When using `NoGC` if the machine runs out of memory, the user will see
an error that says garbage collection failed:

```
ERROR: An MMTk GC thread panicked.  This is a bug.
panicked at /Users/eileencodes/src/github.com/open_source/mmtk-core/src/plan/nogc/global.rs:74:9:
internal error: entered unreachable code: GC triggered in nogc
Backtrace is disabled.
run with `RUST_BACKTRACE=1` environment variable to display a backtrace
running file: test/zlib/test_zlib.rb
```

I added a check in `poll` to see if the plan being used supports GC and
if not, raise an error. I could have updated the `schedule_collection`
function to return this error, but I felt that since that function is
regarding scheduling, it was better to raise if any plan does not
support GC than change the purpose of `schedule_collection`. We still
want to know if a GC ended up being triggered in `schedule_collection`.

The new error looks like this:

```
[2024-07-24T18:43:45Z INFO  mmtk::memory_manager] Initialized MMTk with NoGC (FixedHeapSize(1073741824))
thread '<unnamed>' panicked at /Users/eileencodes/src/github.com/open_source/mmtk-core/src/util/heap/gc_trigger.rs:80:17:
internal error: entered unreachable code: User ran out of space and the plan does not support collecting garbage.
note: run with `RUST_BACKTRACE=1` environment variable to display a backtrace
fatal runtime error: failed to initiate panic, error 5
running file: test/zlib/test_zlib.rb
```