Walking around in Leuven, I always seem to be drawn to the train station and it's square in front.
The light was difficult, not much of it, and very uneven, like here, where the cafe inside is reasonably well lit, but the outside is virtually black.
Difficult light, not easy to see things anyway, so I used the 'easy' camera, the X100s, with it's 35-equivalent lens, which I know by hart. The 14mm on the X-Pro 1 might have given some great results, but that is not an easy lens to use, the (very) wide angle of view tends to distort the subject easily, you need to concentrate, and I had enough to think of with this very difficult light. The 35mm standard is too narrow for me for these kind of pictures, my new 18mm would probably have been ok, but I chose the safer option, so the Pro stayed in the bag. At night, when it is not that easy to predict a situation or to observe people, I prefer to have only one camera visible if possible.
I find that using auto-exposure in such conditions is not reliable, because the light meter wants to find a balance between the dark parts and the light ones, even using an 'intelligent' multi-whatever mode.
In these conditions I usually guesstimate my exposure, based on the part of the photo that is the most important to me, in this case the light inside.
When I find myself in such conditions, I set an exposure by hand and leave it as-is for most of my photographs, unless I really find that conditions have changed considerably. Taking a spot-meter reading on the most important part of the picture would do the trick, but that takes to much time for me, at least for these kind of pictures.
Your histogram or viewing screen are usually not that much help, you will really need some experience to judge the potential results. An EV is a great help by the way, using a reflex viewer like you find on a (D)SLR or even a Leica rangefinder can be very difficult if it is this dark.
That does mean that the files will need some post-processing, but given that the light quality might be a little difficult anyway, with different light sources to which the sensor will react very different, some excessive post is unavoidable for me anyway.
The initial RAW file was developed in Capture One Pro, and I added a mask on the building (without the windows) where I reduced the color saturation, to enhance the contrast between the warm and cosy inside and the cold hostile outside that I was after.
I finished off in Photoshop with more masking (a lot actually) and several adjustment layers. I normally don't do that kind of extensive post-production manipulation, but in this case it that was the only way to get what I was after.
For comparison, this is the original jpeg (which would never have given the result above).
This is more-or-less also how the result looked while viewing in-camera.