] sync_comp_for
+comp_if: 'if' test_nocond [comp_iter]
+
+# not used in grammar, but may appear in "node" passed from Parser to Compiler
+encoding_decl: NAME
+
+yield_expr: 'yield' [yield_arg]
+yield_arg: 'from' test | testlist_star_expr
+
+strings: (STRING | fstring)+
+fstring: FSTRING_START fstring_content* FSTRING_END
+fstring_content: FSTRING_STRING | fstring_expr
+fstring_conversion: '!' NAME
+fstring_expr: '{' testlist ['='] [ fstring_conversion ] [ fstring_format_spec ] '}'
+fstring_format_spec: ':' fstring_content*

parso/python/tokenize.py

@@ -23,6 +23,9 @@ from parso._compatibility import py_version
 from parso.utils import split_lines
 
 
+# Maximum code point of Unicode 6.0: 0x10ffff (1,114,111)
+MAX_UNICODE = '\U0010ffff'
+
 STRING = PythonTokenTypes.STRING
 NAME = PythonTokenTypes.NAME
 NUMBER = PythonTokenTypes.NUMBER
@@ -51,8 +54,13 @@ if py_version >= 30:
     # Python 3 has str.isidentifier() to check if a char is a valid identifier
     is_identifier = str.isidentifier
 else:
-    namechars = string.ascii_letters + '_'
-    is_identifier = lambda s: s in namechars
+    # Python 2 doesn't, but it's not that important anymore and if you tokenize
+    # Python 2 code with this, it's still ok. It's just that parsing Python 3
+    # code with this function is not 100% correct.
+    # This just means that Python 2 code matches a few identifiers too much,
+    # but that doesn't really matter.
+    def is_identifier(s):
+        return True
 
 
 def group(*choices, **kwargs):
@@ -118,9 +126,9 @@ def _get_token_collection(version_info):
         return result
 
 
-fstring_string_single_line = _compile(r'(?:[^{}\r\n]+|\{\{|\}\})+')
+fstring_string_single_line = _compile(r'(?:\{\{|\}\}|\\(?:\r\n?|\n)|[^{}\r\n])+')
 fstring_string_multi_line = _compile(r'(?:[^{}]+|\{\{|\}\})+')
-fstring_format_spec_single_line = _compile(r'[^{}\r\n]+')
+fstring_format_spec_single_line = _compile(r'(?:\\(?:\r\n?|\n)|[^{}\r\n])+')
 fstring_format_spec_multi_line = _compile(r'[^{}]+')
 
 
@@ -130,7 +138,16 @@ def _create_token_collection(version_info):
     Whitespace = r'[ \f\t]*'
     whitespace = _compile(Whitespace)
     Comment = r'#[^\r\n]*'
-    Name = r'\w+'
+    # Python 2 is pretty much not working properly anymore, we just ignore
+    # parsing unicode properly, which is fine, I guess.
+    if version_info[0] == 2:
+        Name = r'([A-Za-z_0-9]+)'
+    elif sys.version_info[0] == 2:
+        # Unfortunately the regex engine cannot deal with the regex below, so
+        # just use this one.
+        Name = r'(\w+)'
+    else:
+        Name = u'([A-Za-z_0-9\u0080-' + MAX_UNICODE + ']+)'
 
     if version_info >= (3, 6):
         Hexnumber = r'0[xX](?:_?[0-9a-fA-F])+'
@@ -297,17 +314,19 @@ class FStringNode(object):
 
 def _close_fstring_if_necessary(fstring_stack, string, start_pos, additional_prefix):
     for fstring_stack_index, node in enumerate(fstring_stack):
-        if string.startswith(node.quote):
+        lstripped_string = string.lstrip()
+        len_lstrip = len(string) - len(lstripped_string)
+        if lstripped_string.startswith(node.quote):
             token = PythonToken(
                 FSTRING_END,
                 node.quote,
                 start_pos,
-                prefix=additional_prefix,
+                prefix=additional_prefix+string[:len_lstrip],
             )
             additional_prefix = ''
             assert not node.previous_lines
             del fstring_stack[fstring_stack_index:]
-            return token, '', len(node.quote)
+            return token, '', len(node.quote) + len_lstrip
     return None, additional_prefix, 0
 
 
@@ -340,7 +359,9 @@ def _find_fstring_string(endpats, fstring_stack, line, lnum, pos):
 
     new_pos = pos
     new_pos += len(string)
-    if allow_multiline and (string.endswith('\n') or string.endswith('\r')):
+    # even if allow_multiline is False, we still need to check for trailing
+    # newlines, because a single-line f-string can contain line continuations
+    if string.endswith('\n') or string.endswith('\r'):
         tos.previous_lines += string
         string = ''
     else:
@@ -463,8 +484,20 @@ def tokenize_lines(lines, version_info, start_pos=(1, 0)):
                     yield fstring_end_token
                     continue
 
-            pseudomatch = pseudo_token.match(line, pos)
-            if not pseudomatch:                             # scan for tokens
+            # in an f-string, match until the end of the string
+            if fstring_stack:
+                string_line = line
+                for fstring_stack_node in fstring_stack:
+                    quote = fstring_stack_node.quote
+                    end_match = endpats[quote].match(line, pos)
+                    if end_match is not None:
+                        end_match_string = end_match.group(0)
+                        if len(end_match_string) - len(quote) + pos < len(string_line):
+                            string_line = line[:pos] + end_match_string[:-len(quote)]
+                pseudomatch = pseudo_token.match(string_line, pos)
+            else:
+                pseudomatch = pseudo_token.match(line, pos)
+            if not pseudomatch:  # scan for tokens
                 match = whitespace.match(line, pos)
                 if pos == 0:
                     for t in dedent_if_necessary(match.end()):
@@ -510,6 +543,24 @@ def tokenize_lines(lines, version_info, start_pos=(1, 0)):
             if (initial in numchars or                      # ordinary number
                     (initial == '.' and token != '.' and token != '...')):
                 yield PythonToken(NUMBER, token, spos, prefix)
+            elif pseudomatch.group(3) is not None:            # ordinary name
+                if token in always_break_tokens:
+                    fstring_stack[:] = []
+                    paren_level = 0
+                    # We only want to dedent if the token is on a new line.
+                    if re.match(r'[ \f\t]*$', line[:start]):
+                        while True:
+                            indent = indents.pop()
+                            if indent > start:
+                                yield PythonToken(DEDENT, '', spos, '')
+                            else:
+                                indents.append(indent)
+                                break
+                if is_identifier(token):
+                    yield PythonToken(NAME, token, spos, prefix)
+                else:
+                    for t in _split_illegal_unicode_name(token, spos, prefix):
+                        yield t  # yield from Python 2
             elif initial in '\r\n':
                 if any(not f.allow_multiline() for f in fstring_stack):
                     # Would use fstring_stack.clear, but that's not available
@@ -523,7 +574,12 @@ def tokenize_lines(lines, version_info, start_pos=(1, 0)):
                 new_line = True
             elif initial == '#':  # Comments
                 assert not token.endswith("\n")
-                additional_prefix = prefix + token
+                if fstring_stack and fstring_stack[-1].is_in_expr():
+                    # `#` is not allowed in f-string expressions
+                    yield PythonToken(ERRORTOKEN, initial, spos, prefix)
+                    pos = start + 1
+                else:
+                    additional_prefix = prefix + token
             elif token in triple_quoted:
                 endprog = endpats[token]
                 endmatch = endprog.match(line, pos)
@@ -564,20 +620,6 @@ def tokenize_lines(lines, version_info, start_pos=(1, 0)):
             elif token in fstring_pattern_map:  # The start of an fstring.
                 fstring_stack.append(FStringNode(fstring_pattern_map[token]))
                 yield PythonToken(FSTRING_START, token, spos, prefix)
-            elif is_identifier(initial):                      # ordinary name
-                if token in always_break_tokens:
-                    fstring_stack[:] = []
-                    paren_level = 0
-                    # We only want to dedent if the token is on a new line.
-                    if re.match(r'[ \f\t]*$', line[:start]):
-                        while True:
-                            indent = indents.pop()
-                            if indent > start:
-                                yield PythonToken(DEDENT, '', spos, '')
-                            else:
-                                indents.append(indent)
-                                break
-                yield PythonToken(NAME, token, spos, prefix)
             elif initial == '\\' and line[start:] in ('\\\n', '\\\r\n', '\\\r'):  # continued stmt
                 additional_prefix += prefix + line[start:]
                 break
@@ -593,10 +635,13 @@ def tokenize_lines(lines, version_info, start_pos=(1, 0)):
                     else:
                         if paren_level:
                             paren_level -= 1
-                elif token == ':' and fstring_stack \
+                elif token.startswith(':') and fstring_stack \
                         and fstring_stack[-1].parentheses_count \
                         - fstring_stack[-1].format_spec_count == 1:
+                    # `:` and `:=` both count
                     fstring_stack[-1].format_spec_count += 1
+                    token = ':'
+                    pos = start + 1
 
                 yield PythonToken(OP, token, spos, prefix)
 
@@ -613,6 +658,39 @@ def tokenize_lines(lines, version_info, start_pos=(1, 0)):
     yield PythonToken(ENDMARKER, '', end_pos, additional_prefix)
 
 
+def _split_illegal_unicode_name(token, start_pos, prefix):
+    def create_token():
+        return PythonToken(ERRORTOKEN if is_illegal else NAME, found, pos, prefix)
+
+    found = ''
+    is_illegal = False
+    pos = start_pos
+    for i, char in enumerate(token):
+        if is_illegal:
+            if is_identifier(char):
+                yield create_token()
+                found = char
+                is_illegal = False
+                prefix = ''
+                pos = start_pos[0], start_pos[1] + i
+            else:
+                found += char
+        else:
+            new_found = found + char
+            if is_identifier(new_found):
+                found = new_found
+            else:
+                if found:
+                    yield create_token()
+                    prefix = ''
+                    pos = start_pos[0], start_pos[1] + i
+                found = char
+                is_illegal = True
+
+    if found:
+        yield create_token()
+
+
 if __name__ == "__main__":
     if len(sys.argv) >= 2:
         path = sys.argv[1]

parso/python/tree.py

@@ -43,7 +43,10 @@ Parser Tree Classes
 """
 
 import re
-from collections import Mapping
+try:
+    from collections.abc import Mapping
+except ImportError:
+    from collections import Mapping
 
 from parso._compatibility import utf8_repr, unicode
 from parso.tree import Node, BaseNode, Leaf, ErrorNode, ErrorLeaf, \
@@ -197,25 +200,22 @@ class Name(_LeafWithoutNewlines):
         return "<%s: %s@%s,%s>" % (type(self).__name__, self.value,
                                    self.line, self.column)
 
-    def is_definition(self):
+    def is_definition(self, include_setitem=False):
         """
         Returns True if the name is being defined.
         """
-        return self.get_definition() is not None
+        return self.get_definition(include_setitem=include_setitem) is not None
 
-    def get_definition(self, import_name_always=False):
+    def get_definition(self, import_name_always=False, include_setitem=False):
         """
-        Returns None if there's on definition for a name.
+        Returns None if there's no definition for a name.
 
-        :param import_name_alway: Specifies if an import name is always a
+        :param import_name_always: Specifies if an import name is always a
             definition. Normally foo in `from foo import bar` is not a
             definition.
         """
         node = self.parent
         type_ = node.type
-        if type_ in ('power', 'atom_expr'):
-            # In `self.x = 3` self is not a definition, but x is.
-            return None
 
         if type_ in ('funcdef', 'classdef'):
             if self == node.name:
@@ -234,7 +234,7 @@ class Name(_LeafWithoutNewlines):
             if node.type == 'suite':
                 return None
             if node.type in _GET_DEFINITION_TYPES:
-                if self in node.get_defined_names():
+                if self in node.get_defined_names(include_setitem):
                     return node
                 if import_name_always and node.type in _IMPORTS:
                     return node
@@ -772,8 +772,8 @@ class ForStmt(Flow):
         """
         return self.children[3]
 
-    def get_defined_names(self):
-        return _defined_names(self.children[1])
+    def get_defined_names(self, include_setitem=False):
+        return _defined_names(self.children[1], include_setitem)
 
 
 class TryStmt(Flow):
@@ -796,7 +796,7 @@ class WithStmt(Flow):
     type = 'with_stmt'
     __slots__ = ()
 
-    def get_defined_names(self):
+    def get_defined_names(self, include_setitem=False):
         """
         Returns the a list of `Name` that the with statement defines. The
         defined names are set after `as`.
@@ -805,7 +805,7 @@ class WithStmt(Flow):
         for with_item in self.children[1:-2:2]:
             # Check with items for 'as' names.
             if with_item.type == 'with_item':
-                names += _defined_names(with_item.children[2])
+                names += _defined_names(with_item.children[2], include_setitem)
         return names
 
     def get_test_node_from_name(self, name):
@@ -846,7 +846,7 @@ class ImportFrom(Import):
     type = 'import_from'
     __slots__ = ()
 
-    def get_defined_names(self):
+    def get_defined_names(self, include_setitem=False):
         """
         Returns the a list of `Name` that the import defines. The
         defined names are set after `import` or in case an alias - `as` - is
@@ -917,7 +917,7 @@ class ImportName(Import):
     type = 'import_name'
     __slots__ = ()
 
-    def get_defined_names(self):
+    def get_defined_names(self, include_setitem=False):
         """
         Returns the a list of `Name` that the import defines. The defined names
         is always the first name after `import` or in case an alias - `as` - is
@@ -1018,7 +1018,7 @@ class YieldExpr(PythonBaseNode):
     __slots__ = ()
 
 
-def _defined_names(current):
+def _defined_names(current, include_setitem):
     """
     A helper function to find the defined names in statements, for loops and
     list comprehensions.
@@ -1026,14 +1026,22 @@ def _defined_names(current):
     names = []
     if current.type in ('testlist_star_expr', 'testlist_comp', 'exprlist', 'testlist'):
         for child in current.children[::2]:
-            names += _defined_names(child)
+            names += _defined_names(child, include_setitem)
     elif current.type in ('atom', 'star_expr'):
-        names += _defined_names(current.children[1])
+        names += _defined_names(current.children[1], include_setitem)
     elif current.type in ('power', 'atom_expr'):
         if current.children[-2] != '**':  # Just if there's no operation
             trailer = current.children[-1]
             if trailer.children[0] == '.':
                 names.append(trailer.children[1])
+            elif trailer.children[0] == '[' and include_setitem:
+                for node in current.children[-2::-1]:
+                    if node.type == 'trailer':
+                        names.append(node.children[1])
+                        break
+                    if node.type == 'name':
+                        names.append(node)
+                        break
     else:
         names.append(current)
     return names
@@ -1043,18 +1051,18 @@ class ExprStmt(PythonBaseNode, DocstringMixin):
     type = 'expr_stmt'
     __slots__ = ()
 
-    def get_defined_names(self):
+    def get_defined_names(self, include_setitem=False):
         """
         Returns a list of `Name` defined before the `=` sign.
         """
         names = []
         if self.children[1].type == 'annassign':
-            names = _defined_names(self.children[0])
+            names = _defined_names(self.children[0], include_setitem)
         return [
             name
             for i in range(0, len(self.children) - 2, 2)
             if '=' in self.children[i + 1].value
-            for name in _defined_names(self.children[i])
+            for name in _defined_names(self.children[i], include_setitem)
         ] + names
 
     def get_rhs(self):
@@ -1147,7 +1155,7 @@ class Param(PythonBaseNode):
         else:
             return self._tfpdef()
 
-    def get_defined_names(self):
+    def get_defined_names(self, include_setitem=False):
         return [self.name]
 
     @property
@@ -1205,12 +1213,12 @@ class SyncCompFor(PythonBaseNode):
     type = 'sync_comp_for'
     __slots__ = ()
 
-    def get_defined_names(self):
+    def get_defined_names(self, include_setitem=False):
         """
         Returns the a list of `Name` that the comprehension defines.
         """
         # allow async for
-        return _defined_names(self.children[1])
+        return _defined_names(self.children[1], include_setitem)
 
 
 # This is simply here so an older Jedi version can work with this new parso

test/failing_examples.py

@@ -319,3 +319,35 @@ if sys.version_info[:2] < (3, 8):
                     continue
             '''),  # 'continue' not supported inside 'finally' clause"
     ]
+
+if sys.version_info[:2] >= (3, 8):
+    # assignment expressions from issue#89
+    FAILING_EXAMPLES += [
+        # Case 2
+        '(lambda: x := 1)',
+        '((lambda: x) := 1)',
+        # Case 3
+        '(a[i] := x)',
+        '((a[i]) := x)',
+        '(a(i) := x)',
+        # Case 4
+        '(a.b := c)',
+        '[(i.i:= 0) for ((i), j) in range(5)]',
+        # Case 5
+        '[i:= 0 for i, j in range(5)]',
+        '[(i:= 0) for ((i), j) in range(5)]',
+        '[(i:= 0) for ((i), j), in range(5)]',
+        '[(i:= 0) for ((i), j.i), in range(5)]',
+        '[[(i:= i) for j in range(5)] for i in range(5)]',
+        '[i for i, j in range(5) if True or (i:= 1)]',
+        '[False and (i:= 0) for i, j in range(5)]',
+        # Case 6
+        '[i+1 for i in (i:= range(5))]',
+        '[i+1 for i in (j:= range(5))]',
+        '[i+1 for i in (lambda: (j:= range(5)))()]',
+        # Case 7
+        'class Example:\n [(j := i) for i in range(5)]',
+        # Not in that issue
+        '(await a := x)',
+        '((await a) := x)',
+    ]

test/test_diff_parser.py

@@ -974,10 +974,12 @@ def test_random_unicode_characters(differ):
     Those issues were all found with the fuzzer.
     """
     differ.initialize('')
-    differ.parse(u'\x1dĔBϞɛˁşʑ˳˻ȣſéÎ\x90̕ȟòwʘ\x1dĔBϞɛˁşʑ˳˻ȣſéÎ', parsers=1, expect_error_leaves=True)
+    differ.parse(u'\x1dĔBϞɛˁşʑ˳˻ȣſéÎ\x90̕ȟòwʘ\x1dĔBϞɛˁşʑ˳˻ȣſéÎ', parsers=1,
+                 expect_error_leaves=True)
     differ.parse(u'\r\r', parsers=1)
     differ.parse(u"˟Ę\x05À\r   rúƣ@\x8a\x15r()\n", parsers=1, expect_error_leaves=True)
-    differ.parse(u'a\ntaǁ\rGĒōns__\n\nb', parsers=1)
+    differ.parse(u'a\ntaǁ\rGĒōns__\n\nb', parsers=1,
+                 expect_error_leaves=sys.version_info[0] == 2)
     s = '        if not (self, "_fi\x02\x0e\x08\n\nle"):'
     differ.parse(s, parsers=1, expect_error_leaves=True)
     differ.parse('')

test/test_fstring.py

@@ -12,33 +12,57 @@ def grammar():
 
 @pytest.mark.parametrize(
     'code', [
-        '{1}',
-        '{1:}',
-        '',
-        '{1!a}',
-        '{1!a:1}',
-        '{1:1}',
-        '{1:1.{32}}',
-        '{1::>4}',
-        '{foo} {bar}',
-        '{x:{y}}',
-        '{x:{y:}}',
-        '{x:{y:1}}',
+        # simple cases
+        'f"{1}"',
+        'f"""{1}"""',
+        'f"{foo} {bar}"',
+
+        # empty string
+        'f""',
+        'f""""""',
+
+        # empty format specifier is okay
+        'f"{1:}"',
+
+        # use of conversion options
+        'f"{1!a}"',
+        'f"{1!a:1}"',
+
+        # format specifiers
+        'f"{1:1}"',
+        'f"{1:1.{32}}"',
+        'f"{1::>4}"',
+        'f"{x:{y}}"',
+        'f"{x:{y:}}"',
+        'f"{x:{y:1}}"',
 
         # Escapes
-        '{{}}',
-        '{{{1}}}',
-        '{{{1}',
-        '1{{2{{3',
-        '}}',
+        'f"{{}}"',
+        'f"{{{1}}}"',
+        'f"{{{1}"',
+        'f"1{{2{{3"',
+        'f"}}"',
 
         # New Python 3.8 syntax f'{a=}'
-        '{a=}',
-        '{a()=}',
+        'f"{a=}"',
+        'f"{a()=}"',
+
+        # multiline f-string
+        'f"""abc\ndef"""',
+        'f"""abc{\n123}def"""',
+
+        # a line continuation inside of an fstring_string
+        'f"abc\\\ndef"',
+        'f"\\\n{123}\\\n"',
+
+        # a line continuation inside of an fstring_expr
+        'f"{\\\n123}"',
+
+        # a line continuation inside of an format spec
+        'f"{123:.2\\\nf}"',
     ]
 )
 def test_valid(code, grammar):
-    code = 'f"""%s"""' % code
     module = grammar.parse(code, error_recovery=False)
     fstring = module.children[0]
     assert fstring.type == 'fstring'
@@ -47,23 +71,34 @@ def test_valid(code, grammar):
 
 @pytest.mark.parametrize(
     'code', [
-        '}',
-        '{',
-        '{1!{a}}',
-        '{!{a}}',
-        '{}',
-        '{:}',
-        '{:}}}',
-        '{:1}',
-        '{!:}',
-        '{!}',
-        '{!a}',
-        '{1:{}}',
-        '{1:{:}}',
+        # an f-string can't contain unmatched curly braces
+        'f"}"',
+        'f"{"',
+        'f"""}"""',
+        'f"""{"""',
+
+        # invalid conversion characters
+        'f"{1!{a}}"',
+        'f"{!{a}}"',
+
+        # The curly braces must contain an expression
+        'f"{}"',
+        'f"{:}"',
+        'f"{:}}}"',
+        'f"{:1}"',
+        'f"{!:}"',
+        'f"{!}"',
+        'f"{!a}"',
+
+        # invalid (empty) format specifiers
+        'f"{1:{}}"',
+        'f"{1:{:}}"',
+
+        # a newline without a line continuation inside a single-line string
+        'f"abc\ndef"',
     ]
 )
 def test_invalid(code, grammar):
-    code = 'f"""%s"""' % code
     with pytest.raises(ParserSyntaxError):
         grammar.parse(code, error_recovery=False)
 
@@ -95,6 +130,7 @@ def test_tokenize_start_pos(code, positions):
             """),
         'f"foo',
         'f"""foo',
+        'f"abc\ndef"',
     ]
 )
 def test_roundtrip(grammar, code):

test/test_parser_tree.py

@@ -180,3 +180,45 @@ def top_function_three():
 
     r = get_raise_stmts(code, 2) #  Lists inside try-catch
     assert len(list(r)) == 2
+
+
+@pytest.mark.parametrize(
+    'code, name_index, is_definition, include_setitem', [
+        ('x = 3', 0, True, False),
+        ('x.y = 3', 0, False, False),
+        ('x.y = 3', 1, True, False),
+        ('x.y = u.v = z', 0, False, False),
+        ('x.y = u.v = z', 1, True, False),
+        ('x.y = u.v = z', 2, False, False),
+        ('x.y = u.v, w = z', 3, True, False),
+        ('x.y = u.v, w = z', 4, True, False),
+        ('x.y = u.v, w = z', 5, False, False),
+
+        ('x, y = z', 0, True, False),
+        ('x, y = z', 1, True, False),
+        ('x, y = z', 2, False, False),
+        ('x, y = z', 2, False, False),
+        ('x[0], y = z', 2, False, False),
+        ('x[0] = z', 0, False, False),
+        ('x[0], y = z', 0, False, False),
+        ('x[0], y = z', 2, False, True),
+        ('x[0] = z', 0, True, True),
+        ('x[0], y = z', 0, True, True),
+        ('x: int = z', 0, True, False),
+        ('x: int = z', 1, False, False),
+        ('x: int = z', 2, False, False),
+        ('x: int', 0, True, False),
+        ('x: int', 1, False, False),
+    ]
+)
+def test_is_definition(code, name_index, is_definition, include_setitem):
+    module = parse(code, version='3.8')
+    name = module.get_first_leaf()
+    while True:
+        if name.type == 'name':
+            if name_index == 0:
+                break
+            name_index -= 1
+        name = name.get_next_leaf()
+
+    assert name.is_definition(include_setitem=include_setitem) == is_definition

test/test_pgen2.py

@@ -292,12 +292,22 @@ def test_left_recursion():
         generate_grammar('foo: foo NAME\n', tokenize.PythonTokenTypes)
 
 
-def test_ambiguities():
-    with pytest.raises(ValueError, match='ambiguous'):
-        generate_grammar('foo: bar | baz\nbar: NAME\nbaz: NAME\n', tokenize.PythonTokenTypes)
-
-    with pytest.raises(ValueError, match='ambiguous'):
-        generate_grammar('''foo: bar | baz\nbar: 'x'\nbaz: "x"\n''', tokenize.PythonTokenTypes)
-
-    with pytest.raises(ValueError, match='ambiguous'):
-        generate_grammar('''foo: bar | 'x'\nbar: 'x'\n''', tokenize.PythonTokenTypes)
+@pytest.mark.parametrize(
+    'grammar, error_match', [
+        ['foo: bar | baz\nbar: NAME\nbaz: NAME\n',
+         r"foo is ambiguous.*given a TokenType\(NAME\).*bar or baz"],
+        ['''foo: bar | baz\nbar: 'x'\nbaz: "x"\n''',
+         r"foo is ambiguous.*given a ReservedString\(x\).*bar or baz"],
+        ['''foo: bar | 'x'\nbar: 'x'\n''',
+         r"foo is ambiguous.*given a ReservedString\(x\).*bar or foo"],
+        # An ambiguity with the second (not the first) child of a production
+        ['outer: "a" [inner] "b" "c"\ninner: "b" "c" [inner]\n',
+         r"outer is ambiguous.*given a ReservedString\(b\).*inner or outer"],
+        # An ambiguity hidden by a level of indirection (middle)
+        ['outer: "a" [middle] "b" "c"\nmiddle: inner\ninner: "b" "c" [inner]\n',
+         r"outer is ambiguous.*given a ReservedString\(b\).*middle or outer"],
+    ]
+)
+def test_ambiguities(grammar, error_match):
+    with pytest.raises(ValueError, match=error_match):
+        generate_grammar(grammar, tokenize.PythonTokenTypes)

test/test_python_errors.py

@@ -293,6 +293,19 @@ def test_valid_fstrings(code):
     assert not _get_error_list(code, version='3.6')
 
 
+@pytest.mark.parametrize(
+    'code', [
+        'a = (b := 1)',
+        '[x4 := x ** 5 for x in range(7)]',
+        '[total := total + v for v in range(10)]',
+        'while chunk := file.read(2):\n pass',
+        'numbers = [y := math.factorial(x), y**2, y**3]',
+    ]
+)
+def test_valid_namedexpr(code):
+    assert not _get_error_list(code, version='3.8')
+
+
 @pytest.mark.parametrize(
     ('code', 'message'), [
         ("f'{1+}'", ('invalid syntax')),
@@ -307,3 +320,15 @@ def test_invalid_fstrings(code, message):
     """
     error, = _get_error_list(code, version='3.6')
     assert message in error.message
+
+
+@pytest.mark.parametrize(
+    'code', [
+        "from foo import (\nbar,\n rab,\n)",
+        "from foo import (bar, rab, )",
+    ]
+)
+def test_trailing_comma(code):
+    errors = _get_error_list(code)
+    assert not errors
+

test/test_tokenize.py

@@ -1,5 +1,6 @@
 # -*- coding: utf-8    # This file contains Unicode characters.
 
+import sys
 from textwrap import dedent
 
 import pytest
@@ -16,6 +17,7 @@ from parso.python.tokenize import PythonToken
 NAME = PythonTokenTypes.NAME
 NEWLINE = PythonTokenTypes.NEWLINE
 STRING = PythonTokenTypes.STRING
+NUMBER = PythonTokenTypes.NUMBER
 INDENT = PythonTokenTypes.INDENT
 DEDENT = PythonTokenTypes.DEDENT
 ERRORTOKEN = PythonTokenTypes.ERRORTOKEN
@@ -140,7 +142,7 @@ def test_identifier_contains_unicode():
     else:
         # Unicode tokens in Python 2 seem to be identified as operators.
         # They will be ignored in the parser, that's ok.
-        assert unicode_token[0] == OP
+        assert unicode_token[0] == ERRORTOKEN
 
 
 def test_quoted_strings():
@@ -228,16 +230,29 @@ def test_endmarker_end_pos():
     check('a\\')
 
 
+xfail_py2 = dict(marks=[pytest.mark.xfail(sys.version_info[0] == 2, reason='Python 2')])
+
+
 @pytest.mark.parametrize(
     ('code', 'types'), [
+        # Indentation
         (' foo', [INDENT, NAME, DEDENT]),
         ('  foo\n bar', [INDENT, NAME, NEWLINE, ERROR_DEDENT, NAME, DEDENT]),
         ('  foo\n bar \n baz', [INDENT, NAME, NEWLINE, ERROR_DEDENT, NAME,
                                 NEWLINE, ERROR_DEDENT, NAME, DEDENT]),
         (' foo\nbar', [INDENT, NAME, NEWLINE, DEDENT, NAME]),
+
+        # Name stuff
+        ('1foo1', [NUMBER, NAME]),
+        pytest.param(
+            u'மெல்லினம்', [NAME],
+            **xfail_py2),
+        pytest.param(u'²', [ERRORTOKEN], **xfail_py2),
+        pytest.param(u'ä²ö', [NAME, ERRORTOKEN, NAME], **xfail_py2),
+        pytest.param(u'ää²¹öö', [NAME, ERRORTOKEN, NAME], **xfail_py2),
     ]
 )
-def test_indentation(code, types):
+def test_token_types(code, types):
     actual_types = [t.type for t in _get_token_list(code)]
     assert actual_types == types + [ENDMARKER]
 
@@ -330,15 +345,72 @@ def test_backslash():
         ('f" "{}', [FSTRING_START, FSTRING_STRING, FSTRING_END, OP, OP]),
         (r'f"\""', [FSTRING_START, FSTRING_STRING, FSTRING_END]),
         (r'f"\""', [FSTRING_START, FSTRING_STRING, FSTRING_END]),
+
+        # format spec
         (r'f"Some {x:.2f}{y}"', [FSTRING_START, FSTRING_STRING, OP, NAME, OP,
                                  FSTRING_STRING, OP, OP, NAME, OP, FSTRING_END]),
+
+        # multiline f-string
+        ('f"""abc\ndef"""', [FSTRING_START, FSTRING_STRING, FSTRING_END]),
+        ('f"""abc{\n123}def"""', [
+            FSTRING_START, FSTRING_STRING, OP, NUMBER, OP, FSTRING_STRING,
+            FSTRING_END
+        ]),
+
+        # a line continuation inside of an fstring_string
+        ('f"abc\\\ndef"', [
+            FSTRING_START, FSTRING_STRING, FSTRING_END
+        ]),
+        ('f"\\\n{123}\\\n"', [
+            FSTRING_START, FSTRING_STRING, OP, NUMBER, OP, FSTRING_STRING,
+            FSTRING_END
+        ]),
+
+        # a line continuation inside of an fstring_expr
+        ('f"{\\\n123}"', [FSTRING_START, OP, NUMBER, OP, FSTRING_END]),
+
+        # a line continuation inside of an format spec
+        ('f"{123:.2\\\nf}"', [
+            FSTRING_START, OP, NUMBER, OP, FSTRING_STRING, OP, FSTRING_END
+        ]),
+
+        # a newline without a line continuation inside a single-line string is
+        # wrong, and will generate an ERRORTOKEN
+        ('f"abc\ndef"', [
+            FSTRING_START, FSTRING_STRING, NEWLINE, NAME, ERRORTOKEN
+        ]),
+
+        # a more complex example
         (r'print(f"Some {x:.2f}a{y}")', [
             NAME, OP, FSTRING_START, FSTRING_STRING, OP, NAME, OP,
             FSTRING_STRING, OP, FSTRING_STRING, OP, NAME, OP, FSTRING_END, OP
         ]),
-
+        # issue #86, a string-like in an f-string expression
+        ('f"{ ""}"', [
+            FSTRING_START, OP, FSTRING_END, STRING
+        ]),
+        ('f"{ f""}"', [
+            FSTRING_START, OP, NAME, FSTRING_END, STRING
+        ]),
     ]
 )
 def test_fstring(code, types, version_ge_py36):
     actual_types = [t.type for t in _get_token_list(code, version_ge_py36)]
     assert types + [ENDMARKER] == actual_types
+
+
+@pytest.mark.parametrize(
+    ('code', 'types'), [
+        # issue #87, `:=` in the outest paratheses should be tokenized
+        # as a format spec marker and part of the format
+        ('f"{x:=10}"', [
+            FSTRING_START, OP, NAME, OP, FSTRING_STRING, OP, FSTRING_END
+        ]),
+        ('f"{(x:=10)}"', [
+            FSTRING_START, OP, OP, NAME, OP, NUMBER, OP, OP, FSTRING_END
+        ]),
+    ]
+)
+def test_fstring_assignment_expression(code, types, version_ge_py38):
+    actual_types = [t.type for t in _get_token_list(code, version_ge_py38)]
+    assert types + [ENDMARKER] == actual_types

tox.ini

@@ -1,9 +1,10 @@
 [tox]
-envlist = {py26,py27,py33,py34,py35,py36,py37}
+envlist = {py26,py27,py33,py34,py35,py36,py37,py38}
 [testenv]
 extras = testing
 deps =
     py26,py33: pytest>=3.0.7,<3.3
+    py27,py34: pytest<3.3
     py26,py33: setuptools<37
     coverage: coverage
 setenv =