basemap plot the opposite longitudinal section

Nobody has claimed this yet.

I'm getting nothing from stackoverflow so I post it here.

The problem is I can't let `bluemarble` to plot a map showing the Indo-Pacific longitudinal section, covering the longitude band of about [60,190]. Below is my code:

```
import matplotlib.pyplot as plt
from mpl_toolkits.basemap import Basemap

lats=[-25,66]
lons=[60,187]

fig=plt.figure();ax=fig.add_subplot(111)

bmap=Basemap(projection='merc',\
llcrnrlat=lats[0],llcrnrlon=lons[0],\
urcrnrlat=lats[1],urcrnrlon=lons[1]-360.,\
lon_0=(lons[0]+lons[1])/2.,\
ax=ax)

bmap.bluemarble()
plt.show()
```

The result has its x-axis flipped, I guess somehow related to this [issue#192](https://github.com/matplotlib/basemap/issues/192)

If I swap the llcrnrlon and urcrnrlon:
```
bmap=Basemap(projection='merc',\
llcrnrlat=lats[0],llcrnrlon=lons[1]-360.,\
urcrnrlat=lats[1],urcrnrlon=lons[0],\
lon_0=(lons[0]+lons[1])/2.,\
ax=ax)
```

the map shows correct orientation, but the wrong domain. I wanted to show the Indian ocean, Euro-asia and west Pacific, however the map shows the remaining longitudes, covering the central-east Pacific, America continents and Atlantic. The lon_0 argument doesn't seem to do anything.

I am aware that when crossing the dateline things get complicated and sometimes you need to `shiftgrid()` or `shiftdata()`, but in this case I have no data yet, just the bluemarble background.

Please help.

Some versions (all in conda with python2.7):
basemap=1.0.7=np111py27_0
matplotlib=2.0.0=np111py27_0

Reproduce the reported code with Basemap 1.0.7 and matplotlib 2.0.0, using the longitude bounds [60,190] and the bluemarble background. Start by checking how the Mercator bounds and lon_0 are handled when crossing the dateline; done means the Indo-Pacific section displays in the requested orientation and domain.

Written by the indexing model from the issue text.