If there were a crack in the glass, then the pressure difference would first cause air to flow into the lamp, which would cause it to stop working (both the mercury and the low pressure are necessary).
If there were a way for mercury to escape from the lamp while it still worked, it would most likely be at the glass-to-metal seal around the filament/electrode terminals. But if that were a problem then lots of bulbs would be leaky and it would be fixed by design.
It is possible for a fluorescent lamp to break (and therefore eventually lose its mercury) without “exploding”. I once witnessed just such an event, which might be of interest:
The lamp was a spiral-style CFL lamp pointing sideways (in an above-the-mirror bathroom fixture).
I don't remember at what point it stopped lighting or if the failure happened when it was switched on. But in any case, the filament at one end glowed brightly and broke. The hot filament fell down against the glass, which then cracked at that point, with a distinct sound, making a small hole. The rest of the glass remained intact.
(The overheated filament would have been caused by a failure in the starting circuit.)