IndexError when searching for fields in files with no newline at the of the file #119

Activity

1. mcepl commented on Jul 27, 2015

   

   mcepl

   on Jul 27, 2015

   Contributor

   More actions

   @aligrudi ... can I get your wisdom here, please?
2. aligrudi commented on Jul 27, 2015

   

   aligrudi

   on Jul 27, 2015

   Last edited by lieryan

   Member

   More actions

   Zach Dwiel notifications@github.com wrote:

    File "/home/ubuntu/anaconda/lib/python2.7/site-packages/rope/base/worder.py", line 131, in _find_word_start
       while current_offset >= 0 and self._is_id_char(current_offset):
     File "/home/ubuntu/anaconda/lib/python2.7/site-packages/rope/base/worder.py", line 163, in _is_id_char
       return self.code[offset].isalnum() or self.code[offset] == '_'
   IndexError: string index out of range
   
   

   I tried to go through and find the right place to fix it, but wasn't sure where to make the change. We don't want to change the FIle.read to enforce newline at end of file. I'm having trouble getting a test case that has this error.

   You may add a test case to WordRangeFinderTest in codeanalyzetest.py.

   Making this change seems to help:

   rope/base/worder.py:162
       def _is_id_char(self, offset):
   +        if offset >= len(self.code):
   +            return False
           return self.code[offset].isalnum() or self.code[offset] == '_'
   

   I'm not sure if this is the right way to go about fixing it though, it seems like it might need to be fixed somewhere else.

   The problem is that in is_from_aliased(), (end + 1) is passed to _find_word_end() to find the end of the next word (end is len(self.code) - 1 and end + 1 goes past self.code). I wonder if is_from_aliased() function should return False if an IndexError is raised (it already does so for ValueError). Does doing so fix the problem?

   Ali
   
3. mcepl

   added

   bugUnexpected or incorrect user-visible behaviorUnexpected or incorrect user-visible behavior

   on Aug 9, 2015
4. mcepl commented on Nov 2, 2026

   

   mcepl

   on Nov 2, 2026

   Contributor

   More actions

   This still reproduces on master (46ebf37) with Python 3.13, and it
   breaks rename as well as find_occurrences():

   import pathlib, tempfile
   from rope.base.project import Project
   from rope.contrib import findit
   from rope.refactor.rename import Rename
   
   root = pathlib.Path(tempfile.mkdtemp())
   (root / "x.py").write_text("from t import system\n\nsystem('hi')\n")
   (root / "t.py").write_text("from os import system")  # no trailing newline
   project = Project(str(root), ropefolder=None)
   resource = project.get_resource("x.py")
   offset = resource.read().rindex("system")
   findit.find_occurrences(project, resource, offset)  # IndexError
   Rename(project, resource, offset).get_changes("run")  # IndexError

   Both end in Worder._is_id_char() via is_from_aliased() and
   _find_word_start(), where self.code[offset] is read past the end of
   t.py, the module whose last token is the name. That matches
   @aligrudi's 2015 diagnosis above. is_from_aliased() now clamps
   as_end with min(..., len(self.code)), but that is still one past
   the last character, and it catches ValueError but not IndexError.
5. mcepl

   added a commit that references this issue on Nov 5, 2026

   Fix IndexError for a from-import at end of file

   ...

   This issue will close once commit d9f8f35 is merged into the 'master' branch.

   d9f8f35