Closes#8981

PR Summary

For images we do some pixel rounding, but if we are doing unsampled images (egimshow(..., interpolation='none')) and saving to a backend that allows rasters (svg, pdf), then we don't want to do the pixel rounding - that just makes the image sizes fractionally too large when that isn't needed at all if we don't resample.

All the test images that changed just model my a pixel, and all haveimshow(..., interpolation='none').

Old

New

code

importnumpyasnpimportmatplotlib.pyplotaspltfig,ax=plt.subplots()ax.plot([0,0,1,1,0], [0,1,1,0,0],color='r',lw=0.1)np.random.seed(19680801)ax.imshow(np.random.randn(16,16),cmap='Blues',extent=(0,1,0,1),interpolation='none',vmin=-1,vmax=1)ax.set_xlim(-0.1,1+0.1)ax.set_ylim(-0.1,1+0.1)

PR Checklist

Documentation and Tests

• Has pytest style unit tests (andpytest passes)
• Documentation is sphinx and numpydoc compliant (the docs shouldbuild without error).
• New plotting related features are documented with examples.

Release Notes

• New features are marked with a.. versionadded:: directive in the docstring and documented indoc/users/next_whats_new/
• API changes are marked with a.. versionchanged:: directive in the docstring and documented indoc/api/next_api_changes/
• Release notes conform with instructions innext_whats_new/README.rst ornext_api_changes/README.rst