Bug in multiple step horizontal histograms

[ycopin]

Title says it all (using matplotlib 1.3.1):

hist([randn(100),randn(100)],orientation='horizontal', histtype='step')

The 2nd histo is just going nuts. This works fine for traditional 'vertical' orientation and/or non-stepped histos.

[ycopin]

I think I found the bug. In axes.py, function hist (line 8467), the following piece of code

                if orientation == 'horizontal':
                    x, y = y, x

                xvals.append(x.copy())
                yvals.append(y.copy())

is wrong: the x-y exchange on the 1st histo will mess up the following ones. It should be replaced by:

                if orientation == 'horizontal':
                    xvals.append(y.copy())
                    yvals.append(x.copy())
                else:
                    xvals.append(x.copy())
                    yvals.append(y.copy())

[tacaswell]

Can you put in a PR with this fix? It would also be great if you could add a test.

[ycopin]

I'm sorry I'm not at ease w/ PR and tests (see issue #841). I was hopping a more instructed developer would just incorporate my simple bugfix... :-/

[tacaswell]

@neggert, can you take a look at this?

[neggert]

Yeah, I can put this in tomorrow. This should go in the v1.3.x branch?

[tacaswell]

Yes to 1.3.x.

We are starting to triage for the 1.4 release and it is unclear if 1.3.2
will ever happen, but off it is no extra work might as well put it on 1.3.x.
On Feb 21, 2014 8:38 PM, "Nic Eggert" notifications@github.com wrote:

Yeah, I can put this in tomorrow. This should go in the v1.3.x branch?

—
Reply to this email directly or view it on GitHubhttps://github.com//issues/2830#issuecomment-35790882
.

[tacaswell]

@neggert Can I pester you about this? If not, I can deal with it, but you are much more familiar with the histogram code.

[ycopin]

Thanks for the PR: I now have a better view of what's called a test (I didn't know you could just compare to a reference image).