ANN: matplotlib-0.99.3

We have uploaded a bugfix release of the matplotlib-0.99 branch.
Source and binaries are available for download from

  https://sourceforge.net/projects/matplotlib/files/matplotlib/matplotlib-0.99.3

This release fixes many minor and at least one critical bug in our
path simplification algorithm. The svn commit changelog since the
last release is below. Thanks to all the developers for their efforts
and to Christoph Gohlke for the windows builds.

Unless something major comes up, this will be our last bugfix release
of the maintenance branch, and we plan to get a trunk release for mpl
1.0 out in the upcoming weeks for testing. The trunk is where all of
the new features have been added, but it is considered less stable
than the maintenance branch which only gets bugfixes.

···

------------------------------------------------------------------------
r8175 | leejjoon | 2010-03-03 12:03:30 -0600 (Wed, 03 Mar 2010) | 1 line

fix arguments of allow_rasterization.draw_wrapper
------------------------------------------------------------------------
r8174 | jdh2358 | 2010-03-03 11:15:58 -0600 (Wed, 03 Mar 2010) | 1 line

added support for favicon in docs build
------------------------------------------------------------------------
r8173 | jdh2358 | 2010-03-03 10:56:16 -0600 (Wed, 03 Mar 2010) | 1 line

applied Mattias get_bounds patch
------------------------------------------------------------------------
r8172 | jdh2358 | 2010-03-03 10:31:42 -0600 (Wed, 03 Mar 2010) | 1 line

fix svnmerge download instructions
------------------------------------------------------------------------
r8171 | jdh2358 | 2010-03-03 09:47:48 -0600 (Wed, 03 Mar 2010) | 1 line

added favicon
------------------------------------------------------------------------
r8160 | mdboom | 2010-02-26 10:27:22 -0600 (Fri, 26 Feb 2010) | 2 lines

Fix offset_copy: the fig argument should be optional.

------------------------------------------------------------------------
r8146 | mdboom | 2010-02-22 10:22:28 -0600 (Mon, 22 Feb 2010) | 1 line

Backporting numpy version check fix to 0.99 branch
------------------------------------------------------------------------
r8145 | jdh2358 | 2010-02-22 08:31:45 -0600 (Mon, 22 Feb 2010) | 1 line

fix setters for regular polygon
------------------------------------------------------------------------
r8135 | leejjoon | 2010-02-16 16:55:27 -0600 (Tue, 16 Feb 2010) | 1 line

fix a bug in Text._get_layout that returns an incorrect information
for an empty string
------------------------------------------------------------------------
r8121 | jdh2358 | 2010-02-08 11:50:27 -0600 (Mon, 08 Feb 2010) | 1 line

added Ariels csd patch for proper scaling at the dc component
------------------------------------------------------------------------
r8116 | mdboom | 2010-02-08 09:57:45 -0600 (Mon, 08 Feb 2010) | 1 line

Fix for libpng-1.4 compatibility
------------------------------------------------------------------------
r8092 | leejjoon | 2010-01-18 18:26:16 -0600 (Mon, 18 Jan 2010) | 1 line

update annotate documentation to explain *annotation_clip* parameter
------------------------------------------------------------------------
r8070 | mdboom | 2010-01-04 08:28:57 -0600 (Mon, 04 Jan 2010) | 1 line

Fix doc 'clean'
------------------------------------------------------------------------
r8068 | mdboom | 2010-01-04 08:14:38 -0600 (Mon, 04 Jan 2010) | 2 lines

Fix bug in PDF, PS, SVG and OS-X backends: do not simplify filled paths.

------------------------------------------------------------------------
r8057 | mdboom | 2009-12-31 09:46:58 -0600 (Thu, 31 Dec 2009) | 2 lines

[2916753] Wrong API signature- yscale

------------------------------------------------------------------------
r8036 | jdh2358 | 2009-12-16 13:21:44 -0600 (Wed, 16 Dec 2009) | 1 line

add mpl book to index sidebar
------------------------------------------------------------------------
r8016 | heeres | 2009-12-09 18:09:03 -0600 (Wed, 09 Dec 2009) | 2 lines

Mplot3d: fix scatter3d markers bug

------------------------------------------------------------------------
r8003 | mdboom | 2009-12-03 13:21:28 -0600 (Thu, 03 Dec 2009) | 1 line

[2896454] subscripts and \mathrm don't get along, aka the
long-standing wiggle baseline problem
------------------------------------------------------------------------
r8001 | mdboom | 2009-12-03 12:54:54 -0600 (Thu, 03 Dec 2009) | 2 lines

[2903460] _path.pyd crashes in 0.99.1.win32-py2.6 on PIII CPU

------------------------------------------------------------------------
r7998 | mdboom | 2009-12-01 13:01:45 -0600 (Tue, 01 Dec 2009) | 2 lines

[2906157] Let sphinx reference matplotlib-created plots (backport)

------------------------------------------------------------------------
r7991 | mdboom | 2009-11-30 12:30:09 -0600 (Mon, 30 Nov 2009) | 2 lines

Fix bug in usage of "contains"

------------------------------------------------------------------------
r7989 | mdboom | 2009-11-30 11:24:49 -0600 (Mon, 30 Nov 2009) | 1 line

Fix SF-2903596 on the 0.99.x branch
------------------------------------------------------------------------
r7983 | mdboom | 2009-11-24 12:08:51 -0600 (Tue, 24 Nov 2009) | 2 lines

[2902715] artist.set_clip_path does not handle tuple

------------------------------------------------------------------------
r7981 | jdh2358 | 2009-11-23 15:16:15 -0600 (Mon, 23 Nov 2009) | 1 line

make IndexFormatter derive from Formatter
------------------------------------------------------------------------
r7972 | mdboom | 2009-11-18 10:19:21 -0600 (Wed, 18 Nov 2009) | 2 lines

Fix plot directive so it handles source files in various encodings.

------------------------------------------------------------------------
r7970 | efiring | 2009-11-16 17:49:11 -0600 (Mon, 16 Nov 2009) | 4 lines

Fix Normalize bug: ensure scalar output for scalar input.
This bug was causing a bug in clabel with a single contour level,
as reported by Brendan Arnold.

------------------------------------------------------------------------
r7952 | mdboom | 2009-11-12 11:27:34 -0600 (Thu, 12 Nov 2009) | 2 lines

[2853659] Patch to fix EMF backend

------------------------------------------------------------------------
r7944 | mdboom | 2009-11-06 13:24:55 -0600 (Fri, 06 Nov 2009) | 2 lines

Add support for \widebar{} in mathtext (Thanks Sean Arms)

------------------------------------------------------------------------
r7942 | jdh2358 | 2009-11-06 09:04:33 -0600 (Fri, 06 Nov 2009) | 1 line

backport afm fix to branch
------------------------------------------------------------------------
r7928 | mdboom | 2009-11-03 14:55:47 -0600 (Tue, 03 Nov 2009) | 1 line

[2889193] Rasterized output offset in pdf and ps
------------------------------------------------------------------------
r7924 | mdboom | 2009-11-03 10:27:13 -0600 (Tue, 03 Nov 2009) | 2 lines

[2880836] Save from toolbar changes directory with WX backend

------------------------------------------------------------------------
r7919 | mdboom | 2009-11-03 09:46:14 -0600 (Tue, 03 Nov 2009) | 2 lines

[2890979] Close paths correctly in PolyCollection

------------------------------------------------------------------------
r7916 | mdboom | 2009-11-02 12:54:29 -0600 (Mon, 02 Nov 2009) | 2 lines

[2890345] pylab.thetagrids() does not accept fmt

------------------------------------------------------------------------
r7906 | leejjoon | 2009-10-25 14:30:43 -0500 (Sun, 25 Oct 2009) | 2 lines

axes_divider.py : fix a bug that axes has a wrong size when pack_start is True.

------------------------------------------------------------------------
r7903 | mdboom | 2009-10-23 10:02:21 -0500 (Fri, 23 Oct 2009) | 2 lines

Support \# in mathtext.

------------------------------------------------------------------------
r7901 | mdboom | 2009-10-22 08:46:08 -0500 (Thu, 22 Oct 2009) | 2 lines

Fix path simplification so the path always starts with a MOVETO.

------------------------------------------------------------------------
r7896 | mdboom | 2009-10-21 08:24:33 -0500 (Wed, 21 Oct 2009) | 2 lines

Fix bug in simplification code that was clipping peaks.

------------------------------------------------------------------------
r7893 | leejjoon | 2009-10-19 04:02:08 -0500 (Mon, 19 Oct 2009) | 1 line

fix a bug in plot_directive
------------------------------------------------------------------------
r7887 | leejjoon | 2009-10-15 23:35:20 -0500 (Thu, 15 Oct 2009) | 1 line

fix Text.get_prop_tup omitting _rotation_mode. patch by Stan West
------------------------------------------------------------------------
r7884 | ryanmay | 2009-10-14 15:27:07 -0500 (Wed, 14 Oct 2009) | 1 line

Add AutoDateFormatter and AutoDateLocator to __all__ and to the
module-level documentation.
------------------------------------------------------------------------
r7880 | leejjoon | 2009-10-12 18:02:27 -0500 (Mon, 12 Oct 2009) | 1 line

fix plot_directive.py that generated incorrect links for some cases
------------------------------------------------------------------------
r7874 | mdboom | 2009-10-12 08:21:07 -0500 (Mon, 12 Oct 2009) | 2 lines

Don't simplify hatched paths in PDF, PS and SVG backends.

------------------------------------------------------------------------
r7867 | astraw | 2009-10-11 13:55:18 -0500 (Sun, 11 Oct 2009) | 2 lines

imshow: only apply axes patch clipping if image doesn't clip itself

------------------------------------------------------------------------
r7866 | astraw | 2009-10-11 13:54:22 -0500 (Sun, 11 Oct 2009) | 2 lines

allow already transformed clip path in artist set_clip_path()

------------------------------------------------------------------------
r7864 | efiring | 2009-10-10 21:51:06 -0500 (Sat, 10 Oct 2009) | 2 lines

docstring fixup: LinearSegmentedColormap.from_list

------------------------------------------------------------------------
r7858 | mdboom | 2009-10-09 10:24:25 -0500 (Fri, 09 Oct 2009) | 2 lines

Don't perform path simplification when hatch is drawn.

------------------------------------------------------------------------
r7849 | mdboom | 2009-10-06 09:07:08 -0500 (Tue, 06 Oct 2009) | 2 lines

Don't do "make.py clean" in terms of svn-clean, since that doesn't
work with a tarball.

------------------------------------------------------------------------
r7847 | efiring | 2009-10-05 18:11:10 -0500 (Mon, 05 Oct 2009) | 2 lines

Add more platforms to setupext.py; patch by Benjamin Drung

------------------------------------------------------------------------
r7846 | efiring | 2009-10-05 17:44:38 -0500 (Mon, 05 Oct 2009) | 8 lines

Tk directory-finding patch by Benjamin Drung.

Fixes the following:
Guess the correct directory in Debian/Ubuntu fails. Replacing all tcl by
tk does not lead to a correct directory:
/usr/share/tcltk/tcl8.5 -> /usr/share/tktk/tk8.5
The correct directory is /usr/share/tcltk/tk8.5

------------------------------------------------------------------------
r7844 | leejjoon | 2009-10-04 21:05:16 -0500 (Sun, 04 Oct 2009) | 1 line

image created by imshow support remove method
------------------------------------------------------------------------
r7837 | mdboom | 2009-10-01 07:37:49 -0500 (Thu, 01 Oct 2009) | 2 lines

Fix alpha of Shadow patch.

------------------------------------------------------------------------
r7834 | mdboom | 2009-09-30 08:26:36 -0500 (Wed, 30 Sep 2009) | 2 lines

Fix Grouper docstring -- it only supports weak-referenceable objects.

------------------------------------------------------------------------
r7827 | efiring | 2009-09-25 18:56:14 -0500 (Fri, 25 Sep 2009) | 2 lines

Fix bug in quiver handling of masked U, V

------------------------------------------------------------------------
r7822 | mdboom | 2009-09-24 12:03:13 -0500 (Thu, 24 Sep 2009) | 2 lines

Clip markers in draw_marker outside of the image rectangle with an
extra border to accomodate the size of the marker.

------------------------------------------------------------------------
r7819 | jdh2358 | 2009-09-23 22:57:31 -0500 (Wed, 23 Sep 2009) | 1 line

applied a fix for sf bug 2865490; is there something more elegant?
------------------------------------------------------------------------
r7818 | jdh2358 | 2009-09-23 22:00:28 -0500 (Wed, 23 Sep 2009) | 1 line

try statically linking in the deps for OSX
------------------------------------------------------------------------
r7813 | jdh2358 | 2009-09-21 12:12:47 -0500 (Mon, 21 Sep 2009) | 1 line

For the binaries, what is the recommended NumPy version?
I assume the current release (1.4.1)?

(Also wondering, not that it's any of my business, whether the
current refactoring project for NumPy has implications for the
upcoming MPL 1.0 relase.)

Thanks!
Alan Isaac

PS Btw, it gets said often but once more won't hurt:
Matplotlib is wonderful!

···

On 6/1/2010 9:06 AM, John Hunter wrote:

We have uploaded a bugfix release of the matplotlib-0.99 branch.

I just wanted to second this. Having suffered many years
with super mongo, matplotlib (and python in general) has
made my work so much easier.

Tommy Grav

···

On Jun 1, 2010, at 9:34 AM, Alan G Isaac wrote:

PS Btw, it gets said often but once more won't hurt:
Matplotlib is wonderful!

+----------------------------------------------------------------------------+
Associate Researcher @ Dept. of Physics and Astronomy
Johns Hopkins University
+----------------------------------------------------------------------------+
tgrav@...3130...
(410) 516-7683
http://web.mac.com/tgrav/Astronomy/Welcome.html
+----------------------------------------------------------------------------+

For the binaries, what is the recommended NumPy version?
I assume the current release (1.4.1)?

Are you on windows? The windows binaries are build against numpy
1.3.0 so any and numpy since then should be good since they are
forward but not backward compatible. At lease that is my
understanding. The OSX binaries are built against numpy 1.2.1 so any
numpy after than should work.

(Also wondering, not that it's any of my business, whether the
current refactoring project for NumPy has implications for the
upcoming MPL 1.0 relase.)

Are you referring to the refactoring to try and make numpy more
accessible as a C library? That has no implications for 1.0, since
we are hoping to get that out soon and the numpy project is
multi-month at least. In any case, mpl 1.0 should be able to be
compiled against all the recent numpys in the wild.

PS Btw, it gets said often but once more won't hurt:
Matplotlib is wonderful!

Thanks! Jae-Joon Lee, Michael Droettboom, Ryan May, and Eric Firing
have been doing the heavy lifting on the regular development work, and
we continue to see lots of fantastic contributions from users. I'm
just trying to keep up :slight_smile: There is really a lot of fantastic stuff
in the upcoming 1.0 so I look forward to getting that out soon.

JDH

···

On Tue, Jun 1, 2010 at 8:34 AM, Alan G Isaac <alan.isaac@...287...> wrote:

On 6/1/2010 9:06 AM, John Hunter wrote:

I just want confirm because I new here

on my server I have this scpy 0.7.2, numpy 1.4.1, matplotlib 0.99.1.1.

may I keep that?

thanks

2010/6/1 John Hunter <jdh2358@…287…>

···

On Tue, Jun 1, 2010 at 8:34 AM, Alan G Isaac <alan.isaac@…287…> wrote:

On 6/1/2010 9:06 AM, John Hunter wrote:

For the binaries, what is the recommended NumPy version?

I assume the current release (1.4.1)?

Are you on windows? The windows binaries are build against numpy

1.3.0 so any and numpy since then should be good since they are

forward but not backward compatible. At lease that is my

understanding. The OSX binaries are built against numpy 1.2.1 so any

numpy after than should work.

(Also wondering, not that it’s any of my business, whether the

current refactoring project for NumPy has implications for the

upcoming MPL 1.0 relase.)

Are you referring to the refactoring to try and make numpy more

accessible as a C library? That has no implications for 1.0, since

we are hoping to get that out soon and the numpy project is

multi-month at least. In any case, mpl 1.0 should be able to be

compiled against all the recent numpys in the wild.

PS Btw, it gets said often but once more won’t hurt:

Matplotlib is wonderful!

Thanks! Jae-Joon Lee, Michael Droettboom, Ryan May, and Eric Firing

have been doing the heavy lifting on the regular development work, and

we continue to see lots of fantastic contributions from users. I’m

just trying to keep up :slight_smile: There is really a lot of fantastic stuff

in the upcoming 1.0 so I look forward to getting that out soon.

JDH



Matplotlib-users mailing list

Matplotlib-users@lists.sourceforge.net

https://lists.sourceforge.net/lists/listinfo/matplotlib-users

You can, but I recommend upgrading mpl to 0.99.3 or setting
'path.simplify : False' in your matplotlibrc because there is a
serious bug in the path simplification algorithm in 0.99.1 that was
fixed in 0.99.3. You can work around the bug by turning off
path.simplify or you can upgrade your mpl.

path.simplify is an rc parameter which can be customized according the
instructions at
http://matplotlib.sourceforge.net/users/customizing.html

JDH

···

On Tue, Jun 1, 2010 at 9:09 AM, Samuel Teixeira Santos <arcanosam@...287...> wrote:

I just want confirm because I new here

on my server I have this scpy 0.7.2, numpy 1.4.1, matplotlib 0.99.1.1.

I will prefer update it very soon

thanks for the response

2010/6/1 John Hunter <jdh2358@…287…>

···

On Tue, Jun 1, 2010 at 9:09 AM, Samuel Teixeira Santos > > <arcanosam@…287…> wrote:

I just want confirm because I new here

on my server I have this scpy 0.7.2, numpy 1.4.1, matplotlib 0.99.1.1.

You can, but I recommend upgrading mpl to 0.99.3 or setting

‘path.simplify : False’ in your matplotlibrc because there is a

serious bug in the path simplification algorithm in 0.99.1 that was

fixed in 0.99.3. You can work around the bug by turning off

path.simplify or you can upgrade your mpl.

path.simplify is an rc parameter which can be customized according the

instructions at

http://matplotlib.sourceforge.net/users/customizing.html

JDH