Fixing pickledb bugs

[StaticallyTypedAnxiety]

Please check one of the following, relating to the CHANGELOG, which should be updated if relevant

• my changes to the code affect some exposed aspect of the developer experience, and I have added an item to relevant 'Added, Fixed, Changed, Removed, Deprecated, or Security' heading under the 'Unreleased' heading of the CHANGELOG, with the format - summary of change [PR#1234](https://github.com/holochain/holochain-rust/pull/1234)
• my changes to the code do not affect any exposed aspect of the developer experience

[StaticallyTypedAnxiety]

Dont merge until the profiler is run(can still review) just need to confirm that the space issue has been fixed.

Edit : This has been confirmed, investigating why pickledb is 3 times larger than file.

Findings : File is larger because a further encoding is done on the content which is an encoded type in itself. This sounds like something that should be talked about architecturally

[lucksus]

I was looking up tempdir's docs because I was wondering if that is a unique directory since we don't want multiple test instances share the same CAS. It says that the tempdir gets removed when TempDir is dropped. This happens a few lines below, right? The conductor will recreate the the path I guess but then nothing will delete it, so it actually is not temporary. This might cause mysterious problems later on.. Not sure if we actually want to merge pickle as default for nodejs conductor into develop.. We might want to if it really is faster than MemoryStorage..

[StaticallyTypedAnxiety]

I agree, I think since memory is the fastest and this is for the use in the nodejs we should still go back to memory. I will revert this back asap

[StaticallyTypedAnxiety]

Benchmarks have been run and flame graph is up on realtimeboard

test bench_file_eav_many_to_one ... bench: 183,641,550 ns/iter (+/- 131,852,367)
test bench_file_eav_one_to_many ... bench: 201,104,300 ns/iter (+/- 119,277,360)
test bench_memory_eav_many_to_one ... bench: 791,201 ns/iter (+/- 77,859)
test bench_memory_eav_one_to_many ... bench: 827,959 ns/iter (+/- 95,228)
test bench_pickle_eav_many_to_one ... bench: 14,167,780 ns/iter (+/- 14,597,584)
test bench_pickle_eav_one_to_many ... bench: 13,985,950 ns/iter (+/- 10,538,156)

[lucksus]

Cool! Hope this gets merged asap! :)
Maybe we should change the title of this PR though since it's now only fixing pickledb without setting it in the nodejs conductor.