quiver is not plotting only 1 vector

[moghimis]

For some reason my old script which was plotting correctly is not reproducing the same figure.
Here is the line of the code:

axq.quiver([xtexth],[ytexth],[uwnd,],[vwnd,],units=units,scale=scalew,width=0.03,color='b')

However if I add 1 element to vector it will plot as before:

axq.quiver([0,xtexth],[0,ytexth],[0,uwnd,],[0,vwnd,],units=units,scale=scalew,width=0.03,color='b')

It seems the first element is ommited.

Saeed.

In [178]: import matplotlib as mpl
In [179]: print mpl.__version__
1.4.2

[TAC edited to add formatting markup]

[tacaswell]

I can not reproduce this on master, but I had to guess at the values of all of the inputs.

Can you provide a complete minimal example of code that does not work (including values, it should be copy-paste runnable).

[moghimis]

Here is my example:

worked

f=figure(1)
axq=f.add_subplot(111)
axq.quiver([5,],[5,],[5,],[5,],pivot='left',color='k')
axq.set_xlim(0,10)
axq.set_ylim(0,10)
plt.show()

works now

f=figure(2)
axq=f.add_subplot(111)
axq.quiver([0,5,],[0,5,],[0,5,],[0,5,],pivot='left',color='k')
axq.set_xlim(0,10)
axq.set_ylim(0,10)
plt.show()

In older version both are working but the scale are differrent.

On Wed, Nov 19, 2014 at 7:00 PM, Thomas A Caswell notifications@github.com
wrote:

I can not reproduce this on master, but I had to guess at the values of
all of the inputs.

Can you provide a complete minimal example of code that does not work
(including values, it should be copy-paste runnable).

—
Reply to this email directly or view it on GitHub
#3817 (comment)
.

---

Saeed Moghimi
College of Earth, Ocean, and Atmospheric Sciences,
Oregon State University,
Corvallis, OR 97331-5503
Office: Burt 104
Email: moghimi@coas.oregonstate.edu
Home page: http://ceoas.oregonstate.edu/profile/moghimi/ (or My personal
homepage https://sites.google.com/site/moghimis/)
Telephone: +1 (541) 737-3402.

---

[tacaswell]

@efiring Thoughts on this?

[efiring]

It works on 1.4.0, but not on 1.4.2. Strange.

[efiring]

Bisecting shows 350413c as the first bad commit. @mdboom would you look at this, please? It's deep in agg.

[efiring]

I don't understand why, but this seems to be fixed by #4025. It must be the elimination of the conditional in the draw method. How that would be related to my bisecting result, I don't know.

[rsignell-usgs]

@tacaswell, this bug was killing me all morning -- my student's notebook was working for him, but not for me. We (with help from @ocefpaf) eventually tracked this down to this bug, due to my newer matplotlib (1.4.2) vs his older one (1.4.0). Any chance of releasing 1.4.3 with this bugfix?

[tacaswell]

@rsignell-usgs Target date for 1.4.3 is Feb 1, but we still have a few out-standing blockers (mostly #3095 ) so with in the next few weeks.

[rsignell-usgs]

@tacaswell. Okay, thanks for this information and thank you also for all the great work you do for the community!