Remove initialize and make finalize optional

[nhz2]

Fixes#70 by initializing on construction and finalizing in a finalizer.

[codecov]

Codecov Report

❌ Patch coverage is 95.83333% with 1 line in your changes missing coverage. Please review.
✅ Project coverage is 60.64%. Comparing base (e7edfed) to head (fdb6ec9).
⚠️ Report is 4 commits behind head on master.

Files with missing lines	Patch %	Lines
src/compression.jl	91.66%	1 Missing ⚠️

Additional details and impacted files

@@ Coverage Diff @@## master #71 +/- ##
==========================================
- Coverage 61.55% 60.64% -0.92% 
==========================================
Files 5 5 Lines 372 371 -1 ==========================================
- Hits 229 225 -4 - Misses 143 146 +3 

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[nhz2]

ZSTD_createCStream_advanced is part of the experimental API though according to facebook/zstd#1734 these functions should not change much in the future.

I am using it because the regular ZSTD_createCStream leads to Julia not running the finalizer before being killed by OOM.

[mkitti]

We should not rely on the GC to finalize external resources.

[nhz2]

We should not rely on the GC to finalize external resources.

Should I try and support both explicit-free when we know the Codec cannot escape and automatic-free for more complex situations?

[nhz2]

atomics require 1.7

[nhz2]

I'm going to close this PR because I don't like that it uses an experimental API that may break in the future. With Zstd_jll being a standard library, this means a future version of Julia could be incompatible with this PR.

Also, with some options, the zstd compression and decompression context can use multiple GB's of memory, so users may need more control over when that gets allocated and freed.