hist with stackedboxed anomaly

Hello,

Is there some reason the first and second figure in the simple program below should show any difference other than color? (On my machine, the first plot is missing the 3 and 4 datapoints.)

import pylab as p
x1=[1, 2, 3, 4, 2, 2, 2, 3, 3, 3]
x2=[1, 2]
p.figure()
p.hist([x2, x1], histtype='barstacked')
p.figure()
p.hist([x1, x2], histtype='barstacked')
p.show()

It seems that the xrange is computed from the first data in the sequence and then anything in the second sequence that is outside that range is simply tossed. If I force the bins to encompass the entire range with bins= in the hist() call, then all data is displayed in both plots. Is this desired/expected?

I am running matplotlib 0.98.5.2-1ubuntu3.

Thanks for any help, and thanks for matplotlib!

Rich