One thing I've noticed during my recent PRs is that the code base that does streaming xlsx reading and writing is quite different from the 'sync' reading and writing of xlsx.js. But the streaming implementation seems to have less functionality implemented (see #1143)

It seems to me that it would be easier to maintain and to write tests if we have only one implementation. The only real difference I see is that the two approaches use different zip libraries (because the sync xlsx.js implementation needs to work in the browser). Assuming we can write an interface that accepts both zip librariers, can't we make the stream implementation the master implementation and the sync version is just a simple Promise wrapper around that?

@guyonroche I'm sure you know the history better, maybe you could shed some light on the two implementations and what your plans are/were for them? Thanks!