matplotlib · jkseppan · Jan 12, 2018 · Feb 16, 2018 · Feb 18, 2018 · Jan 12, 2018
diff --git a/doc/api/next_api_changes/2018-02-16-JKS.rst b/doc/api/next_api_changes/2018-02-16-JKS.rst
@@ -0,0 +1,8 @@
+dviread changes
+---------------
+
+The ``format`` keyword argument to ``dviread.find_tex_file`` has been
+deprecated. The function without the ``format`` argument, as well as
+the new ``dviread.find_tex_files`` function, cache their results in
+``texsupport.N.db`` in the cache directory to speed up dvi file
+processing.
diff --git a/doc/users/next_whats_new/texsupport_cache.rst b/doc/users/next_whats_new/texsupport_cache.rst
@@ -0,0 +1,22 @@
+TeX support cache
+-----------------
+
+The `usetex` feature sends snippets of TeX code to LaTeX and related
+external tools for processing. This causes a nontrivial number of
+helper processes to be spawned, which can be slow on some platforms.
+A new cache database helps reduce the need to spawn these helper
+processes, which should improve `usetex` processing speed.
+
+The new cache files
+~~~~~~~~~~~~~~~~~~~
+
+The cache database is stored in a file named `texsupport.N.db` in the
+standard cache directory (traditionally `$HOME/.matplotlib` but
+possibly `$HOME/.cache/matplotlib`), where `N` stands for a version
+number. The version number is incremented when new kinds of items are
+added to the caching code, in order to avoid version clashes when
+using multiple different versions of Matplotlib. The auxiliary files
+`texsupport.N.db-wal` and `texsupport.N.db-shm` help coordinate usage
+of the cache between concurrently running instances. All of these
+cache files may be deleted when Matplotlib is not running, and
+subsequent calls to the `usetex` code will recompute the TeX results.
diff --git a/doc/users/whats_new.rst b/doc/users/whats_new.rst
@@ -14,12 +14,12 @@ revision, see the :ref:`github-stats`.
 ..
    For a release, add a new section after this, then comment out the include
    and toctree below by indenting them. Uncomment them after the release.
-   .. include:: next_whats_new/README.rst
-   .. toctree::
-      :glob:
-      :maxdepth: 1
+.. include:: next_whats_new/README.rst
+.. toctree::
+   :glob:
+   :maxdepth: 1
 
-      next_whats_new/*
+   next_whats_new/*
 
 
 New in Matplotlib 2.2

diff --git a/lib/matplotlib/dviread.py b/lib/matplotlib/dviread.py
diff --git a/lib/matplotlib/tests/baseline_images/dviread/broken1.dvi b/lib/matplotlib/tests/baseline_images/dviread/broken1.dvi
diff --git a/lib/matplotlib/tests/baseline_images/dviread/broken2.dvi b/lib/matplotlib/tests/baseline_images/dviread/broken2.dvi
diff --git a/lib/matplotlib/tests/baseline_images/dviread/broken3.dvi b/lib/matplotlib/tests/baseline_images/dviread/broken3.dvi
diff --git a/lib/matplotlib/tests/baseline_images/dviread/broken4.dvi b/lib/matplotlib/tests/baseline_images/dviread/broken4.dvi
diff --git a/lib/matplotlib/tests/baseline_images/dviread/broken5.dvi b/lib/matplotlib/tests/baseline_images/dviread/broken5.dvi
diff --git a/lib/matplotlib/tests/baseline_images/dviread/virtual.vf b/lib/matplotlib/tests/baseline_images/dviread/virtual.vf
diff --git a/lib/matplotlib/tests/baseline_images/dviread/virtual.vpl b/lib/matplotlib/tests/baseline_images/dviread/virtual.vpl
@@ -0,0 +1,23 @@
+(FAMILY TESTING)
+(COMMENT Test data for matplotlib)
+(COMMENT Run vptovf virtual.vpl to obtain virtual.vf)
+(FACE O 352)
+(CODINGSCHEME TEX TEXT)
+(DESIGNSIZE R 10.0)
+(FONTDIMEN
+   (SLANT R 0.0)
+   (SPACE R 0.333334)
+   (STRETCH R 0.166667)
+   (SHRINK R 0.111112)
+   (XHEIGHT R 0.430555)
+   (QUAD R 1.000003)
+   (EXTRASPACE R 0.111112)
+   )
+(MAPFONT D 0
+   (FONTNAME cmr10)
+   (FONTDSIZE R 10.0)
+   )
+(MAPFONT D 1
+   (FONTNAME cmex10)
+   (FONTDSIZE R 10.0)
+   )
diff --git a/lib/matplotlib/tests/test_dviread.py b/lib/matplotlib/tests/test_dviread.py
@@ -1,9 +1,16 @@
 from matplotlib.testing.decorators import skip_if_command_unavailable
 
+try:
+    from unittest import mock
+except ImportError:
+    import mock
+
 import matplotlib.dviread as dr
 import os.path
 import json
 import pytest
+import sqlite3
+import warnings
 
 
 def test_PsfontsMap(monkeypatch):
@@ -68,3 +75,119 @@ def test_dviread():
                  'boxes': [[b.x, b.y, b.height, b.width] for b in page.boxes]}
                 for page in dvi]
     assert data == correct
+
+
+@skip_if_command_unavailable(["kpsewhich", "-version"])
+def test_dviread_get_fonts():
+    dir = os.path.join(os.path.dirname(__file__), 'baseline_images', 'dviread')
+    dvi = dr._DviReader(os.path.join(dir, 'test.dvi'), None)
+    assert dvi.fontnames == \
+        {'cmex10', 'cmmi10', 'cmmi5', 'cmr10', 'cmr5', 'cmr7'}
+    vf = dr.Vf(os.path.join(dir, 'virtual.vf'))
+    assert vf.fontnames == {'cmex10', 'cmr10'}
+
+
+def test_dviread_get_fonts_error_handling():
+    dir = os.path.join(os.path.dirname(__file__), 'baseline_images', 'dviread')
+    for n, message in [(1, "too few 223 bytes"),
+                       (2, "post-postamble identification"),
+                       (3, "postamble offset"),
+                       (4, "postamble not found"),
+                       (5, "opcode 127 in postamble")]:
+        with pytest.raises(ValueError) as e:
+            dr.Dvi(os.path.join(dir, "broken%d.dvi" % n), None)
+        assert message in str(e.value)
+
+
+def test_TeXSupportCache(tmpdir):
+    dbfile = str(tmpdir / "test.db")
+    cache = dr.TeXSupportCache(filename=dbfile)
+    assert cache.get_pathnames(['foo', 'bar']) == {}
+    with cache.connection as transaction:
+        cache.update_pathnames({'foo': '/tmp/foo',
+                                'xyzzy': '/xyzzy.dat',
+                                'fontfile': None}, transaction)
+    assert cache.get_pathnames(['foo', 'bar']) == {'foo': '/tmp/foo'}
+    assert cache.get_pathnames(['xyzzy', 'fontfile']) == \
+        {'xyzzy': '/xyzzy.dat', 'fontfile': None}
+
+    # check that modifying a dvi file invalidates the cache
+    filename = str(tmpdir / "file.dvi")
+    with open(filename, "wb") as f:
+        f.write(b'qwerty')
+    os.utime(filename, (0, 0))
+    with cache.connection as t:
+        id1 = cache.dvi_new_file(filename, t)
+    assert cache.dvi_id(filename) == id1
+
+    with open(filename, "wb") as f:
+        f.write(b'asfdg')
+    os.utime(filename, (0, 0))
+    assert cache.dvi_id(filename) is None
+    with cache.connection as t:
+        id2 = cache.dvi_new_file(filename, t)
+    assert cache.dvi_id(filename) == id2
+
+
+def test_TeXSupportCache_versioning(tmpdir):
+    dbfile = str(tmpdir / "test.db")
+    cache1 = dr.TeXSupportCache(dbfile)
+    with cache1.connection as transaction:
+        cache1.update_pathnames({'foo': '/tmp/foo'}, transaction)
+
+    with sqlite3.connect(dbfile, isolation_level="DEFERRED") as conn:
+        conn.executescript('PRAGMA user_version=1000000000;')
+
+    with pytest.raises(dr.TeXSupportCacheError):
+        cache2 = dr.TeXSupportCache(dbfile)
+
+
+def test_find_tex_files(tmpdir):
+    with mock.patch('matplotlib.dviread.subprocess.Popen') as mock_popen:
+        mock_proc = mock.Mock()
+        stdout = '{s}tmp{s}foo.pfb\n{s}tmp{s}bar.map\n'.\
+                 format(s=os.path.sep).encode('ascii')
+        mock_proc.configure_mock(**{'communicate.return_value': (stdout, b'')})
+        mock_popen.return_value = mock_proc
+
+        # first call uses the results from kpsewhich
+        cache = dr.TeXSupportCache(filename=str(tmpdir / "test.db"))
+        assert dr.find_tex_files(
+            ['foo.pfb', 'cmsy10.pfb', 'bar.tmp', 'bar.map'], cache) \
+            == {'foo.pfb': '{s}tmp{s}foo.pfb'.format(s=os.path.sep),
+                'bar.map': '{s}tmp{s}bar.map'.format(s=os.path.sep),
+                'cmsy10.pfb': None, 'bar.tmp': None}
+        assert mock_popen.called
+
+        # second call (subset of the first one) uses only the cache
+        mock_popen.reset_mock()
+        assert dr.find_tex_files(['foo.pfb', 'cmsy10.pfb'], cache) \
+            == {'foo.pfb': '{s}tmp{s}foo.pfb'.format(s=os.path.sep),
+                'cmsy10.pfb': None}
+        assert not mock_popen.called
+
+        # third call (includes more than the first one) uses kpsewhich again
+        mock_popen.reset_mock()
+        stdout = '{s}usr{s}local{s}cmr10.tfm\n'.\
+                 format(s=os.path.sep).encode('ascii')
+        mock_proc.configure_mock(**{'communicate.return_value': (stdout, b'')})
+        mock_popen.return_value = mock_proc
+        assert dr.find_tex_files(['foo.pfb', 'cmr10.tfm'], cache) == \
+            {'foo.pfb': '{s}tmp{s}foo.pfb'.format(s=os.path.sep),
+             'cmr10.tfm': '{s}usr{s}local{s}cmr10.tfm'.format(s=os.path.sep)}
+        assert mock_popen.called
+
+
+def test_find_tex_file_format():
+    with mock.patch('matplotlib.dviread.subprocess.Popen') as mock_popen:
+        mock_proc = mock.Mock()
+        stdout = b'/foo/bar/baz\n'
+        mock_proc.configure_mock(**{'communicate.return_value': (stdout, b'')})
+        mock_popen.return_value = mock_proc
+
+        warnings.filterwarnings(
+            'ignore',
+            'The format option to find_tex_file is deprecated.*',
+            UserWarning)
+        assert dr.find_tex_file('foobar', format='tfm') == '/foo/bar/baz'
+        assert mock_popen.called