Well, in the normal case, you wouldn't do it. These variable length structures need to be created on the heap to be able to be used in a variable length way, for the most part, so you'd just put a pointer to them into an array.
However, I can think of a couple of methods, such as packing into an array, and using a second one to index it, like so:
a = [aaaa,bb,ccc,dd]
idx = [0,4,6,9]
To get to the i'th element of a, accesses would go through idx like so: a[idx[i]]. In general, of course, there's no way to allow O(1) access and updates without occasional repacking.
This is exactly the problem I'm getting at, you're not actually working within the confines of C here, you're creating funny workarounds for the fact that certain C features don't work how you want them to, and you're violating the type system in the process.
The concept of VLAs does not fit the language well, they're inherently something of an anomaly. Allowing them in structs would simply multiply the anomalies.
I'm frankly shocked that they were codified in C99 at all, rather than codifying something akin to alloca() with implementation-defined behavior, but I'm infinitely grateful the committee did not elect to make them anything more than they are -- which is a semi-portable mechanism for allocating arbitrary amounts of automatic memory.
However, I can think of a couple of methods, such as packing into an array, and using a second one to index it, like so:
To get to the i'th element of a, accesses would go through idx like so: a[idx[i]]. In general, of course, there's no way to allow O(1) access and updates without occasional repacking.