Batch ghostscript converter.

anntzer · anntzer · commit 5d8e1e0bcf9d · 2017-10-17T00:57:23.000-07:00
diff --git a/lib/matplotlib/__init__.py b/lib/matplotlib/__init__.py
@@ -334,10 +334,10 @@ def checkdep_ghostscript():
             gs_execs = ['gswin32c', 'gswin64c', 'mgs', 'gs']
         else:
             gs_execs = ['gs']
-        for gs_exec in gs_execs:
+        for gs_exec in map(str, gs_execs):
             try:
                 s = subprocess.Popen(
-                    [str(gs_exec), '--version'], stdout=subprocess.PIPE,
+                    [gs_exec, '--version'], stdout=subprocess.PIPE,
                     stderr=subprocess.PIPE)
                 stdout, stderr = s.communicate()
                 if s.returncode == 0:
diff --git a/lib/matplotlib/testing/compare.py b/lib/matplotlib/testing/compare.py
@@ -10,7 +10,6 @@
 import atexit
 import functools
 import hashlib
-import itertools
 import os
 import re
 import shutil
@@ -133,6 +132,13 @@ def convert(old, new):
     return convert
 
 
+try:
+    _fsencode = os.fsencode
+except AttributeError:  # Py2.
+    def _fsencode(s):
+        return s.encode(sys.getfilesystemencoding())
+
+
 # Modified from https://bugs.python.org/issue25567.
 _find_unsafe_bytes = re.compile(br'[^a-zA-Z0-9_@%+=:,./-]').search
 
@@ -142,38 +148,82 @@ def _shlex_quote_bytes(b):
             else b"'" + b.replace(b"'", b"'\"'\"'") + b"'")
 
 
-class _SVGConverter(object):
+class _ConverterError(Exception):
+    pass
+
+
+class _Converter(object):
     def __init__(self):
         self._proc = None
-        # We cannot rely on the GC to trigger `__del__` at exit because
-        # other modules (e.g. `subprocess`) may already have their globals
-        # set to `None`, which make `proc.communicate` or `proc.terminate`
-        # fail.  By relying on `atexit` we ensure the destructor runs before
-        # `None`-setting occurs.
+        # Explicitly register deletion from an atexit handler because if we
+        # wait until the object is GC'd (which occurs later), then some module
+        # globals (e.g. signal.SIGKILL) has already been set to None, and
+        # kill() doesn't work anymore...
         atexit.register(self.__del__)
 
-    def _read_to_prompt(self):
-        """Did Inkscape reach the prompt without crashing?
-        """
-        stream = iter(functools.partial(self._proc.stdout.read, 1), b"")
-        prompt = (b"\n", b">")
-        n = len(prompt)
-        its = itertools.tee(stream, n)
-        for i, it in enumerate(its):
-            next(itertools.islice(it, i, i), None)  # Advance `it` by `i`.
+    def __del__(self):
+        if self._proc:
+            self._proc.kill()
+            self._proc.wait()
+            for stream in filter(None, [self._proc.stdin,
+                                        self._proc.stdout,
+                                        self._proc.stderr]):
+                stream.close()
+            self._proc = None
+
+    def _read_until(self, terminator):
+        """Read until the prompt is reached."""
+        terminator = [six.int2byte(c) for c in six.iterbytes(terminator)]
+        buf = []
         while True:
-            window = tuple(map(next, its))
-            if len(window) != n:
-                # Ran out of data -- one of the `next(it)` raised
-                # StopIteration, so the tuple is shorter.
-                return False
-            if self._proc.poll() is not None:
-                # Inkscape exited.
-                return False
-            if window == prompt:
-                # Successfully read until prompt.
-                return True
+            c = self._proc.stdout.read(1)
+            if not c:
+                raise _ConverterError
+            buf.append(c)
+            if buf[-len(terminator):] == terminator:
+                return b"".join(buf[:-len(terminator)])
+
+
+class _GSConverter(_Converter):
+    def __call__(self, orig, dest):
+        if not self._proc:
+            self._stdout = TemporaryFile()
+            self._proc = subprocess.Popen(
+                [matplotlib.checkdep_ghostscript.executable,
+                 "-dNOPAUSE", "-sDEVICE=png16m"],
+                # As far as I can see, ghostscript never outputs to stderr.
+                stdin=subprocess.PIPE, stdout=subprocess.PIPE)
+            try:
+                self._read_until(b"\nGS")
+            except _ConverterError:
+                raise OSError("Failed to start Ghostscript")
+
+        def encode_and_escape(name):
+            return (_fsencode(name)
+                    .replace(b"(", br"\(")
+                    .replace(b")", br"\)")
+                    .replace(b"\\", b"\\\\"))
+
+        self._proc.stdin.write(
+            b"<< /OutputFile ("
+            + encode_and_escape(dest)
+            + b") >> setpagedevice ("
+            + encode_and_escape(orig)
+            + b") run flush\n")
+        self._proc.stdin.flush()
+        # GS> if nothing left on the stack; GS<n> if n items left on the stack.
+        err = self._read_until(b"GS")
+        stack = self._read_until(b">")
+        if stack or not os.path.exists(dest):
+            if stack:
+                self._proc.stdin.write(b"pop\n" * int(stack[1:]))
+            # Using the systemencoding should at least get the filenames right.
+            raise ImageComparisonFailure(
+                (err + b"GS" + stack + b">")
+                .decode(sys.getfilesystemencoding(), "replace"))
+
 
+class _SVGConverter(_Converter):
     def __call__(self, orig, dest):
         if (not self._proc  # First run.
                 or self._proc.poll() is not None):  # Inkscape terminated.
@@ -191,23 +241,21 @@ def __call__(self, orig, dest):
             # seem to sometimes deadlock when stderr is redirected to a pipe,
             # so we redirect it to a temporary file instead.  This is not
             # necessary anymore as of Inkscape 0.92.1.
-            self._stderr = TemporaryFile()
+            stderr = TemporaryFile()
             self._proc = subprocess.Popen(
                 [str("inkscape"), "--without-gui", "--shell"],
                 stdin=subprocess.PIPE, stdout=subprocess.PIPE,
-                stderr=self._stderr, env=env)
-            if not self._read_to_prompt():
-                raise OSError("Failed to start Inkscape")
-
-        try:
-            fsencode = os.fsencode
-        except AttributeError:  # Py2.
-            def fsencode(s):
-                return s.encode(sys.getfilesystemencoding())
+                stderr=stderr, env=env)
+            # Slight abuse, but makes shutdown handling easier.
+            self._proc.stderr = stderr
+            try:
+                self._read_until(b"\n>")
+            except _ConverterError:
+                raise OSError("Failed to start Inkscape in interactive mode")
 
         # Inkscape uses glib's `g_shell_parse_argv`, which has a consistent
         # behavior across platforms, so we can just use `shlex.quote`.
-        orig_b, dest_b = map(_shlex_quote_bytes, map(fsencode, [orig, dest]))
+        orig_b, dest_b = map(_shlex_quote_bytes, map(_fsencode, [orig, dest]))
         if b"\n" in orig_b or b"\n" in dest_b:
             # Who knows whether the current folder name has a newline, or if
             # our encoding is even ASCII compatible...  Just fall back on the
@@ -217,35 +265,22 @@ def fsencode(s):
                 str('inkscape'), '-z', old, '--export-png', new])(orig, dest)
         self._proc.stdin.write(orig_b + b" --export-png=" + dest_b + b"\n")
         self._proc.stdin.flush()
-        if not self._read_to_prompt():
-            # Inkscape's output is not localized but gtk's is, so the
-            # output stream probably has a mixed encoding.  Using
-            # `getfilesystemencoding` should at least get the filenames
-            # right...
+        try:
+            self._read_until(b"\n>")
+        except _ConverterError:
+            # Inkscape's output is not localized but gtk's is, so the output
+            # stream probably has a mixed encoding.  Using the filesystem
+            # encoding should at least get the filenames right...
             self._stderr.seek(0)
             raise ImageComparisonFailure(
                 self._stderr.read().decode(
                     sys.getfilesystemencoding(), "replace"))
 
-    def __del__(self):
-        if self._proc:
-            if self._proc.poll() is None:  # Not exited yet.
-                self._proc.communicate(b"quit\n")
-                self._proc.wait()
-            self._proc.stdin.close()
-            self._proc.stdout.close()
-            self._stderr.close()
-
 
 def _update_converter():
     gs, gs_v = matplotlib.checkdep_ghostscript()
     if gs_v is not None:
-        def cmd(old, new):
-            return [str(gs), '-q', '-sDEVICE=png16m', '-dNOPAUSE', '-dBATCH',
-             '-sOutputFile=' + new, old]
-        converter['pdf'] = make_external_conversion_command(cmd)
-        converter['eps'] = make_external_conversion_command(cmd)
-
+        converter['pdf'] = converter['eps'] = _GSConverter()
     if matplotlib.checkdep_inkscape() is not None:
         converter['svg'] = _SVGConverter()
 
