blob: 6b56be488eeb0319a8621d2edf33bcfe8f07c2ec [file] [log] [blame]
Tim Peters6cd6a822001-08-17 22:11:27 +00001r"""Utilities to compile possibly incomplete Python source code.
Guido van Rossumc41c1a91998-10-22 21:56:15 +00002
Tim Peters6cd6a822001-08-17 22:11:27 +00003This module provides two interfaces, broadly similar to the builtin
Walter Dörwald4df30682003-11-20 13:38:01 +00004function compile(), which take program text, a filename and a 'mode'
Tim Peters6cd6a822001-08-17 22:11:27 +00005and:
Skip Montanaroe99d5ea2001-01-20 19:54:20 +00006
Walter Dörwald4df30682003-11-20 13:38:01 +00007- Return code object if the command is complete and valid
8- Return None if the command is incomplete
9- Raise SyntaxError, ValueError or OverflowError if the command is a
Tim Peters6cd6a822001-08-17 22:11:27 +000010 syntax error (OverflowError and ValueError can be produced by
11 malformed literals).
Guido van Rossumc41c1a91998-10-22 21:56:15 +000012
Tim Peters6cd6a822001-08-17 22:11:27 +000013Approach:
Guido van Rossumc41c1a91998-10-22 21:56:15 +000014
Tim Peters6cd6a822001-08-17 22:11:27 +000015First, check if the source consists entirely of blank lines and
16comments; if so, replace it with 'pass', because the built-in
17parser doesn't always do the right thing for these.
Guido van Rossumc41c1a91998-10-22 21:56:15 +000018
Tim Peters6cd6a822001-08-17 22:11:27 +000019Compile three times: as is, with \n, and with \n\n appended. If it
20compiles as is, it's complete. If it compiles with one \n appended,
21we expect more. If it doesn't compile either way, we compare the
22error we get when compiling with \n or \n\n appended. If the errors
23are the same, the code is broken. But if the errors are different, we
24expect more. Not intuitive; not even guaranteed to hold in future
25releases; but this matches the compiler's behavior from Python 1.4
26through 2.2, at least.
Guido van Rossumc41c1a91998-10-22 21:56:15 +000027
Tim Peters6cd6a822001-08-17 22:11:27 +000028Caveat:
Guido van Rossumc41c1a91998-10-22 21:56:15 +000029
Tim Peters6cd6a822001-08-17 22:11:27 +000030It is possible (but not likely) that the parser stops parsing with a
31successful outcome before reaching the end of the source; in this
32case, trailing symbols may be ignored instead of causing an error.
33For example, a backslash followed by two newlines may be followed by
34arbitrary garbage. This will be fixed once the API for the parser is
35better.
Guido van Rossumc41c1a91998-10-22 21:56:15 +000036
Tim Peters6cd6a822001-08-17 22:11:27 +000037The two interfaces are:
Guido van Rossumc41c1a91998-10-22 21:56:15 +000038
Tim Peters6cd6a822001-08-17 22:11:27 +000039compile_command(source, filename, symbol):
Guido van Rossumc41c1a91998-10-22 21:56:15 +000040
Tim Peters6cd6a822001-08-17 22:11:27 +000041 Compiles a single command in the manner described above.
Guido van Rossumc41c1a91998-10-22 21:56:15 +000042
Tim Peters6cd6a822001-08-17 22:11:27 +000043CommandCompiler():
Guido van Rossumc41c1a91998-10-22 21:56:15 +000044
Tim Peters6cd6a822001-08-17 22:11:27 +000045 Instances of this class have __call__ methods identical in
46 signature to compile_command; the difference is that if the
47 instance compiles program text containing a __future__ statement,
48 the instance 'remembers' and compiles all subsequent program texts
49 with the statement in force.
Guido van Rossumc41c1a91998-10-22 21:56:15 +000050
Tim Peters6cd6a822001-08-17 22:11:27 +000051The module also provides another class:
52
53Compile():
54
55 Instances of this class act like the built-in function compile,
56 but with 'memory' in the sense described above.
57"""
58
59import __future__
Cheryl Sabella052d3fc2020-06-04 19:40:24 -040060import warnings
Tim Peters6cd6a822001-08-17 22:11:27 +000061
62_features = [getattr(__future__, fname)
63 for fname in __future__.all_feature_names]
64
65__all__ = ["compile_command", "Compile", "CommandCompiler"]
66
Terry Jan Reedyb676f5f2021-02-13 01:49:18 -050067PyCF_DONT_IMPLY_DEDENT = 0x200 # Matches pythonrun.h.
Guido van Rossum4b499dd32003-02-13 22:07:59 +000068
Tim Peters6cd6a822001-08-17 22:11:27 +000069def _maybe_compile(compiler, source, filename, symbol):
Terry Jan Reedyb676f5f2021-02-13 01:49:18 -050070 # Check for source consisting of only blank lines and comments.
Eric S. Raymond6b71e742001-02-09 08:56:30 +000071 for line in source.split("\n"):
72 line = line.strip()
Guido van Rossumc41c1a91998-10-22 21:56:15 +000073 if line and line[0] != '#':
Terry Jan Reedyb676f5f2021-02-13 01:49:18 -050074 break # Leave it alone.
Guido van Rossumc41c1a91998-10-22 21:56:15 +000075 else:
Guido van Rossum993bc3a2003-05-16 01:24:30 +000076 if symbol != "eval":
77 source = "pass" # Replace it with a 'pass' statement
Guido van Rossumc41c1a91998-10-22 21:56:15 +000078
Guido van Rossumc41c1a91998-10-22 21:56:15 +000079 try:
Terry Jan Reedy2068b262021-02-11 19:31:10 -050080 return compiler(source, filename, symbol)
Terry Jan Reedyb676f5f2021-02-13 01:49:18 -050081 except SyntaxError: # Let other compile() errors propagate.
Guido van Rossumc41c1a91998-10-22 21:56:15 +000082 pass
83
Victor Stinner369a1cb2020-08-12 14:53:28 +020084 # Catch syntax warnings after the first compile
Terry Jan Reedyc818b152020-08-13 13:18:49 -040085 # to emit warnings (SyntaxWarning, DeprecationWarning) at most once.
Cheryl Sabella052d3fc2020-06-04 19:40:24 -040086 with warnings.catch_warnings():
Terry Jan Reedyc818b152020-08-13 13:18:49 -040087 warnings.simplefilter("error")
Victor Stinner369a1cb2020-08-12 14:53:28 +020088
Terry Jan Reedyb676f5f2021-02-13 01:49:18 -050089 code1 = err1 = err2 = None
Cheryl Sabella052d3fc2020-06-04 19:40:24 -040090 try:
91 code1 = compiler(source + "\n", filename, symbol)
92 except SyntaxError as e:
93 err1 = e
Guido van Rossumc41c1a91998-10-22 21:56:15 +000094
Cheryl Sabella052d3fc2020-06-04 19:40:24 -040095 try:
96 code2 = compiler(source + "\n\n", filename, symbol)
97 except SyntaxError as e:
98 err2 = e
Guido van Rossumc41c1a91998-10-22 21:56:15 +000099
Mario Corcherob64334c2019-12-06 14:27:38 +0000100 try:
Pablo Galindodbb22812021-02-09 20:07:38 +0000101 if not code1 and _is_syntax_error(err1, err2):
Mario Corcherob64334c2019-12-06 14:27:38 +0000102 raise err1
Terry Jan Reedyb676f5f2021-02-13 01:49:18 -0500103 else:
104 return None
Mario Corcherob64334c2019-12-06 14:27:38 +0000105 finally:
106 err1 = err2 = None
Tim Peters6cd6a822001-08-17 22:11:27 +0000107
Pablo Galindodbb22812021-02-09 20:07:38 +0000108def _is_syntax_error(err1, err2):
109 rep1 = repr(err1)
110 rep2 = repr(err2)
111 if "was never closed" in rep1 and "was never closed" in rep2:
112 return False
113 if rep1 == rep2:
114 return True
115 return False
116
Guido van Rossum4b499dd32003-02-13 22:07:59 +0000117def _compile(source, filename, symbol):
118 return compile(source, filename, symbol, PyCF_DONT_IMPLY_DEDENT)
119
Tim Peters6cd6a822001-08-17 22:11:27 +0000120def compile_command(source, filename="<input>", symbol="single"):
121 r"""Compile a command and determine whether it is incomplete.
122
123 Arguments:
124
125 source -- the source string; may contain \n characters
126 filename -- optional filename from which source was read; default
127 "<input>"
Joannah Nanjekye7ba1f752020-05-14 21:59:46 -0300128 symbol -- optional grammar start symbol; "single" (default), "exec"
129 or "eval"
Tim Peters6cd6a822001-08-17 22:11:27 +0000130
131 Return value / exceptions raised:
132
133 - Return a code object if the command is complete and valid
134 - Return None if the command is incomplete
135 - Raise SyntaxError, ValueError or OverflowError if the command is a
136 syntax error (OverflowError and ValueError can be produced by
137 malformed literals).
138 """
Guido van Rossum4b499dd32003-02-13 22:07:59 +0000139 return _maybe_compile(_compile, source, filename, symbol)
Tim Peters6cd6a822001-08-17 22:11:27 +0000140
141class Compile:
142 """Instances of this class behave much like the built-in compile
143 function, but if one is used to compile text containing a future
144 statement, it "remembers" and compiles all subsequent program texts
145 with the statement in force."""
146 def __init__(self):
Guido van Rossum4b499dd32003-02-13 22:07:59 +0000147 self.flags = PyCF_DONT_IMPLY_DEDENT
Tim Peters6cd6a822001-08-17 22:11:27 +0000148
149 def __call__(self, source, filename, symbol):
Serhiy Storchaka1f21eaa2019-09-01 12:16:51 +0300150 codeob = compile(source, filename, symbol, self.flags, True)
Tim Peters6cd6a822001-08-17 22:11:27 +0000151 for feature in _features:
152 if codeob.co_flags & feature.compiler_flag:
153 self.flags |= feature.compiler_flag
154 return codeob
155
156class CommandCompiler:
157 """Instances of this class have __call__ methods identical in
158 signature to compile_command; the difference is that if the
159 instance compiles program text containing a __future__ statement,
160 the instance 'remembers' and compiles all subsequent program texts
161 with the statement in force."""
162
163 def __init__(self,):
164 self.compiler = Compile()
165
166 def __call__(self, source, filename="<input>", symbol="single"):
167 r"""Compile a command and determine whether it is incomplete.
168
169 Arguments:
170
171 source -- the source string; may contain \n characters
172 filename -- optional filename from which source was read;
173 default "<input>"
174 symbol -- optional grammar start symbol; "single" (default) or
175 "eval"
176
177 Return value / exceptions raised:
178
179 - Return a code object if the command is complete and valid
180 - Return None if the command is incomplete
181 - Raise SyntaxError, ValueError or OverflowError if the command is a
182 syntax error (OverflowError and ValueError can be produced by
183 malformed literals).
184 """
185 return _maybe_compile(self.compiler, source, filename, symbol)