Buffer overflow in xkcd example with pdf backend.

[jenshnielsen]

I am seeing a buffer overflow bug when building the docs with pdf files in the XKCD example.

import matplotlib.pyplot as plt
import numpy as np

with plt.xkcd():
    # Based on "The Data So Far" from XKCD by Randall Monroe
    # http://xkcd.com/373/

    fig = plt.figure()
    ax = fig.add_axes((0.1, 0.2, 0.8, 0.7))
    ax.bar([-0.125, 1.0 - 0.125], [0, 100], 0.25)
    ax.spines['right'].set_color('none')
    ax.spines['top'].set_color('none')
    ax.xaxis.set_ticks_position('bottom')
    ax.set_xticks([0, 1])
    ax.set_xlim([-0.5, 1.5])
    ax.set_ylim([0, 110])
    ax.set_xticklabels(['CONFIRMED BY\nEXPERIMENT', 'REFUTED BY\nEXPERIMENT'])
    plt.yticks([])

    plt.title("CLAIMS OF SUPERNATURAL POWERS")

    fig.text(
        0.5, 0.05,
        '"The Data So Far" from xkcd by Randall Monroe',
        ha='center')
    plt.savefig('test.pdf')

Results in:

Traceback (most recent call last):
  File "xkcd.py", line 26, in <module>
    plt.savefig('test.pdf')
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/pyplot.py", line 576, in savefig
    res = fig.savefig(*args, **kwargs)
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/figure.py", line 1490, in savefig
    self.canvas.print_figure(*args, **kwargs)
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/backend_bases.py", line 2213, in print_figure
    **kwargs)
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/backends/backend_pdf.py", line 2509, in print_pdf
    self.figure.draw(renderer)
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/artist.py", line 60, in draw_wrapper
    draw(artist, renderer, *args, **kwargs)
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/figure.py", line 1032, in draw
    self.patch.draw(renderer)
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/artist.py", line 60, in draw_wrapper
    draw(artist, renderer, *args, **kwargs)
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/patches.py", line 484, in draw
    renderer.draw_path(gc, tpath, affine, rgbFace)
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/patheffects.py", line 107, in draw_path
    rgbFace)
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/patheffects.py", line 208, in draw_path
    Stroke.draw_path(self, renderer, gc, tpath, affine, rgbFace)
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/patheffects.py", line 197, in draw_path
    renderer.draw_path(gc0, tpath, trans, rgbFace)
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/backends/backend_pdf.py", line 1611, in draw_path
    gc.get_sketch_params())
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/backends/backend_pdf.py", line 1416, in writePath
    sketch=sketch)
  File "/Users/jhn/Envs/mplstable/lib/python3.4/site-packages/matplotlib-1.5.dev1-py3.4-macosx-10.10-x86_64.egg/matplotlib/backends/backend_pdf.py", line 1406, in pathOperations
    True))]
MemoryError: Buffer overflow

I have bisected this down to the merge of #4197 specifically cc41bff

[tacaswell]

I'm thinking this is the 'correct' behavior and the buffer just needs to be bigger.

[jenshnielsen]

Sounds plausible. I have not yet had a chance to dig into that code but it makes sense that a larger buffer is needed for xkcd plots

[mdboom]

Doh. That's right. I'll work on a solution.