Merge branch 'master' of https://github.com/gward/pep8 into gward-master

5 files changed, 98 insertions, 14 deletions

diff --git a/pep8.py b/pep8.py
index e38ab26..47ea867 100755
--- a/pep8.py
+++ b/pep8.py
@@ -200,7 +200,7 @@ def missing_newline(physical_line):
         return len(physical_line), "W292 no newline at end of file"
 
 
-def maximum_line_length(physical_line, max_line_length):
+def maximum_line_length(physical_line, max_line_length, multiline):
     """
     Limit all lines to a maximum of 79 characters.
 
@@ -216,6 +216,10 @@ def maximum_line_length(physical_line, max_line_length):
     line = physical_line.rstrip()
     length = len(line)
     if length > max_line_length and not noqa(line):
+        # Sometimes, long lines in docstrings are hard to avoid -- like,
+        # a long URL that can't be wrapped because it has no whitespace.
+        if multiline and re.match(r'^\s*\S+$', line):
+            return
         if hasattr(line, 'decode'):   # Python 2
             # The line could contain multi-byte characters
             try:
@@ -1189,6 +1193,7 @@ class Checker(object):
         self._logical_checks = options.logical_checks
         self._ast_checks = options.ast_checks
         self.max_line_length = options.max_line_length
+        self.multiline = False  # in a multiline string?
         self.hang_closing = options.hang_closing
         self.verbose = options.verbose
         self.filename = filename
@@ -1237,16 +1242,9 @@ class Checker(object):
         self.line_number += 1
         if self.line_number > len(self.lines):
             return ''
-        return self.lines[self.line_number - 1]
-
-    def readline_check_physical(self):
-        """
-        Check and return the next physical line. This method can be
-        used to feed tokenize.generate_tokens.
-        """
-        line = self.readline()
-        if line:
-            self.check_physical(line)
+        line = self.lines[self.line_number - 1]
+        if self.indent_char is None and line[:1] in WHITESPACE:
+            self.indent_char = line[0]
         return line
 
     def run_check(self, check, argument_names):
@@ -1263,8 +1261,6 @@ class Checker(object):
         Run all physical checks on a raw input line.
         """
         self.physical_line = line
-        if self.indent_char is None and line[:1] in WHITESPACE:
-            self.indent_char = line[0]
         for name, check, argument_names in self._physical_checks:
             result = self.run_check(check, argument_names)
             if result is not None:
@@ -1352,13 +1348,47 @@ class Checker(object):
     def generate_tokens(self):
         if self._io_error:
             self.report_error(1, 0, 'E902 %s' % self._io_error, readlines)
-        tokengen = tokenize.generate_tokens(self.readline_check_physical)
+        tokengen = tokenize.generate_tokens(self.readline)
         try:
             for token in tokengen:
                 yield token
+                self.maybe_check_physical(token)
         except (SyntaxError, tokenize.TokenError):
             self.report_invalid_syntax()
 
+    def maybe_check_physical(self, token):
+        """
+        If appropriate (based on token), check current physical line(s).
+        """
+        # This is called after every token, but we only want to take action
+        # after a token that ends a line.
+        if token[0] in (tokenize.NEWLINE, tokenize.NL):
+            # Obviously, a newline token ends a single physical line.
+            self.check_physical(token[4])
+        elif token[0] == tokenize.STRING and token[1].count('\n'):
+            # Less obviously, a string that contains newlines is a
+            # multiline string, either triple-quoted or with internal
+            # newlines backslash-escaped. Check every physical line in the
+            # string *except* for the last one: its newline is outside of
+            # the multiline string, so we consider it a regular physical
+            # line, and will check it like any other physical line.
+            #
+            # Subtleties:
+            # - we don't *completely* ignore the last line; if it contains
+            #   the magical "# noqa" comment, we disable all physical
+            #   checks for the entire multiline string
+            # - have to wind self.line_number back because initially it
+            #   points to the last line of the string, and we want
+            #   check_physical() to give accurate feedback
+            if noqa(token[4]):
+                return
+            self.multiline = True
+            self.line_number = token[2][0]
+            for line in token[1].split('\n')[:-1]:
+                self.check_physical(line + '\n')
+                self.line_number += 1
+            self.multiline = False
+
     def check_all(self, expected=None, line_offset=0):
         """
         Run all checks on the input file.
diff --git a/testsuite/E50.py b/testsuite/E50.py
index 4cc4383..6d62b38 100644
--- a/testsuite/E50.py
+++ b/testsuite/E50.py
@@ -45,3 +45,26 @@ ddd = \
 ('''
     ''' + ' \
 ')
+#: E501 E225 E226
+very_long_identifiers=and_terrible_whitespace_habits(are_no_excuse+for_long_lines)
+#
+#: E501
+'''multiline string
+with a long long long long long long long long long long long long long long long long line
+'''
+#: E501
+'''same thing, but this time without a terminal newline in the string
+long long long long long long long long long long long long long long long long line'''
+#
+# issue 224 (unavoidable long lines in docstrings)
+#: Okay
+"""
+I'm some great documentation.  Because I'm some great documentation, I'm
+going to give you a reference to some valuable information about some API
+that I'm calling:
+
+    http://msdn.microsoft.com/en-us/library/windows/desktop/aa363858(v=vs.85).aspx
+"""
+#: E501
+"""
+longnospaceslongnospaceslongnospaceslongnospaceslongnospaceslongnospaceslongnospaceslongnospaces"""
diff --git a/testsuite/W19.py b/testsuite/W19.py
index c81f46f..2209e38 100644
--- a/testsuite/W19.py
+++ b/testsuite/W19.py
@@ -97,6 +97,28 @@ if length > options.max_line_length:
 if os.path.exists(os.path.join(path, PEP8_BIN)):
 	cmd = ([os.path.join(path, PEP8_BIN)] +
 	       self._pep8_options(targetfile))
+#: W191
+'''
+	multiline string with tab in it'''
+#: E101 W191
+'''multiline string
+	with tabs
+   and spaces
+'''
+#: Okay
+'''sometimes, you just need to go nuts in a multiline string
+	and allow all sorts of crap
+  like mixed tabs and spaces
+      
+or trailing whitespace  
+or long long long long long long long long long long long long long long long long long lines
+'''  # nopep8
+#: Okay
+'''this one
+	will get no warning
+even though the noqa comment is not immediately after the string
+''' + foo  # noqa
+#
 #: E101 W191
 if foo is None and bar is "frop" and \
         blah == 'yeah':
diff --git a/testsuite/W29.py b/testsuite/W29.py
index 2578f4f..42802ca 100644
--- a/testsuite/W29.py
+++ b/testsuite/W29.py
@@ -6,5 +6,8 @@ print
 class Foo(object):
     
     bang = 12
+#: W291
+'''multiline
+string with trailing whitespace'''   
 #: W292
 # This line doesn't have a linefeed
\ No newline at end of file
diff --git a/testsuite/W39no.py b/testsuite/W39no.py
new file mode 100644
index 0000000..2948b82
--- /dev/null
+++ b/testsuite/W39no.py
@@ -0,0 +1,6 @@
+#: Okay
+'''there is nothing wrong
+with a multiline string at EOF
+
+that happens to have a blank line in it
+'''