You say
"there’s a clear difference between the uniformly spherical bubbles in glasses and the shape of the voids in quartz". I have to disagree that this is a clear difference that can be in any way diagnostic. I have seen them uniformally spherical in some Victorian quartz from gold deposits (more commonly they are irregular and often elongated). Likewise, while they are never crystal shapes in glass, they are often elongated and irregular like this example in glass.
View attachment 1366
There are all sorts of reasons why bubble inclusions have a particular shape - for example some were simple fluid bubbles when the quartz formed, others subsequently modified their shape as the quartz cooled, others formed along fracture planes in the cooling quartz, others disaggregated after intitial formation of one big bubble, forming many tiny bubbles. And a person looking at an inclusion in a hand specimen will commonly see the almost-always present spherical shape of the vapour bubble in the liquid, rather than the outline of the liquid.
View attachment 1367
And of course quartz is much harder than glass, so the simple scratch test is far more diagnostic of whether a specimen is quartz or glass.
And as you say, negative crystal shapes when present can be quite spectacular!
View attachment 1365
I first saw the enhydros when a 14 year old prospector at Beechworth, And yes, they were not in crystals but had taken the shape of cavities - I think they were actually chalcedonic in the main.