The TreeNodeIterator decodes nodes while iterating over a tree blob.
This should reduce peak memory usage as now only the serialized tree
blob and a single node have to alive at the same time. Using the
iterator has implications for the error handling however. Now it is
necessary that all loops that iterate through a tree check for errors
before using the node returned by the iterator.
The other change is that it is no longer possible to iterate over a tree
multiple times. Instead it must be loaded a second time. This only
affects the tree rewriting code.
This changes Dumper.writeNode to spawn loader goroutines as needed
instead of as a pool. The code is shorter, fewer goroutines are spawned
for small files, and crash dumps (also for unrelated errors) should be
smaller.
It was only used in two places:
- stats: apparently as a minor performance optimization, which is
unlikely to be important
- find: filtered directories would be ignored. However, this
optimization missed that it is possible that two directories have the
exact same content. Such directories would be incorrectly ignored too.
Example:
```
mkdir test test/a test/b
restic backup test
restic find latest test/b
-> incorrectly does not return anything
```
Thus, remove the functionality as it's apparently too complex to use
correctly.
Package internal/dump has been reworked so its API consists of a single
type Dumper that handles tar and zip formats. Tree loading and node
writing happen concurrently.
