Remove many unused variables. #13932

anntzer · 2019-04-11T17:06:18Z

They can be detected by removing the exclusion of F841 in .flake8.

AFM parsing was made slightly stricter as a consequence (see changelog).

Some other unused variables have been left in, either for symmetry in
the code or because their lack of use may possibly indicate a bug that
needs to be investigated.

PR Summary

PR Checklist

Has Pytest style unit tests
Code is Flake 8 compliant
New features are documented, with examples if plot related
Documentation is sphinx and numpydoc compliant
Added an entry to doc/users/next_whats_new/ if major new feature (follow instructions in README.rst there)
Documented in doc/api/api_changes.rst if API changed in a backward-incompatible way

dstansby · 2019-04-11T18:17:33Z

I'd be nervous about removing any of these that don't have coverage in the test suite, lets see what happens though.

timhoffm · 2019-04-13T14:58:55Z

doc/api/next_api_changes/2019-04-11-AL.rst

@@ -0,0 +1,5 @@
+Stricter parsing of the Composites section of AFM files


The PR title and commit message suggest that it does not have any effect on running code. However, here you state an API change. I'd feel much safer if this was a separate PR, not least because this PR is large.

anntzer · 2019-04-13T16:21:05Z

Reverted the changes to afm.py.

timhoffm

IMHO this PR is too big, and thus makes reviewing difficult. I just got 1/3 though and then had to stop for doing other things. My motivation to take it up again is quite low.

For reviewing it's better to have multiple smaller PRs. You could just do it for a limited number of files in the first step.

timhoffm · 2019-04-13T19:08:06Z

examples/axes_grid1/demo_axes_grid.py

-
-    # This only affects axes in first column and second row as share_all =
-    # False.
+        ax.imshow(Z, extent=extent, interpolation="nearest")


I loosely recall that there was a discussion if examples should have return values even though they are not used.

Advantage: The user sees that this returns an object that can be used further.
Disadvantage: In practical use, you would only define the return value if you actually need it. So the example does not represent how you would use Matplotlib in real life. And it's more cluttered than need be.

Personally, I'm fine with removing.

I hope no one is reading axes_grid demos to learn about basic mpl concepts such as artists being returned by most calls :)

timhoffm · 2019-04-13T19:17:48Z

lib/matplotlib/dviread.py

@@ -455,7 +455,7 @@ def _fnt_def_real(self, k, c, s, d, a, l):

    @_dispatch(247, state=_dvistate.pre, args=('u1', 'u4', 'u4', 'u4', 'u1'))
    def _pre(self, i, num, den, mag, k):
-        comment = self.file.read(k)
+        self.file.read(k)  # comment


This change removes semantics. Either use _comment = ... or make the comment message more clear.

that variable is not used, this is just reading a comment and throwing it away. slightly edited the comment about the comment :)

anntzer · 2019-04-14T19:31:00Z

Sounds fair re: PR size; split out changes to tests (which are actually the majority of changes) to #13957.

tacaswell · 2019-04-21T02:35:03Z

lib/matplotlib/backends/backend_svg.py

            attrib = {}
            # Must add "px" to workaround a Firefox bug
-            style['font-size'] = short_float_fmt(fontsize) + 'px'
-            style['font-family'] = str(fontfamily)
+            style['font-size'] = short_float_fmt(prop.get_size()) + 'px'


.get_size_in_points and .get_size are methods with identical bodies....

I know... planning to deprecate get_size_in_points but there are bugger issues with fontproperties anyways.

tacaswell · 2019-04-21T02:54:19Z

I think this has gotten much smaller from when @timhoffm first reviewed it.

efiring · 2019-04-21T20:37:08Z

lib/matplotlib/backends/backend_gtk3.py

@@ -582,7 +582,7 @@ def configure_subplots(self, button):
        toolfig = Figure(figsize=(6, 3))
        canvas = self._get_canvas(toolfig)
        toolfig.subplots_adjust(top=0.9)
-        tool = SubplotTool(self.canvas.figure, toolfig)
+        SubplotTool(self.canvas.figure, toolfig)


The question this raises is, what keeps the SubplotTool instance alive when no reference to it is saved--not even until the end of this method? Presumably something external does get a reference, and I just don't see it. Otherwise, even without this change, the reference would last only until the end of the method.

I also don't see who is holding the reference.

It's actually a bit curious because the Tool object needs to stay alive even after the function has exited (so that the callbacks are not garbage-collected), but it's not clear to me what makes this possible.
In any case I checked, we do need to maintain a referrence to tool in case e.g. a gc collect()ion occurs in the body of the function, so I reverted this change and left a comment.

timhoffm

It would be good if we can find out who's holing the SubplotTool reference. Otherwise, this looks ok.

timhoffm · 2019-04-23T10:23:34Z

lib/matplotlib/backends/backend_macosx.py

@@ -160,6 +160,7 @@ def prepare_configure_subplots(self):
        toolfig = Figure(figsize=(6, 3))
        canvas = FigureCanvasMac(toolfig)
        toolfig.subplots_adjust(top=0.9)
+        # Need to keep a reference to the tool.
        tool = SubplotTool(self.canvas.figure, toolfig)


Suggested change

tool = SubplotTool(self.canvas.figure, toolfig)

_tool = SubplotTool(self.canvas.figure, toolfig)

The underscore prefix is commonly used as "intentionally unused"; e.g. Pycharm does not mark the variable as unused anymore with this change.

They can be detected by removing the exclusion of F841 in .flake8. Some other unused variables have been left in, either for symmetry in the code or because their lack of use may possibly indicate a bug that needs to be investigated.

anntzer force-pushed the unused branch 2 times, most recently from 041ddf1 to 088dbe1 Compare April 11, 2019 18:11

anntzer force-pushed the unused branch from 088dbe1 to 1a9e385 Compare April 11, 2019 22:49

timhoffm reviewed Apr 13, 2019

View reviewed changes

anntzer force-pushed the unused branch from 1a9e385 to 191ea63 Compare April 13, 2019 16:21

timhoffm reviewed Apr 14, 2019

View reviewed changes

anntzer force-pushed the unused branch from 191ea63 to c91e406 Compare April 14, 2019 19:29

anntzer mentioned this pull request Apr 14, 2019

Remove many unused variables in tests. #13957

Merged

6 tasks

tacaswell reviewed Apr 21, 2019

View reviewed changes

tacaswell added this to the v3.2.0 milestone Apr 21, 2019

tacaswell approved these changes Apr 21, 2019

View reviewed changes

efiring reviewed Apr 21, 2019

View reviewed changes

timhoffm approved these changes Apr 22, 2019

View reviewed changes

anntzer force-pushed the unused branch 2 times, most recently from 2b3ce74 to abd9b67 Compare April 23, 2019 07:58

timhoffm reviewed Apr 23, 2019

View reviewed changes

Remove many unused variables.

c53017e

They can be detected by removing the exclusion of F841 in .flake8. Some other unused variables have been left in, either for symmetry in the code or because their lack of use may possibly indicate a bug that needs to be investigated.

anntzer force-pushed the unused branch from abd9b67 to c53017e Compare April 23, 2019 11:57

timhoffm merged commit 3ba7b19 into matplotlib:master Apr 23, 2019

anntzer deleted the unused branch April 23, 2019 13:40

		@@ -0,0 +1,5 @@
		Stricter parsing of the Composites section of AFM files

	tool = SubplotTool(self.canvas.figure, toolfig)
	_tool = SubplotTool(self.canvas.figure, toolfig)

Uh oh!

Remove many unused variables. #13932

Remove many unused variables. #13932

Uh oh!

Conversation

anntzer commented Apr 11, 2019

PR Summary

PR Checklist

Uh oh!

dstansby commented Apr 11, 2019

Uh oh!

Choose a reason for hiding this comment

Uh oh!

anntzer commented Apr 13, 2019

Uh oh!

timhoffm left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

anntzer commented Apr 14, 2019

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

tacaswell commented Apr 21, 2019

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

timhoffm left a comment

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!