From 0708302464a6c280098e23a921ee7cd195cb7e0f Mon Sep 17 00:00:00 2001
From: "Brian C. Lane" <bcl@redhat.com>
Date: Wed, 10 Jul 2019 17:30:10 -0700
Subject: [PATCH] Use binary mode to tail the file

In python 3 f.seek() on text doesn't work like it does in py2/C because
text is now unicode. So change read_tail to use byte mode and take
unicode into account. Also add tests for it.
---
 src/pylorax/sysutils.py        | 37 +++++++++++------
 tests/pylorax/test_sysutils.py | 72 ++++++++++++++++++++++++++++++++++
 2 files changed, 98 insertions(+), 11 deletions(-)

diff --git a/src/pylorax/sysutils.py b/src/pylorax/sysutils.py
index c0bc4421..c445fae1 100644
--- a/src/pylorax/sysutils.py
+++ b/src/pylorax/sysutils.py
@@ -1,7 +1,7 @@
 #
 # sysutils.py
 #
-# Copyright (C) 2009-2015 Red Hat, Inc.
+# Copyright (C) 2009-2019 Red Hat, Inc.
 #
 # This program is free software; you can redistribute it and/or modify
 # it under the terms of the GNU General Public License as published by
@@ -133,13 +133,28 @@ def flatconfig(filename):
 
 def read_tail(path, size):
     """Read up to `size` kibibytes from the end of a file"""
-    with open(path, "r") as f:
-        f.seek(0, 2)
-        end = f.tell()
-        if end < 1024 * size:
-            f.seek(0, 0)
-        else:
-            f.seek(end - (1024 * size))
-        # Find the start of the next line and return the rest
-        f.readline()
-        return f.read()
+
+    # NOTE: In py3 text files are unicode, not bytes so we have to open it as bytes
+    with open(path, "rb") as f:
+        return _read_file_end(f, size)
+
+def _read_file_end(f, size):
+    """Read the end of a file
+
+    This skips to the next line to avoid starting in the middle of a unicode character.
+    And returns "" in the case of a UnicodeDecodeError
+    """
+    f.seek(0, 2)
+    end = f.tell()
+    if end < 1024 * size:
+        f.seek(0, 0)
+    else:
+        f.seek(end - (1024 * size))
+    data = f.read()
+    try:
+        # Find the first newline in the block
+        newline = min(1+data.find(b'\n'), len(data))
+        text = data[newline:].decode("UTF-8")
+    except UnicodeDecodeError:
+        return ""
+    return text
diff --git a/tests/pylorax/test_sysutils.py b/tests/pylorax/test_sysutils.py
index 8b33e7a0..818709d9 100644
--- a/tests/pylorax/test_sysutils.py
+++ b/tests/pylorax/test_sysutils.py
@@ -14,11 +14,13 @@
 # You should have received a copy of the GNU General Public License
 # along with this program.  If not, see <http://www.gnu.org/licenses/>.
 #
+import io
 import unittest
 import tempfile
 import os
 
 from pylorax.sysutils import joinpaths, touch, replace, chown_, chmod_, remove, linktree
+from pylorax.sysutils import _read_file_end
 
 class SysUtilsTest(unittest.TestCase):
     def joinpaths_test(self):
@@ -72,3 +74,73 @@ class SysUtilsTest(unittest.TestCase):
             linktree(os.path.join(tdname, "one"), os.path.join(tdname, "copy"))
 
             self.assertTrue(os.path.exists(os.path.join(tdname, "copy", "two", "three", "lorax-link-test-file")))
+
+    def _generate_lines(self, unicode=False):
+        # helper to generate several KiB of lines of text
+        bio = io.BytesIO()
+        for i in range(0,1024):
+            if not unicode:
+                bio.write(b"Here is another line to test. It is line #%d\n" % i)
+            else:
+                bio.write(b"Here is \xc3\xa0n\xc3\xb2ther line t\xc3\xb2 test. It is line #%d\n" % i)
+        bio.seek(0)
+        return bio
+
+    def read_file_end_test(self):
+        """Test reading from the end of a file"""
+        self.maxDiff = None
+
+        # file of just lines
+        f = self._generate_lines()
+
+        # Grab the end of the 'file' to compare with, starting at the next line (hard-coded)
+        f.seek(-987, 2)
+        result = f.read().decode("utf-8")
+        f.seek(0)
+        self.assertEqual(_read_file_end(f, 1), result)
+
+        # file of lines with no final \n, chop off the trailing \n
+        f.seek(-1,2)
+        f.truncate()
+        f.seek(0)
+        self.assertEqual(_read_file_end(f, 1), result[:-1])
+
+        # short file, truncate it at 1023 characters
+        f.seek(1023)
+        f.truncate()
+        # Grab the end of the file, starting at the next line (hard-coded)
+        f.seek(44)
+        result = f.read().decode("utf-8")
+        f.seek(0)
+        self.assertEqual(_read_file_end(f, 1), result)
+
+        # short file with no line endings
+        f.seek(43)
+        f.truncate()
+        # Grab the whole file
+        f.seek(0)
+        result = f.read().decode("utf-8")
+        f.seek(0)
+        self.assertEqual(_read_file_end(f, 1), result)
+
+        # file with unicode in it
+        f = self._generate_lines(unicode=True)
+
+        # Grab the end of the 'file' to compare with, starting at the next line (hard-coded)
+        f.seek(-1000, 2)
+        result = f.read().decode("utf-8")
+        f.seek(0)
+        self.assertEqual(_read_file_end(f, 1), result)
+
+        # file with unicode right on block boundary, so that a decode of it would fail if it didn't
+        # move to the next line.
+        f.seek(-1000, 2)
+        result = f.read().decode("utf-8")
+        f.seek(-1025, 2)
+        f.write(b"\xc3\xb2")
+        f.seek(0)
+        self.assertEqual(_read_file_end(f, 1), result)
+
+        # Test for UnicodeDecodeError returning an empty string
+        f = io.BytesIO(b"\xff\xff\xffHere is a string with invalid unicode in it.")
+        self.assertEqual(_read_file_end(f, 1), "")