venv: do not prepend a truncated shebang interpreter (#2203)

docs/changelog/2208.bugfix.rst

@@ -0,0 +1,2 @@
+Prevent tox from using a truncated interpreter when using
+``TOX_LIMITED_SHEBANG`` -- by :user:`jdknight`.

docs/config.rst

@@ -1097,9 +1097,9 @@ Handle interpreter directives with long lengths
 For systems supporting executable text files (scripts with a shebang), the
 system will attempt to parse the interpreter directive to determine the program
 to execute on the target text file. When ``tox`` prepares a virtual environment
-in a file container which has a large length (e.x. using Jenkins Pipelines), the
+in a file container which has a large length (e.g. using Jenkins Pipelines), the
 system might not be able to invoke shebang scripts which define interpreters
-beyond system limits (e.x. Linux as a limit of 128; ``BINPRM_BUF_SIZE``). To
+beyond system limits (e.g. Linux has a limit of 128; ``BINPRM_BUF_SIZE``). To
 workaround an environment which suffers from an interpreter directive limit, a
 user can bypass the system's interpreter parser by defining the
 ``TOX_LIMITED_SHEBANG`` environment variable before invoking ``tox``::

src/tox/venv.py

@@ -19,6 +19,9 @@
 
 from .config import DepConfig
 
+#: maximum parsed shebang interpreter length (see: prepend_shebang_interpreter)
+MAXINTERP = 2048
+
 
 class CreationConfig:
     def __init__(
@@ -672,7 +675,7 @@ def prepend_shebang_interpreter(args):
     #
     # When preparing virtual environments in a file container which has large
     # length, the system might not be able to invoke shebang scripts which
-    # define interpreters beyond system limits (e.x. Linux as a limit of 128;
+    # define interpreters beyond system limits (e.g. Linux has a limit of 128;
     # BINPRM_BUF_SIZE). This method can be used to check if the executable is
     # a script containing a shebang line. If so, extract the interpreter (and
     # possible optional argument) and prepend the values to the provided
@@ -682,8 +685,9 @@ def prepend_shebang_interpreter(args):
     try:
         with open(args[0], "rb") as f:
             if f.read(1) == b"#" and f.read(1) == b"!":
-                MAXINTERP = 2048
-                interp = f.readline(MAXINTERP).rstrip().decode("UTF-8")
+                interp = f.readline(MAXINTERP + 1).rstrip().decode("UTF-8")
+                if len(interp) > MAXINTERP:  # avoid a truncated interpreter
+                    return args
                 interp_args = interp.split(None, 1)[:2]
                 return interp_args + args
     except (UnicodeDecodeError, IOError):

tests/unit/test_venv.py

@@ -9,6 +9,7 @@
 from tox.interpreters import NoInterpreterInfo
 from tox.session.commands.run.sequential import installpkg, runtestenv
 from tox.venv import (
+    MAXINTERP,
     CreationConfig,
     VirtualEnv,
     getdigest,
@@ -1149,6 +1150,18 @@ def test_tox_testenv_interpret_shebang_long_example(tmpdir):
     assert args == expected + base_args
 
 
+@pytest.mark.skipif("sys.platform == 'win32'", reason="no shebang on Windows")
+def test_tox_testenv_interpret_shebang_skip_truncated(tmpdir):
+    testfile = tmpdir.join("check_shebang_truncation.py")
+    original_args = [str(testfile), "arg1", "arg2", "arg3"]
+
+    # interpreter (too long example)
+    testfile.write("#!" + ("x" * (MAXINTERP + 1)))
+    args = prepend_shebang_interpreter(original_args)
+
+    assert args == original_args
+
+
 @pytest.mark.parametrize("download", [True, False, None])
 def test_create_download(mocksession, newconfig, download):
     config = newconfig(