From 5ed05e08b1f71e663316a4b2c8b1c84967c22701 Mon Sep 17 00:00:00 2001
From: "James R. Barlow" <jim@purplerock.ca>
Date: Fri, 9 Nov 2018 01:40:01 -0800
Subject: [PATCH] Fix "no languages" test and misuse of os.environ

---
 src/ocrmypdf/exec/tesseract.py | 31 +++++++++++++++++++++----------
 tests/test_tess4.py            | 22 +++++++++++++++++++---
 2 files changed, 40 insertions(+), 13 deletions(-)

diff --git a/src/ocrmypdf/exec/tesseract.py b/src/ocrmypdf/exec/tesseract.py
index 1dce012c..fde6b066 100644
--- a/src/ocrmypdf/exec/tesseract.py
+++ b/src/ocrmypdf/exec/tesseract.py
@@ -21,7 +21,7 @@ import shutil
 from functools import lru_cache
 from collections import namedtuple
 from textwrap import dedent
-from subprocess import CalledProcessError, TimeoutExpired, check_output, STDOUT
+from subprocess import CalledProcessError, TimeoutExpired, check_output, STDOUT, run, PIPE
 from contextlib import suppress
 
 from ..exceptions import MissingDependencyError, TesseractConfigError
@@ -92,22 +92,33 @@ def psm():
 
 @lru_cache(maxsize=1)
 def languages():
+    def lang_error(output):
+        msg = dedent("""Tesseract failed to report available languages.
+        Output from Tesseract:
+        -----------
+        """)
+        msg += output
+        print(msg, file=sys.stderr)
+
     args_tess = [
         'tesseract',
         '--list-langs'
     ]
     try:
-        langs = check_output(
-                args_tess, universal_newlines=True, stderr=STDOUT)
+        proc = run(
+            args_tess, universal_newlines=True, stdout=PIPE, stderr=STDOUT,
+            check=True
+        )
+        output = proc.stdout
     except CalledProcessError as e:
-        msg = dedent("""Tesseract failed to report available languages.
-        Output from Tesseract:
-        -----------
-        """)
-        msg += e.output
-        print(msg, file=sys.stderr)
+        lang_error(e.output)
         raise MissingDependencyError from e
-    return set(lang.strip() for lang in langs.splitlines()[1:])
+
+    header, *rest = output.splitlines()
+    if not header.startswith('List of available languages'):
+        lang_error(output)
+        raise MissingDependencyError
+    return set(lang.strip() for lang in rest)
 
 
 def tess_base_args(langs, engine_mode):
diff --git a/tests/test_tess4.py b/tests/test_tess4.py
index 859b3cbb..0af3b7e6 100644
--- a/tests/test_tess4.py
+++ b/tests/test_tess4.py
@@ -16,8 +16,9 @@
 # along with OCRmyPDF.  If not, see <http://www.gnu.org/licenses/>.
 
 import pytest
-from ocrmypdf.exceptions import ExitCode
+from ocrmypdf.exceptions import ExitCode, MissingDependencyError
 from ocrmypdf.exec import tesseract
+from ocrmypdf.helpers import fspath
 from ocrmypdf import pdfinfo
 import sys
 import os
@@ -43,6 +44,7 @@ def _ensure_tess4():
         tess4 = Path(os.environ['OCRMYPDF_TESS4'])
         assert tess4.is_file()
         env['PATH'] = tess4.parent + ':' + env['PATH']
+        env['OCRMYPDF_TESS4'] = os.environ['OCRMYPDF_TESS4']
         return env
 
     raise EnvironmentError("Can't find Tesseract 4")
@@ -56,9 +58,12 @@ def ensure_tess4():
 @contextmanager
 def modified_os_environ(env):
     old_env = os.environ.copy()
-    os.environ = env
+    os.environ.update(env)
     yield
-    os.environ = old_env
+    for key in env:
+        del os.environ[key]
+        if key in old_env:
+            os.environ[key] = old_env[key]
 
 
 def tess4_available():
@@ -149,3 +154,14 @@ def test_content_preservation(ensure_tess4, resources, outpdf):
     info = pdfinfo.PdfInfo(outpdf)
     page = info[0]
     assert len(page.images) > 1, "masks were rasterized"
+
+
+def test_no_languages(ensure_tess4, tmpdir):
+    env = ensure_tess4
+    (tmpdir / 'tessdata').mkdir()
+    env['TESSDATA_PREFIX'] = fspath(tmpdir)
+
+    with modified_os_environ(env):
+        with pytest.raises(MissingDependencyError):
+            tesseract.languages.cache_clear()
+            tesseract.languages()