Show More
@@ -1,762 +1,763 | |||||
1 | #!/usr/bin/env python |
|
1 | #!/usr/bin/env python | |
2 |
|
2 | |||
3 | from __future__ import absolute_import, print_function |
|
3 | from __future__ import absolute_import, print_function | |
4 |
|
4 | |||
5 | import ast |
|
5 | import ast | |
6 | import collections |
|
6 | import collections | |
7 | import os |
|
7 | import os | |
8 | import re |
|
8 | import re | |
9 | import sys |
|
9 | import sys | |
10 |
|
10 | |||
11 | # Import a minimal set of stdlib modules needed for list_stdlib_modules() |
|
11 | # Import a minimal set of stdlib modules needed for list_stdlib_modules() | |
12 | # to work when run from a virtualenv. The modules were chosen empirically |
|
12 | # to work when run from a virtualenv. The modules were chosen empirically | |
13 | # so that the return value matches the return value without virtualenv. |
|
13 | # so that the return value matches the return value without virtualenv. | |
14 | if True: # disable lexical sorting checks |
|
14 | if True: # disable lexical sorting checks | |
15 | try: |
|
15 | try: | |
16 | import BaseHTTPServer as basehttpserver |
|
16 | import BaseHTTPServer as basehttpserver | |
17 | except ImportError: |
|
17 | except ImportError: | |
18 | basehttpserver = None |
|
18 | basehttpserver = None | |
19 | import zlib |
|
19 | import zlib | |
20 |
|
20 | |||
21 | # Whitelist of modules that symbols can be directly imported from. |
|
21 | # Whitelist of modules that symbols can be directly imported from. | |
22 | allowsymbolimports = ( |
|
22 | allowsymbolimports = ( | |
23 | '__future__', |
|
23 | '__future__', | |
|
24 | 'mercurial', | |||
24 | 'mercurial.hgweb.common', |
|
25 | 'mercurial.hgweb.common', | |
25 | 'mercurial.hgweb.request', |
|
26 | 'mercurial.hgweb.request', | |
26 | 'mercurial.i18n', |
|
27 | 'mercurial.i18n', | |
27 | 'mercurial.node', |
|
28 | 'mercurial.node', | |
28 | # for cffi modules to re-export pure functions |
|
29 | # for cffi modules to re-export pure functions | |
29 | 'mercurial.pure.base85', |
|
30 | 'mercurial.pure.base85', | |
30 | 'mercurial.pure.bdiff', |
|
31 | 'mercurial.pure.bdiff', | |
31 | 'mercurial.pure.diffhelpers', |
|
32 | 'mercurial.pure.diffhelpers', | |
32 | 'mercurial.pure.mpatch', |
|
33 | 'mercurial.pure.mpatch', | |
33 | 'mercurial.pure.osutil', |
|
34 | 'mercurial.pure.osutil', | |
34 | 'mercurial.pure.parsers', |
|
35 | 'mercurial.pure.parsers', | |
35 | ) |
|
36 | ) | |
36 |
|
37 | |||
37 | # Whitelist of symbols that can be directly imported. |
|
38 | # Whitelist of symbols that can be directly imported. | |
38 | directsymbols = ( |
|
39 | directsymbols = ( | |
39 | 'demandimport', |
|
40 | 'demandimport', | |
40 | ) |
|
41 | ) | |
41 |
|
42 | |||
42 | # Modules that must be aliased because they are commonly confused with |
|
43 | # Modules that must be aliased because they are commonly confused with | |
43 | # common variables and can create aliasing and readability issues. |
|
44 | # common variables and can create aliasing and readability issues. | |
44 | requirealias = { |
|
45 | requirealias = { | |
45 | 'ui': 'uimod', |
|
46 | 'ui': 'uimod', | |
46 | } |
|
47 | } | |
47 |
|
48 | |||
48 | def usingabsolute(root): |
|
49 | def usingabsolute(root): | |
49 | """Whether absolute imports are being used.""" |
|
50 | """Whether absolute imports are being used.""" | |
50 | if sys.version_info[0] >= 3: |
|
51 | if sys.version_info[0] >= 3: | |
51 | return True |
|
52 | return True | |
52 |
|
53 | |||
53 | for node in ast.walk(root): |
|
54 | for node in ast.walk(root): | |
54 | if isinstance(node, ast.ImportFrom): |
|
55 | if isinstance(node, ast.ImportFrom): | |
55 | if node.module == '__future__': |
|
56 | if node.module == '__future__': | |
56 | for n in node.names: |
|
57 | for n in node.names: | |
57 | if n.name == 'absolute_import': |
|
58 | if n.name == 'absolute_import': | |
58 | return True |
|
59 | return True | |
59 |
|
60 | |||
60 | return False |
|
61 | return False | |
61 |
|
62 | |||
62 | def walklocal(root): |
|
63 | def walklocal(root): | |
63 | """Recursively yield all descendant nodes but not in a different scope""" |
|
64 | """Recursively yield all descendant nodes but not in a different scope""" | |
64 | todo = collections.deque(ast.iter_child_nodes(root)) |
|
65 | todo = collections.deque(ast.iter_child_nodes(root)) | |
65 | yield root, False |
|
66 | yield root, False | |
66 | while todo: |
|
67 | while todo: | |
67 | node = todo.popleft() |
|
68 | node = todo.popleft() | |
68 | newscope = isinstance(node, ast.FunctionDef) |
|
69 | newscope = isinstance(node, ast.FunctionDef) | |
69 | if not newscope: |
|
70 | if not newscope: | |
70 | todo.extend(ast.iter_child_nodes(node)) |
|
71 | todo.extend(ast.iter_child_nodes(node)) | |
71 | yield node, newscope |
|
72 | yield node, newscope | |
72 |
|
73 | |||
73 | def dotted_name_of_path(path): |
|
74 | def dotted_name_of_path(path): | |
74 | """Given a relative path to a source file, return its dotted module name. |
|
75 | """Given a relative path to a source file, return its dotted module name. | |
75 |
|
76 | |||
76 | >>> dotted_name_of_path('mercurial/error.py') |
|
77 | >>> dotted_name_of_path('mercurial/error.py') | |
77 | 'mercurial.error' |
|
78 | 'mercurial.error' | |
78 | >>> dotted_name_of_path('zlibmodule.so') |
|
79 | >>> dotted_name_of_path('zlibmodule.so') | |
79 | 'zlib' |
|
80 | 'zlib' | |
80 | """ |
|
81 | """ | |
81 | parts = path.replace(os.sep, '/').split('/') |
|
82 | parts = path.replace(os.sep, '/').split('/') | |
82 | parts[-1] = parts[-1].split('.', 1)[0] # remove .py and .so and .ARCH.so |
|
83 | parts[-1] = parts[-1].split('.', 1)[0] # remove .py and .so and .ARCH.so | |
83 | if parts[-1].endswith('module'): |
|
84 | if parts[-1].endswith('module'): | |
84 | parts[-1] = parts[-1][:-6] |
|
85 | parts[-1] = parts[-1][:-6] | |
85 | return '.'.join(parts) |
|
86 | return '.'.join(parts) | |
86 |
|
87 | |||
87 | def fromlocalfunc(modulename, localmods): |
|
88 | def fromlocalfunc(modulename, localmods): | |
88 | """Get a function to examine which locally defined module the |
|
89 | """Get a function to examine which locally defined module the | |
89 | target source imports via a specified name. |
|
90 | target source imports via a specified name. | |
90 |
|
91 | |||
91 | `modulename` is an `dotted_name_of_path()`-ed source file path, |
|
92 | `modulename` is an `dotted_name_of_path()`-ed source file path, | |
92 | which may have `.__init__` at the end of it, of the target source. |
|
93 | which may have `.__init__` at the end of it, of the target source. | |
93 |
|
94 | |||
94 | `localmods` is a set of absolute `dotted_name_of_path()`-ed source file |
|
95 | `localmods` is a set of absolute `dotted_name_of_path()`-ed source file | |
95 | paths of locally defined (= Mercurial specific) modules. |
|
96 | paths of locally defined (= Mercurial specific) modules. | |
96 |
|
97 | |||
97 | This function assumes that module names not existing in |
|
98 | This function assumes that module names not existing in | |
98 | `localmods` are from the Python standard library. |
|
99 | `localmods` are from the Python standard library. | |
99 |
|
100 | |||
100 | This function returns the function, which takes `name` argument, |
|
101 | This function returns the function, which takes `name` argument, | |
101 | and returns `(absname, dottedpath, hassubmod)` tuple if `name` |
|
102 | and returns `(absname, dottedpath, hassubmod)` tuple if `name` | |
102 | matches against locally defined module. Otherwise, it returns |
|
103 | matches against locally defined module. Otherwise, it returns | |
103 | False. |
|
104 | False. | |
104 |
|
105 | |||
105 | It is assumed that `name` doesn't have `.__init__`. |
|
106 | It is assumed that `name` doesn't have `.__init__`. | |
106 |
|
107 | |||
107 | `absname` is an absolute module name of specified `name` |
|
108 | `absname` is an absolute module name of specified `name` | |
108 | (e.g. "hgext.convert"). This can be used to compose prefix for sub |
|
109 | (e.g. "hgext.convert"). This can be used to compose prefix for sub | |
109 | modules or so. |
|
110 | modules or so. | |
110 |
|
111 | |||
111 | `dottedpath` is a `dotted_name_of_path()`-ed source file path |
|
112 | `dottedpath` is a `dotted_name_of_path()`-ed source file path | |
112 | (e.g. "hgext.convert.__init__") of `name`. This is used to look |
|
113 | (e.g. "hgext.convert.__init__") of `name`. This is used to look | |
113 | module up in `localmods` again. |
|
114 | module up in `localmods` again. | |
114 |
|
115 | |||
115 | `hassubmod` is whether it may have sub modules under it (for |
|
116 | `hassubmod` is whether it may have sub modules under it (for | |
116 | convenient, even though this is also equivalent to "absname != |
|
117 | convenient, even though this is also equivalent to "absname != | |
117 | dottednpath") |
|
118 | dottednpath") | |
118 |
|
119 | |||
119 | >>> localmods = {'foo.__init__', 'foo.foo1', |
|
120 | >>> localmods = {'foo.__init__', 'foo.foo1', | |
120 | ... 'foo.bar.__init__', 'foo.bar.bar1', |
|
121 | ... 'foo.bar.__init__', 'foo.bar.bar1', | |
121 | ... 'baz.__init__', 'baz.baz1'} |
|
122 | ... 'baz.__init__', 'baz.baz1'} | |
122 | >>> fromlocal = fromlocalfunc('foo.xxx', localmods) |
|
123 | >>> fromlocal = fromlocalfunc('foo.xxx', localmods) | |
123 | >>> # relative |
|
124 | >>> # relative | |
124 | >>> fromlocal('foo1') |
|
125 | >>> fromlocal('foo1') | |
125 | ('foo.foo1', 'foo.foo1', False) |
|
126 | ('foo.foo1', 'foo.foo1', False) | |
126 | >>> fromlocal('bar') |
|
127 | >>> fromlocal('bar') | |
127 | ('foo.bar', 'foo.bar.__init__', True) |
|
128 | ('foo.bar', 'foo.bar.__init__', True) | |
128 | >>> fromlocal('bar.bar1') |
|
129 | >>> fromlocal('bar.bar1') | |
129 | ('foo.bar.bar1', 'foo.bar.bar1', False) |
|
130 | ('foo.bar.bar1', 'foo.bar.bar1', False) | |
130 | >>> # absolute |
|
131 | >>> # absolute | |
131 | >>> fromlocal('baz') |
|
132 | >>> fromlocal('baz') | |
132 | ('baz', 'baz.__init__', True) |
|
133 | ('baz', 'baz.__init__', True) | |
133 | >>> fromlocal('baz.baz1') |
|
134 | >>> fromlocal('baz.baz1') | |
134 | ('baz.baz1', 'baz.baz1', False) |
|
135 | ('baz.baz1', 'baz.baz1', False) | |
135 | >>> # unknown = maybe standard library |
|
136 | >>> # unknown = maybe standard library | |
136 | >>> fromlocal('os') |
|
137 | >>> fromlocal('os') | |
137 | False |
|
138 | False | |
138 | >>> fromlocal(None, 1) |
|
139 | >>> fromlocal(None, 1) | |
139 | ('foo', 'foo.__init__', True) |
|
140 | ('foo', 'foo.__init__', True) | |
140 | >>> fromlocal('foo1', 1) |
|
141 | >>> fromlocal('foo1', 1) | |
141 | ('foo.foo1', 'foo.foo1', False) |
|
142 | ('foo.foo1', 'foo.foo1', False) | |
142 | >>> fromlocal2 = fromlocalfunc('foo.xxx.yyy', localmods) |
|
143 | >>> fromlocal2 = fromlocalfunc('foo.xxx.yyy', localmods) | |
143 | >>> fromlocal2(None, 2) |
|
144 | >>> fromlocal2(None, 2) | |
144 | ('foo', 'foo.__init__', True) |
|
145 | ('foo', 'foo.__init__', True) | |
145 | >>> fromlocal2('bar2', 1) |
|
146 | >>> fromlocal2('bar2', 1) | |
146 | False |
|
147 | False | |
147 | >>> fromlocal2('bar', 2) |
|
148 | >>> fromlocal2('bar', 2) | |
148 | ('foo.bar', 'foo.bar.__init__', True) |
|
149 | ('foo.bar', 'foo.bar.__init__', True) | |
149 | """ |
|
150 | """ | |
150 | if not isinstance(modulename, str): |
|
151 | if not isinstance(modulename, str): | |
151 | modulename = modulename.decode('ascii') |
|
152 | modulename = modulename.decode('ascii') | |
152 | prefix = '.'.join(modulename.split('.')[:-1]) |
|
153 | prefix = '.'.join(modulename.split('.')[:-1]) | |
153 | if prefix: |
|
154 | if prefix: | |
154 | prefix += '.' |
|
155 | prefix += '.' | |
155 | def fromlocal(name, level=0): |
|
156 | def fromlocal(name, level=0): | |
156 | # name is false value when relative imports are used. |
|
157 | # name is false value when relative imports are used. | |
157 | if not name: |
|
158 | if not name: | |
158 | # If relative imports are used, level must not be absolute. |
|
159 | # If relative imports are used, level must not be absolute. | |
159 | assert level > 0 |
|
160 | assert level > 0 | |
160 | candidates = ['.'.join(modulename.split('.')[:-level])] |
|
161 | candidates = ['.'.join(modulename.split('.')[:-level])] | |
161 | else: |
|
162 | else: | |
162 | if not level: |
|
163 | if not level: | |
163 | # Check relative name first. |
|
164 | # Check relative name first. | |
164 | candidates = [prefix + name, name] |
|
165 | candidates = [prefix + name, name] | |
165 | else: |
|
166 | else: | |
166 | candidates = ['.'.join(modulename.split('.')[:-level]) + |
|
167 | candidates = ['.'.join(modulename.split('.')[:-level]) + | |
167 | '.' + name] |
|
168 | '.' + name] | |
168 |
|
169 | |||
169 | for n in candidates: |
|
170 | for n in candidates: | |
170 | if n in localmods: |
|
171 | if n in localmods: | |
171 | return (n, n, False) |
|
172 | return (n, n, False) | |
172 | dottedpath = n + '.__init__' |
|
173 | dottedpath = n + '.__init__' | |
173 | if dottedpath in localmods: |
|
174 | if dottedpath in localmods: | |
174 | return (n, dottedpath, True) |
|
175 | return (n, dottedpath, True) | |
175 | return False |
|
176 | return False | |
176 | return fromlocal |
|
177 | return fromlocal | |
177 |
|
178 | |||
178 | def populateextmods(localmods): |
|
179 | def populateextmods(localmods): | |
179 | """Populate C extension modules based on pure modules""" |
|
180 | """Populate C extension modules based on pure modules""" | |
180 | newlocalmods = set(localmods) |
|
181 | newlocalmods = set(localmods) | |
181 | for n in localmods: |
|
182 | for n in localmods: | |
182 | if n.startswith('mercurial.pure.'): |
|
183 | if n.startswith('mercurial.pure.'): | |
183 | m = n[len('mercurial.pure.'):] |
|
184 | m = n[len('mercurial.pure.'):] | |
184 | newlocalmods.add('mercurial.cext.' + m) |
|
185 | newlocalmods.add('mercurial.cext.' + m) | |
185 | newlocalmods.add('mercurial.cffi._' + m) |
|
186 | newlocalmods.add('mercurial.cffi._' + m) | |
186 | return newlocalmods |
|
187 | return newlocalmods | |
187 |
|
188 | |||
188 | def list_stdlib_modules(): |
|
189 | def list_stdlib_modules(): | |
189 | """List the modules present in the stdlib. |
|
190 | """List the modules present in the stdlib. | |
190 |
|
191 | |||
191 | >>> py3 = sys.version_info[0] >= 3 |
|
192 | >>> py3 = sys.version_info[0] >= 3 | |
192 | >>> mods = set(list_stdlib_modules()) |
|
193 | >>> mods = set(list_stdlib_modules()) | |
193 | >>> 'BaseHTTPServer' in mods or py3 |
|
194 | >>> 'BaseHTTPServer' in mods or py3 | |
194 | True |
|
195 | True | |
195 |
|
196 | |||
196 | os.path isn't really a module, so it's missing: |
|
197 | os.path isn't really a module, so it's missing: | |
197 |
|
198 | |||
198 | >>> 'os.path' in mods |
|
199 | >>> 'os.path' in mods | |
199 | False |
|
200 | False | |
200 |
|
201 | |||
201 | sys requires special treatment, because it's baked into the |
|
202 | sys requires special treatment, because it's baked into the | |
202 | interpreter, but it should still appear: |
|
203 | interpreter, but it should still appear: | |
203 |
|
204 | |||
204 | >>> 'sys' in mods |
|
205 | >>> 'sys' in mods | |
205 | True |
|
206 | True | |
206 |
|
207 | |||
207 | >>> 'collections' in mods |
|
208 | >>> 'collections' in mods | |
208 | True |
|
209 | True | |
209 |
|
210 | |||
210 | >>> 'cStringIO' in mods or py3 |
|
211 | >>> 'cStringIO' in mods or py3 | |
211 | True |
|
212 | True | |
212 |
|
213 | |||
213 | >>> 'cffi' in mods |
|
214 | >>> 'cffi' in mods | |
214 | True |
|
215 | True | |
215 | """ |
|
216 | """ | |
216 | for m in sys.builtin_module_names: |
|
217 | for m in sys.builtin_module_names: | |
217 | yield m |
|
218 | yield m | |
218 | # These modules only exist on windows, but we should always |
|
219 | # These modules only exist on windows, but we should always | |
219 | # consider them stdlib. |
|
220 | # consider them stdlib. | |
220 | for m in ['msvcrt', '_winreg']: |
|
221 | for m in ['msvcrt', '_winreg']: | |
221 | yield m |
|
222 | yield m | |
222 | yield 'builtins' # python3 only |
|
223 | yield 'builtins' # python3 only | |
223 | for m in 'fcntl', 'grp', 'pwd', 'termios': # Unix only |
|
224 | for m in 'fcntl', 'grp', 'pwd', 'termios': # Unix only | |
224 | yield m |
|
225 | yield m | |
225 | for m in 'cPickle', 'datetime': # in Python (not C) on PyPy |
|
226 | for m in 'cPickle', 'datetime': # in Python (not C) on PyPy | |
226 | yield m |
|
227 | yield m | |
227 | for m in ['cffi']: |
|
228 | for m in ['cffi']: | |
228 | yield m |
|
229 | yield m | |
229 | stdlib_prefixes = {sys.prefix, sys.exec_prefix} |
|
230 | stdlib_prefixes = {sys.prefix, sys.exec_prefix} | |
230 | # We need to supplement the list of prefixes for the search to work |
|
231 | # We need to supplement the list of prefixes for the search to work | |
231 | # when run from within a virtualenv. |
|
232 | # when run from within a virtualenv. | |
232 | for mod in (basehttpserver, zlib): |
|
233 | for mod in (basehttpserver, zlib): | |
233 | if mod is None: |
|
234 | if mod is None: | |
234 | continue |
|
235 | continue | |
235 | try: |
|
236 | try: | |
236 | # Not all module objects have a __file__ attribute. |
|
237 | # Not all module objects have a __file__ attribute. | |
237 | filename = mod.__file__ |
|
238 | filename = mod.__file__ | |
238 | except AttributeError: |
|
239 | except AttributeError: | |
239 | continue |
|
240 | continue | |
240 | dirname = os.path.dirname(filename) |
|
241 | dirname = os.path.dirname(filename) | |
241 | for prefix in stdlib_prefixes: |
|
242 | for prefix in stdlib_prefixes: | |
242 | if dirname.startswith(prefix): |
|
243 | if dirname.startswith(prefix): | |
243 | # Then this directory is redundant. |
|
244 | # Then this directory is redundant. | |
244 | break |
|
245 | break | |
245 | else: |
|
246 | else: | |
246 | stdlib_prefixes.add(dirname) |
|
247 | stdlib_prefixes.add(dirname) | |
247 | for libpath in sys.path: |
|
248 | for libpath in sys.path: | |
248 | # We want to walk everything in sys.path that starts with |
|
249 | # We want to walk everything in sys.path that starts with | |
249 | # something in stdlib_prefixes. |
|
250 | # something in stdlib_prefixes. | |
250 | if not any(libpath.startswith(p) for p in stdlib_prefixes): |
|
251 | if not any(libpath.startswith(p) for p in stdlib_prefixes): | |
251 | continue |
|
252 | continue | |
252 | for top, dirs, files in os.walk(libpath): |
|
253 | for top, dirs, files in os.walk(libpath): | |
253 | for i, d in reversed(list(enumerate(dirs))): |
|
254 | for i, d in reversed(list(enumerate(dirs))): | |
254 | if (not os.path.exists(os.path.join(top, d, '__init__.py')) |
|
255 | if (not os.path.exists(os.path.join(top, d, '__init__.py')) | |
255 | or top == libpath and d in ('hgdemandimport', 'hgext', |
|
256 | or top == libpath and d in ('hgdemandimport', 'hgext', | |
256 | 'mercurial')): |
|
257 | 'mercurial')): | |
257 | del dirs[i] |
|
258 | del dirs[i] | |
258 | for name in files: |
|
259 | for name in files: | |
259 | if not name.endswith(('.py', '.so', '.pyc', '.pyo', '.pyd')): |
|
260 | if not name.endswith(('.py', '.so', '.pyc', '.pyo', '.pyd')): | |
260 | continue |
|
261 | continue | |
261 | if name.startswith('__init__.py'): |
|
262 | if name.startswith('__init__.py'): | |
262 | full_path = top |
|
263 | full_path = top | |
263 | else: |
|
264 | else: | |
264 | full_path = os.path.join(top, name) |
|
265 | full_path = os.path.join(top, name) | |
265 | rel_path = full_path[len(libpath) + 1:] |
|
266 | rel_path = full_path[len(libpath) + 1:] | |
266 | mod = dotted_name_of_path(rel_path) |
|
267 | mod = dotted_name_of_path(rel_path) | |
267 | yield mod |
|
268 | yield mod | |
268 |
|
269 | |||
269 | stdlib_modules = set(list_stdlib_modules()) |
|
270 | stdlib_modules = set(list_stdlib_modules()) | |
270 |
|
271 | |||
271 | def imported_modules(source, modulename, f, localmods, ignore_nested=False): |
|
272 | def imported_modules(source, modulename, f, localmods, ignore_nested=False): | |
272 | """Given the source of a file as a string, yield the names |
|
273 | """Given the source of a file as a string, yield the names | |
273 | imported by that file. |
|
274 | imported by that file. | |
274 |
|
275 | |||
275 | Args: |
|
276 | Args: | |
276 | source: The python source to examine as a string. |
|
277 | source: The python source to examine as a string. | |
277 | modulename: of specified python source (may have `__init__`) |
|
278 | modulename: of specified python source (may have `__init__`) | |
278 | localmods: set of locally defined module names (may have `__init__`) |
|
279 | localmods: set of locally defined module names (may have `__init__`) | |
279 | ignore_nested: If true, import statements that do not start in |
|
280 | ignore_nested: If true, import statements that do not start in | |
280 | column zero will be ignored. |
|
281 | column zero will be ignored. | |
281 |
|
282 | |||
282 | Returns: |
|
283 | Returns: | |
283 | A list of absolute module names imported by the given source. |
|
284 | A list of absolute module names imported by the given source. | |
284 |
|
285 | |||
285 | >>> f = 'foo/xxx.py' |
|
286 | >>> f = 'foo/xxx.py' | |
286 | >>> modulename = 'foo.xxx' |
|
287 | >>> modulename = 'foo.xxx' | |
287 | >>> localmods = {'foo.__init__': True, |
|
288 | >>> localmods = {'foo.__init__': True, | |
288 | ... 'foo.foo1': True, 'foo.foo2': True, |
|
289 | ... 'foo.foo1': True, 'foo.foo2': True, | |
289 | ... 'foo.bar.__init__': True, 'foo.bar.bar1': True, |
|
290 | ... 'foo.bar.__init__': True, 'foo.bar.bar1': True, | |
290 | ... 'baz.__init__': True, 'baz.baz1': True } |
|
291 | ... 'baz.__init__': True, 'baz.baz1': True } | |
291 | >>> # standard library (= not locally defined ones) |
|
292 | >>> # standard library (= not locally defined ones) | |
292 | >>> sorted(imported_modules( |
|
293 | >>> sorted(imported_modules( | |
293 | ... 'from stdlib1 import foo, bar; import stdlib2', |
|
294 | ... 'from stdlib1 import foo, bar; import stdlib2', | |
294 | ... modulename, f, localmods)) |
|
295 | ... modulename, f, localmods)) | |
295 | [] |
|
296 | [] | |
296 | >>> # relative importing |
|
297 | >>> # relative importing | |
297 | >>> sorted(imported_modules( |
|
298 | >>> sorted(imported_modules( | |
298 | ... 'import foo1; from bar import bar1', |
|
299 | ... 'import foo1; from bar import bar1', | |
299 | ... modulename, f, localmods)) |
|
300 | ... modulename, f, localmods)) | |
300 | ['foo.bar.bar1', 'foo.foo1'] |
|
301 | ['foo.bar.bar1', 'foo.foo1'] | |
301 | >>> sorted(imported_modules( |
|
302 | >>> sorted(imported_modules( | |
302 | ... 'from bar.bar1 import name1, name2, name3', |
|
303 | ... 'from bar.bar1 import name1, name2, name3', | |
303 | ... modulename, f, localmods)) |
|
304 | ... modulename, f, localmods)) | |
304 | ['foo.bar.bar1'] |
|
305 | ['foo.bar.bar1'] | |
305 | >>> # absolute importing |
|
306 | >>> # absolute importing | |
306 | >>> sorted(imported_modules( |
|
307 | >>> sorted(imported_modules( | |
307 | ... 'from baz import baz1, name1', |
|
308 | ... 'from baz import baz1, name1', | |
308 | ... modulename, f, localmods)) |
|
309 | ... modulename, f, localmods)) | |
309 | ['baz.__init__', 'baz.baz1'] |
|
310 | ['baz.__init__', 'baz.baz1'] | |
310 | >>> # mixed importing, even though it shouldn't be recommended |
|
311 | >>> # mixed importing, even though it shouldn't be recommended | |
311 | >>> sorted(imported_modules( |
|
312 | >>> sorted(imported_modules( | |
312 | ... 'import stdlib, foo1, baz', |
|
313 | ... 'import stdlib, foo1, baz', | |
313 | ... modulename, f, localmods)) |
|
314 | ... modulename, f, localmods)) | |
314 | ['baz.__init__', 'foo.foo1'] |
|
315 | ['baz.__init__', 'foo.foo1'] | |
315 | >>> # ignore_nested |
|
316 | >>> # ignore_nested | |
316 | >>> sorted(imported_modules( |
|
317 | >>> sorted(imported_modules( | |
317 | ... '''import foo |
|
318 | ... '''import foo | |
318 | ... def wat(): |
|
319 | ... def wat(): | |
319 | ... import bar |
|
320 | ... import bar | |
320 | ... ''', modulename, f, localmods)) |
|
321 | ... ''', modulename, f, localmods)) | |
321 | ['foo.__init__', 'foo.bar.__init__'] |
|
322 | ['foo.__init__', 'foo.bar.__init__'] | |
322 | >>> sorted(imported_modules( |
|
323 | >>> sorted(imported_modules( | |
323 | ... '''import foo |
|
324 | ... '''import foo | |
324 | ... def wat(): |
|
325 | ... def wat(): | |
325 | ... import bar |
|
326 | ... import bar | |
326 | ... ''', modulename, f, localmods, ignore_nested=True)) |
|
327 | ... ''', modulename, f, localmods, ignore_nested=True)) | |
327 | ['foo.__init__'] |
|
328 | ['foo.__init__'] | |
328 | """ |
|
329 | """ | |
329 | fromlocal = fromlocalfunc(modulename, localmods) |
|
330 | fromlocal = fromlocalfunc(modulename, localmods) | |
330 | for node in ast.walk(ast.parse(source, f)): |
|
331 | for node in ast.walk(ast.parse(source, f)): | |
331 | if ignore_nested and getattr(node, 'col_offset', 0) > 0: |
|
332 | if ignore_nested and getattr(node, 'col_offset', 0) > 0: | |
332 | continue |
|
333 | continue | |
333 | if isinstance(node, ast.Import): |
|
334 | if isinstance(node, ast.Import): | |
334 | for n in node.names: |
|
335 | for n in node.names: | |
335 | found = fromlocal(n.name) |
|
336 | found = fromlocal(n.name) | |
336 | if not found: |
|
337 | if not found: | |
337 | # this should import standard library |
|
338 | # this should import standard library | |
338 | continue |
|
339 | continue | |
339 | yield found[1] |
|
340 | yield found[1] | |
340 | elif isinstance(node, ast.ImportFrom): |
|
341 | elif isinstance(node, ast.ImportFrom): | |
341 | found = fromlocal(node.module, node.level) |
|
342 | found = fromlocal(node.module, node.level) | |
342 | if not found: |
|
343 | if not found: | |
343 | # this should import standard library |
|
344 | # this should import standard library | |
344 | continue |
|
345 | continue | |
345 |
|
346 | |||
346 | absname, dottedpath, hassubmod = found |
|
347 | absname, dottedpath, hassubmod = found | |
347 | if not hassubmod: |
|
348 | if not hassubmod: | |
348 | # "dottedpath" is not a package; must be imported |
|
349 | # "dottedpath" is not a package; must be imported | |
349 | yield dottedpath |
|
350 | yield dottedpath | |
350 | # examination of "node.names" should be redundant |
|
351 | # examination of "node.names" should be redundant | |
351 | # e.g.: from mercurial.node import nullid, nullrev |
|
352 | # e.g.: from mercurial.node import nullid, nullrev | |
352 | continue |
|
353 | continue | |
353 |
|
354 | |||
354 | modnotfound = False |
|
355 | modnotfound = False | |
355 | prefix = absname + '.' |
|
356 | prefix = absname + '.' | |
356 | for n in node.names: |
|
357 | for n in node.names: | |
357 | found = fromlocal(prefix + n.name) |
|
358 | found = fromlocal(prefix + n.name) | |
358 | if not found: |
|
359 | if not found: | |
359 | # this should be a function or a property of "node.module" |
|
360 | # this should be a function or a property of "node.module" | |
360 | modnotfound = True |
|
361 | modnotfound = True | |
361 | continue |
|
362 | continue | |
362 | yield found[1] |
|
363 | yield found[1] | |
363 | if modnotfound: |
|
364 | if modnotfound: | |
364 | # "dottedpath" is a package, but imported because of non-module |
|
365 | # "dottedpath" is a package, but imported because of non-module | |
365 | # lookup |
|
366 | # lookup | |
366 | yield dottedpath |
|
367 | yield dottedpath | |
367 |
|
368 | |||
368 | def verify_import_convention(module, source, localmods): |
|
369 | def verify_import_convention(module, source, localmods): | |
369 | """Verify imports match our established coding convention. |
|
370 | """Verify imports match our established coding convention. | |
370 |
|
371 | |||
371 | We have 2 conventions: legacy and modern. The modern convention is in |
|
372 | We have 2 conventions: legacy and modern. The modern convention is in | |
372 | effect when using absolute imports. |
|
373 | effect when using absolute imports. | |
373 |
|
374 | |||
374 | The legacy convention only looks for mixed imports. The modern convention |
|
375 | The legacy convention only looks for mixed imports. The modern convention | |
375 | is much more thorough. |
|
376 | is much more thorough. | |
376 | """ |
|
377 | """ | |
377 | root = ast.parse(source) |
|
378 | root = ast.parse(source) | |
378 | absolute = usingabsolute(root) |
|
379 | absolute = usingabsolute(root) | |
379 |
|
380 | |||
380 | if absolute: |
|
381 | if absolute: | |
381 | return verify_modern_convention(module, root, localmods) |
|
382 | return verify_modern_convention(module, root, localmods) | |
382 | else: |
|
383 | else: | |
383 | return verify_stdlib_on_own_line(root) |
|
384 | return verify_stdlib_on_own_line(root) | |
384 |
|
385 | |||
385 | def verify_modern_convention(module, root, localmods, root_col_offset=0): |
|
386 | def verify_modern_convention(module, root, localmods, root_col_offset=0): | |
386 | """Verify a file conforms to the modern import convention rules. |
|
387 | """Verify a file conforms to the modern import convention rules. | |
387 |
|
388 | |||
388 | The rules of the modern convention are: |
|
389 | The rules of the modern convention are: | |
389 |
|
390 | |||
390 | * Ordering is stdlib followed by local imports. Each group is lexically |
|
391 | * Ordering is stdlib followed by local imports. Each group is lexically | |
391 | sorted. |
|
392 | sorted. | |
392 | * Importing multiple modules via "import X, Y" is not allowed: use |
|
393 | * Importing multiple modules via "import X, Y" is not allowed: use | |
393 | separate import statements. |
|
394 | separate import statements. | |
394 | * Importing multiple modules via "from X import ..." is allowed if using |
|
395 | * Importing multiple modules via "from X import ..." is allowed if using | |
395 | parenthesis and one entry per line. |
|
396 | parenthesis and one entry per line. | |
396 | * Only 1 relative import statement per import level ("from .", "from ..") |
|
397 | * Only 1 relative import statement per import level ("from .", "from ..") | |
397 | is allowed. |
|
398 | is allowed. | |
398 | * Relative imports from higher levels must occur before lower levels. e.g. |
|
399 | * Relative imports from higher levels must occur before lower levels. e.g. | |
399 | "from .." must be before "from .". |
|
400 | "from .." must be before "from .". | |
400 | * Imports from peer packages should use relative import (e.g. do not |
|
401 | * Imports from peer packages should use relative import (e.g. do not | |
401 | "import mercurial.foo" from a "mercurial.*" module). |
|
402 | "import mercurial.foo" from a "mercurial.*" module). | |
402 | * Symbols can only be imported from specific modules (see |
|
403 | * Symbols can only be imported from specific modules (see | |
403 | `allowsymbolimports`). For other modules, first import the module then |
|
404 | `allowsymbolimports`). For other modules, first import the module then | |
404 | assign the symbol to a module-level variable. In addition, these imports |
|
405 | assign the symbol to a module-level variable. In addition, these imports | |
405 | must be performed before other local imports. This rule only |
|
406 | must be performed before other local imports. This rule only | |
406 | applies to import statements outside of any blocks. |
|
407 | applies to import statements outside of any blocks. | |
407 | * Relative imports from the standard library are not allowed. |
|
408 | * Relative imports from the standard library are not allowed. | |
408 | * Certain modules must be aliased to alternate names to avoid aliasing |
|
409 | * Certain modules must be aliased to alternate names to avoid aliasing | |
409 | and readability problems. See `requirealias`. |
|
410 | and readability problems. See `requirealias`. | |
410 | """ |
|
411 | """ | |
411 | if not isinstance(module, str): |
|
412 | if not isinstance(module, str): | |
412 | module = module.decode('ascii') |
|
413 | module = module.decode('ascii') | |
413 | topmodule = module.split('.')[0] |
|
414 | topmodule = module.split('.')[0] | |
414 | fromlocal = fromlocalfunc(module, localmods) |
|
415 | fromlocal = fromlocalfunc(module, localmods) | |
415 |
|
416 | |||
416 | # Whether a local/non-stdlib import has been performed. |
|
417 | # Whether a local/non-stdlib import has been performed. | |
417 | seenlocal = None |
|
418 | seenlocal = None | |
418 | # Whether a local/non-stdlib, non-symbol import has been seen. |
|
419 | # Whether a local/non-stdlib, non-symbol import has been seen. | |
419 | seennonsymbollocal = False |
|
420 | seennonsymbollocal = False | |
420 | # The last name to be imported (for sorting). |
|
421 | # The last name to be imported (for sorting). | |
421 | lastname = None |
|
422 | lastname = None | |
422 | laststdlib = None |
|
423 | laststdlib = None | |
423 | # Relative import levels encountered so far. |
|
424 | # Relative import levels encountered so far. | |
424 | seenlevels = set() |
|
425 | seenlevels = set() | |
425 |
|
426 | |||
426 | for node, newscope in walklocal(root): |
|
427 | for node, newscope in walklocal(root): | |
427 | def msg(fmt, *args): |
|
428 | def msg(fmt, *args): | |
428 | return (fmt % args, node.lineno) |
|
429 | return (fmt % args, node.lineno) | |
429 | if newscope: |
|
430 | if newscope: | |
430 | # Check for local imports in function |
|
431 | # Check for local imports in function | |
431 | for r in verify_modern_convention(module, node, localmods, |
|
432 | for r in verify_modern_convention(module, node, localmods, | |
432 | node.col_offset + 4): |
|
433 | node.col_offset + 4): | |
433 | yield r |
|
434 | yield r | |
434 | elif isinstance(node, ast.Import): |
|
435 | elif isinstance(node, ast.Import): | |
435 | # Disallow "import foo, bar" and require separate imports |
|
436 | # Disallow "import foo, bar" and require separate imports | |
436 | # for each module. |
|
437 | # for each module. | |
437 | if len(node.names) > 1: |
|
438 | if len(node.names) > 1: | |
438 | yield msg('multiple imported names: %s', |
|
439 | yield msg('multiple imported names: %s', | |
439 | ', '.join(n.name for n in node.names)) |
|
440 | ', '.join(n.name for n in node.names)) | |
440 |
|
441 | |||
441 | name = node.names[0].name |
|
442 | name = node.names[0].name | |
442 | asname = node.names[0].asname |
|
443 | asname = node.names[0].asname | |
443 |
|
444 | |||
444 | stdlib = name in stdlib_modules |
|
445 | stdlib = name in stdlib_modules | |
445 |
|
446 | |||
446 | # Ignore sorting rules on imports inside blocks. |
|
447 | # Ignore sorting rules on imports inside blocks. | |
447 | if node.col_offset == root_col_offset: |
|
448 | if node.col_offset == root_col_offset: | |
448 | if lastname and name < lastname and laststdlib == stdlib: |
|
449 | if lastname and name < lastname and laststdlib == stdlib: | |
449 | yield msg('imports not lexically sorted: %s < %s', |
|
450 | yield msg('imports not lexically sorted: %s < %s', | |
450 | name, lastname) |
|
451 | name, lastname) | |
451 |
|
452 | |||
452 | lastname = name |
|
453 | lastname = name | |
453 | laststdlib = stdlib |
|
454 | laststdlib = stdlib | |
454 |
|
455 | |||
455 | # stdlib imports should be before local imports. |
|
456 | # stdlib imports should be before local imports. | |
456 | if stdlib and seenlocal and node.col_offset == root_col_offset: |
|
457 | if stdlib and seenlocal and node.col_offset == root_col_offset: | |
457 | yield msg('stdlib import "%s" follows local import: %s', |
|
458 | yield msg('stdlib import "%s" follows local import: %s', | |
458 | name, seenlocal) |
|
459 | name, seenlocal) | |
459 |
|
460 | |||
460 | if not stdlib: |
|
461 | if not stdlib: | |
461 | seenlocal = name |
|
462 | seenlocal = name | |
462 |
|
463 | |||
463 | # Import of sibling modules should use relative imports. |
|
464 | # Import of sibling modules should use relative imports. | |
464 | topname = name.split('.')[0] |
|
465 | topname = name.split('.')[0] | |
465 | if topname == topmodule: |
|
466 | if topname == topmodule: | |
466 | yield msg('import should be relative: %s', name) |
|
467 | yield msg('import should be relative: %s', name) | |
467 |
|
468 | |||
468 | if name in requirealias and asname != requirealias[name]: |
|
469 | if name in requirealias and asname != requirealias[name]: | |
469 | yield msg('%s module must be "as" aliased to %s', |
|
470 | yield msg('%s module must be "as" aliased to %s', | |
470 | name, requirealias[name]) |
|
471 | name, requirealias[name]) | |
471 |
|
472 | |||
472 | elif isinstance(node, ast.ImportFrom): |
|
473 | elif isinstance(node, ast.ImportFrom): | |
473 | # Resolve the full imported module name. |
|
474 | # Resolve the full imported module name. | |
474 | if node.level > 0: |
|
475 | if node.level > 0: | |
475 | fullname = '.'.join(module.split('.')[:-node.level]) |
|
476 | fullname = '.'.join(module.split('.')[:-node.level]) | |
476 | if node.module: |
|
477 | if node.module: | |
477 | fullname += '.%s' % node.module |
|
478 | fullname += '.%s' % node.module | |
478 | else: |
|
479 | else: | |
479 | assert node.module |
|
480 | assert node.module | |
480 | fullname = node.module |
|
481 | fullname = node.module | |
481 |
|
482 | |||
482 | topname = fullname.split('.')[0] |
|
483 | topname = fullname.split('.')[0] | |
483 | if topname == topmodule: |
|
484 | if topname == topmodule: | |
484 | yield msg('import should be relative: %s', fullname) |
|
485 | yield msg('import should be relative: %s', fullname) | |
485 |
|
486 | |||
486 | # __future__ is special since it needs to come first and use |
|
487 | # __future__ is special since it needs to come first and use | |
487 | # symbol import. |
|
488 | # symbol import. | |
488 | if fullname != '__future__': |
|
489 | if fullname != '__future__': | |
489 | if not fullname or fullname in stdlib_modules: |
|
490 | if not fullname or fullname in stdlib_modules: | |
490 | yield msg('relative import of stdlib module') |
|
491 | yield msg('relative import of stdlib module') | |
491 | else: |
|
492 | else: | |
492 | seenlocal = fullname |
|
493 | seenlocal = fullname | |
493 |
|
494 | |||
494 | # Direct symbol import is only allowed from certain modules and |
|
495 | # Direct symbol import is only allowed from certain modules and | |
495 | # must occur before non-symbol imports. |
|
496 | # must occur before non-symbol imports. | |
496 | found = fromlocal(node.module, node.level) |
|
497 | found = fromlocal(node.module, node.level) | |
497 | if found and found[2]: # node.module is a package |
|
498 | if found and found[2]: # node.module is a package | |
498 | prefix = found[0] + '.' |
|
499 | prefix = found[0] + '.' | |
499 | symbols = (n.name for n in node.names |
|
500 | symbols = (n.name for n in node.names | |
500 | if not fromlocal(prefix + n.name)) |
|
501 | if not fromlocal(prefix + n.name)) | |
501 | else: |
|
502 | else: | |
502 | symbols = (n.name for n in node.names) |
|
503 | symbols = (n.name for n in node.names) | |
503 | symbols = [sym for sym in symbols if sym not in directsymbols] |
|
504 | symbols = [sym for sym in symbols if sym not in directsymbols] | |
504 | if node.module and node.col_offset == root_col_offset: |
|
505 | if node.module and node.col_offset == root_col_offset: | |
505 | if symbols and fullname not in allowsymbolimports: |
|
506 | if symbols and fullname not in allowsymbolimports: | |
506 | yield msg('direct symbol import %s from %s', |
|
507 | yield msg('direct symbol import %s from %s', | |
507 | ', '.join(symbols), fullname) |
|
508 | ', '.join(symbols), fullname) | |
508 |
|
509 | |||
509 | if symbols and seennonsymbollocal: |
|
510 | if symbols and seennonsymbollocal: | |
510 | yield msg('symbol import follows non-symbol import: %s', |
|
511 | yield msg('symbol import follows non-symbol import: %s', | |
511 | fullname) |
|
512 | fullname) | |
512 | if not symbols and fullname not in stdlib_modules: |
|
513 | if not symbols and fullname not in stdlib_modules: | |
513 | seennonsymbollocal = True |
|
514 | seennonsymbollocal = True | |
514 |
|
515 | |||
515 | if not node.module: |
|
516 | if not node.module: | |
516 | assert node.level |
|
517 | assert node.level | |
517 |
|
518 | |||
518 | # Only allow 1 group per level. |
|
519 | # Only allow 1 group per level. | |
519 | if (node.level in seenlevels |
|
520 | if (node.level in seenlevels | |
520 | and node.col_offset == root_col_offset): |
|
521 | and node.col_offset == root_col_offset): | |
521 | yield msg('multiple "from %s import" statements', |
|
522 | yield msg('multiple "from %s import" statements', | |
522 | '.' * node.level) |
|
523 | '.' * node.level) | |
523 |
|
524 | |||
524 | # Higher-level groups come before lower-level groups. |
|
525 | # Higher-level groups come before lower-level groups. | |
525 | if any(node.level > l for l in seenlevels): |
|
526 | if any(node.level > l for l in seenlevels): | |
526 | yield msg('higher-level import should come first: %s', |
|
527 | yield msg('higher-level import should come first: %s', | |
527 | fullname) |
|
528 | fullname) | |
528 |
|
529 | |||
529 | seenlevels.add(node.level) |
|
530 | seenlevels.add(node.level) | |
530 |
|
531 | |||
531 | # Entries in "from .X import ( ... )" lists must be lexically |
|
532 | # Entries in "from .X import ( ... )" lists must be lexically | |
532 | # sorted. |
|
533 | # sorted. | |
533 | lastentryname = None |
|
534 | lastentryname = None | |
534 |
|
535 | |||
535 | for n in node.names: |
|
536 | for n in node.names: | |
536 | if lastentryname and n.name < lastentryname: |
|
537 | if lastentryname and n.name < lastentryname: | |
537 | yield msg('imports from %s not lexically sorted: %s < %s', |
|
538 | yield msg('imports from %s not lexically sorted: %s < %s', | |
538 | fullname, n.name, lastentryname) |
|
539 | fullname, n.name, lastentryname) | |
539 |
|
540 | |||
540 | lastentryname = n.name |
|
541 | lastentryname = n.name | |
541 |
|
542 | |||
542 | if n.name in requirealias and n.asname != requirealias[n.name]: |
|
543 | if n.name in requirealias and n.asname != requirealias[n.name]: | |
543 | yield msg('%s from %s must be "as" aliased to %s', |
|
544 | yield msg('%s from %s must be "as" aliased to %s', | |
544 | n.name, fullname, requirealias[n.name]) |
|
545 | n.name, fullname, requirealias[n.name]) | |
545 |
|
546 | |||
546 | def verify_stdlib_on_own_line(root): |
|
547 | def verify_stdlib_on_own_line(root): | |
547 | """Given some python source, verify that stdlib imports are done |
|
548 | """Given some python source, verify that stdlib imports are done | |
548 | in separate statements from relative local module imports. |
|
549 | in separate statements from relative local module imports. | |
549 |
|
550 | |||
550 | >>> list(verify_stdlib_on_own_line(ast.parse('import sys, foo'))) |
|
551 | >>> list(verify_stdlib_on_own_line(ast.parse('import sys, foo'))) | |
551 | [('mixed imports\\n stdlib: sys\\n relative: foo', 1)] |
|
552 | [('mixed imports\\n stdlib: sys\\n relative: foo', 1)] | |
552 | >>> list(verify_stdlib_on_own_line(ast.parse('import sys, os'))) |
|
553 | >>> list(verify_stdlib_on_own_line(ast.parse('import sys, os'))) | |
553 | [] |
|
554 | [] | |
554 | >>> list(verify_stdlib_on_own_line(ast.parse('import foo, bar'))) |
|
555 | >>> list(verify_stdlib_on_own_line(ast.parse('import foo, bar'))) | |
555 | [] |
|
556 | [] | |
556 | """ |
|
557 | """ | |
557 | for node in ast.walk(root): |
|
558 | for node in ast.walk(root): | |
558 | if isinstance(node, ast.Import): |
|
559 | if isinstance(node, ast.Import): | |
559 | from_stdlib = {False: [], True: []} |
|
560 | from_stdlib = {False: [], True: []} | |
560 | for n in node.names: |
|
561 | for n in node.names: | |
561 | from_stdlib[n.name in stdlib_modules].append(n.name) |
|
562 | from_stdlib[n.name in stdlib_modules].append(n.name) | |
562 | if from_stdlib[True] and from_stdlib[False]: |
|
563 | if from_stdlib[True] and from_stdlib[False]: | |
563 | yield ('mixed imports\n stdlib: %s\n relative: %s' % |
|
564 | yield ('mixed imports\n stdlib: %s\n relative: %s' % | |
564 | (', '.join(sorted(from_stdlib[True])), |
|
565 | (', '.join(sorted(from_stdlib[True])), | |
565 | ', '.join(sorted(from_stdlib[False]))), node.lineno) |
|
566 | ', '.join(sorted(from_stdlib[False]))), node.lineno) | |
566 |
|
567 | |||
567 | class CircularImport(Exception): |
|
568 | class CircularImport(Exception): | |
568 | pass |
|
569 | pass | |
569 |
|
570 | |||
570 | def checkmod(mod, imports): |
|
571 | def checkmod(mod, imports): | |
571 | shortest = {} |
|
572 | shortest = {} | |
572 | visit = [[mod]] |
|
573 | visit = [[mod]] | |
573 | while visit: |
|
574 | while visit: | |
574 | path = visit.pop(0) |
|
575 | path = visit.pop(0) | |
575 | for i in sorted(imports.get(path[-1], [])): |
|
576 | for i in sorted(imports.get(path[-1], [])): | |
576 | if len(path) < shortest.get(i, 1000): |
|
577 | if len(path) < shortest.get(i, 1000): | |
577 | shortest[i] = len(path) |
|
578 | shortest[i] = len(path) | |
578 | if i in path: |
|
579 | if i in path: | |
579 | if i == path[0]: |
|
580 | if i == path[0]: | |
580 | raise CircularImport(path) |
|
581 | raise CircularImport(path) | |
581 | continue |
|
582 | continue | |
582 | visit.append(path + [i]) |
|
583 | visit.append(path + [i]) | |
583 |
|
584 | |||
584 | def rotatecycle(cycle): |
|
585 | def rotatecycle(cycle): | |
585 | """arrange a cycle so that the lexicographically first module listed first |
|
586 | """arrange a cycle so that the lexicographically first module listed first | |
586 |
|
587 | |||
587 | >>> rotatecycle(['foo', 'bar']) |
|
588 | >>> rotatecycle(['foo', 'bar']) | |
588 | ['bar', 'foo', 'bar'] |
|
589 | ['bar', 'foo', 'bar'] | |
589 | """ |
|
590 | """ | |
590 | lowest = min(cycle) |
|
591 | lowest = min(cycle) | |
591 | idx = cycle.index(lowest) |
|
592 | idx = cycle.index(lowest) | |
592 | return cycle[idx:] + cycle[:idx] + [lowest] |
|
593 | return cycle[idx:] + cycle[:idx] + [lowest] | |
593 |
|
594 | |||
594 | def find_cycles(imports): |
|
595 | def find_cycles(imports): | |
595 | """Find cycles in an already-loaded import graph. |
|
596 | """Find cycles in an already-loaded import graph. | |
596 |
|
597 | |||
597 | All module names recorded in `imports` should be absolute one. |
|
598 | All module names recorded in `imports` should be absolute one. | |
598 |
|
599 | |||
599 | >>> from __future__ import print_function |
|
600 | >>> from __future__ import print_function | |
600 | >>> imports = {'top.foo': ['top.bar', 'os.path', 'top.qux'], |
|
601 | >>> imports = {'top.foo': ['top.bar', 'os.path', 'top.qux'], | |
601 | ... 'top.bar': ['top.baz', 'sys'], |
|
602 | ... 'top.bar': ['top.baz', 'sys'], | |
602 | ... 'top.baz': ['top.foo'], |
|
603 | ... 'top.baz': ['top.foo'], | |
603 | ... 'top.qux': ['top.foo']} |
|
604 | ... 'top.qux': ['top.foo']} | |
604 | >>> print('\\n'.join(sorted(find_cycles(imports)))) |
|
605 | >>> print('\\n'.join(sorted(find_cycles(imports)))) | |
605 | top.bar -> top.baz -> top.foo -> top.bar |
|
606 | top.bar -> top.baz -> top.foo -> top.bar | |
606 | top.foo -> top.qux -> top.foo |
|
607 | top.foo -> top.qux -> top.foo | |
607 | """ |
|
608 | """ | |
608 | cycles = set() |
|
609 | cycles = set() | |
609 | for mod in sorted(imports.keys()): |
|
610 | for mod in sorted(imports.keys()): | |
610 | try: |
|
611 | try: | |
611 | checkmod(mod, imports) |
|
612 | checkmod(mod, imports) | |
612 | except CircularImport as e: |
|
613 | except CircularImport as e: | |
613 | cycle = e.args[0] |
|
614 | cycle = e.args[0] | |
614 | cycles.add(" -> ".join(rotatecycle(cycle))) |
|
615 | cycles.add(" -> ".join(rotatecycle(cycle))) | |
615 | return cycles |
|
616 | return cycles | |
616 |
|
617 | |||
617 | def _cycle_sortkey(c): |
|
618 | def _cycle_sortkey(c): | |
618 | return len(c), c |
|
619 | return len(c), c | |
619 |
|
620 | |||
620 | def embedded(f, modname, src): |
|
621 | def embedded(f, modname, src): | |
621 | """Extract embedded python code |
|
622 | """Extract embedded python code | |
622 |
|
623 | |||
623 | >>> def _forcestr(thing): |
|
624 | >>> def _forcestr(thing): | |
624 | ... if not isinstance(thing, str): |
|
625 | ... if not isinstance(thing, str): | |
625 | ... return thing.decode('ascii') |
|
626 | ... return thing.decode('ascii') | |
626 | ... return thing |
|
627 | ... return thing | |
627 | >>> def test(fn, lines): |
|
628 | >>> def test(fn, lines): | |
628 | ... for s, m, f, l in embedded(fn, b"example", lines): |
|
629 | ... for s, m, f, l in embedded(fn, b"example", lines): | |
629 | ... print("%s %s %d" % (_forcestr(m), _forcestr(f), l)) |
|
630 | ... print("%s %s %d" % (_forcestr(m), _forcestr(f), l)) | |
630 | ... print(repr(_forcestr(s))) |
|
631 | ... print(repr(_forcestr(s))) | |
631 | >>> lines = [ |
|
632 | >>> lines = [ | |
632 | ... b'comment', |
|
633 | ... b'comment', | |
633 | ... b' >>> from __future__ import print_function', |
|
634 | ... b' >>> from __future__ import print_function', | |
634 | ... b" >>> ' multiline", |
|
635 | ... b" >>> ' multiline", | |
635 | ... b" ... string'", |
|
636 | ... b" ... string'", | |
636 | ... b' ', |
|
637 | ... b' ', | |
637 | ... b'comment', |
|
638 | ... b'comment', | |
638 | ... b' $ cat > foo.py <<EOF', |
|
639 | ... b' $ cat > foo.py <<EOF', | |
639 | ... b' > from __future__ import print_function', |
|
640 | ... b' > from __future__ import print_function', | |
640 | ... b' > EOF', |
|
641 | ... b' > EOF', | |
641 | ... ] |
|
642 | ... ] | |
642 | >>> test(b"example.t", lines) |
|
643 | >>> test(b"example.t", lines) | |
643 | example[2] doctest.py 2 |
|
644 | example[2] doctest.py 2 | |
644 | "from __future__ import print_function\\n' multiline\\nstring'\\n" |
|
645 | "from __future__ import print_function\\n' multiline\\nstring'\\n" | |
645 | example[7] foo.py 7 |
|
646 | example[7] foo.py 7 | |
646 | 'from __future__ import print_function\\n' |
|
647 | 'from __future__ import print_function\\n' | |
647 | """ |
|
648 | """ | |
648 | inlinepython = 0 |
|
649 | inlinepython = 0 | |
649 | shpython = 0 |
|
650 | shpython = 0 | |
650 | script = [] |
|
651 | script = [] | |
651 | prefix = 6 |
|
652 | prefix = 6 | |
652 | t = '' |
|
653 | t = '' | |
653 | n = 0 |
|
654 | n = 0 | |
654 | for l in src: |
|
655 | for l in src: | |
655 | n += 1 |
|
656 | n += 1 | |
656 | if not l.endswith(b'\n'): |
|
657 | if not l.endswith(b'\n'): | |
657 | l += b'\n' |
|
658 | l += b'\n' | |
658 | if l.startswith(b' >>> '): # python inlines |
|
659 | if l.startswith(b' >>> '): # python inlines | |
659 | if shpython: |
|
660 | if shpython: | |
660 | print("%s:%d: Parse Error" % (f, n)) |
|
661 | print("%s:%d: Parse Error" % (f, n)) | |
661 | if not inlinepython: |
|
662 | if not inlinepython: | |
662 | # We've just entered a Python block. |
|
663 | # We've just entered a Python block. | |
663 | inlinepython = n |
|
664 | inlinepython = n | |
664 | t = b'doctest.py' |
|
665 | t = b'doctest.py' | |
665 | script.append(l[prefix:]) |
|
666 | script.append(l[prefix:]) | |
666 | continue |
|
667 | continue | |
667 | if l.startswith(b' ... '): # python inlines |
|
668 | if l.startswith(b' ... '): # python inlines | |
668 | script.append(l[prefix:]) |
|
669 | script.append(l[prefix:]) | |
669 | continue |
|
670 | continue | |
670 | cat = re.search(br"\$ \s*cat\s*>\s*(\S+\.py)\s*<<\s*EOF", l) |
|
671 | cat = re.search(br"\$ \s*cat\s*>\s*(\S+\.py)\s*<<\s*EOF", l) | |
671 | if cat: |
|
672 | if cat: | |
672 | if inlinepython: |
|
673 | if inlinepython: | |
673 | yield b''.join(script), (b"%s[%d]" % |
|
674 | yield b''.join(script), (b"%s[%d]" % | |
674 | (modname, inlinepython)), t, inlinepython |
|
675 | (modname, inlinepython)), t, inlinepython | |
675 | script = [] |
|
676 | script = [] | |
676 | inlinepython = 0 |
|
677 | inlinepython = 0 | |
677 | shpython = n |
|
678 | shpython = n | |
678 | t = cat.group(1) |
|
679 | t = cat.group(1) | |
679 | continue |
|
680 | continue | |
680 | if shpython and l.startswith(b' > '): # sh continuation |
|
681 | if shpython and l.startswith(b' > '): # sh continuation | |
681 | if l == b' > EOF\n': |
|
682 | if l == b' > EOF\n': | |
682 | yield b''.join(script), (b"%s[%d]" % |
|
683 | yield b''.join(script), (b"%s[%d]" % | |
683 | (modname, shpython)), t, shpython |
|
684 | (modname, shpython)), t, shpython | |
684 | script = [] |
|
685 | script = [] | |
685 | shpython = 0 |
|
686 | shpython = 0 | |
686 | else: |
|
687 | else: | |
687 | script.append(l[4:]) |
|
688 | script.append(l[4:]) | |
688 | continue |
|
689 | continue | |
689 | if inlinepython and l == b' \n': |
|
690 | if inlinepython and l == b' \n': | |
690 | yield b''.join(script), (b"%s[%d]" % |
|
691 | yield b''.join(script), (b"%s[%d]" % | |
691 | (modname, inlinepython)), t, inlinepython |
|
692 | (modname, inlinepython)), t, inlinepython | |
692 | script = [] |
|
693 | script = [] | |
693 | inlinepython = 0 |
|
694 | inlinepython = 0 | |
694 | continue |
|
695 | continue | |
695 |
|
696 | |||
696 | def sources(f, modname): |
|
697 | def sources(f, modname): | |
697 | """Yields possibly multiple sources from a filepath |
|
698 | """Yields possibly multiple sources from a filepath | |
698 |
|
699 | |||
699 | input: filepath, modulename |
|
700 | input: filepath, modulename | |
700 | yields: script(string), modulename, filepath, linenumber |
|
701 | yields: script(string), modulename, filepath, linenumber | |
701 |
|
702 | |||
702 | For embedded scripts, the modulename and filepath will be different |
|
703 | For embedded scripts, the modulename and filepath will be different | |
703 | from the function arguments. linenumber is an offset relative to |
|
704 | from the function arguments. linenumber is an offset relative to | |
704 | the input file. |
|
705 | the input file. | |
705 | """ |
|
706 | """ | |
706 | py = False |
|
707 | py = False | |
707 | if not f.endswith('.t'): |
|
708 | if not f.endswith('.t'): | |
708 | with open(f, 'rb') as src: |
|
709 | with open(f, 'rb') as src: | |
709 | yield src.read(), modname, f, 0 |
|
710 | yield src.read(), modname, f, 0 | |
710 | py = True |
|
711 | py = True | |
711 | if py or f.endswith('.t'): |
|
712 | if py or f.endswith('.t'): | |
712 | with open(f, 'rb') as src: |
|
713 | with open(f, 'rb') as src: | |
713 | for script, modname, t, line in embedded(f, modname, src): |
|
714 | for script, modname, t, line in embedded(f, modname, src): | |
714 | yield script, modname, t, line |
|
715 | yield script, modname, t, line | |
715 |
|
716 | |||
716 | def main(argv): |
|
717 | def main(argv): | |
717 | if len(argv) < 2 or (argv[1] == '-' and len(argv) > 2): |
|
718 | if len(argv) < 2 or (argv[1] == '-' and len(argv) > 2): | |
718 | print('Usage: %s {-|file [file] [file] ...}') |
|
719 | print('Usage: %s {-|file [file] [file] ...}') | |
719 | return 1 |
|
720 | return 1 | |
720 | if argv[1] == '-': |
|
721 | if argv[1] == '-': | |
721 | argv = argv[:1] |
|
722 | argv = argv[:1] | |
722 | argv.extend(l.rstrip() for l in sys.stdin.readlines()) |
|
723 | argv.extend(l.rstrip() for l in sys.stdin.readlines()) | |
723 | localmodpaths = {} |
|
724 | localmodpaths = {} | |
724 | used_imports = {} |
|
725 | used_imports = {} | |
725 | any_errors = False |
|
726 | any_errors = False | |
726 | for source_path in argv[1:]: |
|
727 | for source_path in argv[1:]: | |
727 | modname = dotted_name_of_path(source_path) |
|
728 | modname = dotted_name_of_path(source_path) | |
728 | localmodpaths[modname] = source_path |
|
729 | localmodpaths[modname] = source_path | |
729 | localmods = populateextmods(localmodpaths) |
|
730 | localmods = populateextmods(localmodpaths) | |
730 | for localmodname, source_path in sorted(localmodpaths.items()): |
|
731 | for localmodname, source_path in sorted(localmodpaths.items()): | |
731 | if not isinstance(localmodname, bytes): |
|
732 | if not isinstance(localmodname, bytes): | |
732 | # This is only safe because all hg's files are ascii |
|
733 | # This is only safe because all hg's files are ascii | |
733 | localmodname = localmodname.encode('ascii') |
|
734 | localmodname = localmodname.encode('ascii') | |
734 | for src, modname, name, line in sources(source_path, localmodname): |
|
735 | for src, modname, name, line in sources(source_path, localmodname): | |
735 | try: |
|
736 | try: | |
736 | used_imports[modname] = sorted( |
|
737 | used_imports[modname] = sorted( | |
737 | imported_modules(src, modname, name, localmods, |
|
738 | imported_modules(src, modname, name, localmods, | |
738 | ignore_nested=True)) |
|
739 | ignore_nested=True)) | |
739 | for error, lineno in verify_import_convention(modname, src, |
|
740 | for error, lineno in verify_import_convention(modname, src, | |
740 | localmods): |
|
741 | localmods): | |
741 | any_errors = True |
|
742 | any_errors = True | |
742 | print('%s:%d: %s' % (source_path, lineno + line, error)) |
|
743 | print('%s:%d: %s' % (source_path, lineno + line, error)) | |
743 | except SyntaxError as e: |
|
744 | except SyntaxError as e: | |
744 | print('%s:%d: SyntaxError: %s' % |
|
745 | print('%s:%d: SyntaxError: %s' % | |
745 | (source_path, e.lineno + line, e)) |
|
746 | (source_path, e.lineno + line, e)) | |
746 | cycles = find_cycles(used_imports) |
|
747 | cycles = find_cycles(used_imports) | |
747 | if cycles: |
|
748 | if cycles: | |
748 | firstmods = set() |
|
749 | firstmods = set() | |
749 | for c in sorted(cycles, key=_cycle_sortkey): |
|
750 | for c in sorted(cycles, key=_cycle_sortkey): | |
750 | first = c.split()[0] |
|
751 | first = c.split()[0] | |
751 | # As a rough cut, ignore any cycle that starts with the |
|
752 | # As a rough cut, ignore any cycle that starts with the | |
752 | # same module as some other cycle. Otherwise we see lots |
|
753 | # same module as some other cycle. Otherwise we see lots | |
753 | # of cycles that are effectively duplicates. |
|
754 | # of cycles that are effectively duplicates. | |
754 | if first in firstmods: |
|
755 | if first in firstmods: | |
755 | continue |
|
756 | continue | |
756 | print('Import cycle:', c) |
|
757 | print('Import cycle:', c) | |
757 | firstmods.add(first) |
|
758 | firstmods.add(first) | |
758 | any_errors = True |
|
759 | any_errors = True | |
759 | return any_errors != 0 |
|
760 | return any_errors != 0 | |
760 |
|
761 | |||
761 | if __name__ == '__main__': |
|
762 | if __name__ == '__main__': | |
762 | sys.exit(int(main(sys.argv))) |
|
763 | sys.exit(int(main(sys.argv))) |
General Comments 0
You need to be logged in to leave comments.
Login now