Use GC.preserve around string parsing

[quinnj]

Full implementation of the idea/discussion in #391 . This ensures our buf is GC-preserved before every call to parsestring and the preservation lasts until the PtrString goes out of scope.

Many thanks to @yuyichao for helping around this issue.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 89.99%. Comparing base (cba1b4a) to head (501cc88).
⚠️ Report is 4 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #395      +/-   ##
==========================================
+ Coverage   89.89%   89.99%   +0.09%     
==========================================
  Files           7        7              
  Lines        1316     1329      +13     
==========================================
+ Hits         1183     1196      +13     
  Misses        133      133              

☔ 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.

[yuyichao]

Since buf isn't explicitly used and getbuf is just an accessor, GC.@preserve x is probably more clear in this case. The main difference is that the other fields of x would also be preserved.

[quinnj]

Ah, so GC.@preserve x will ensure the field is also preserved?

[yuyichao]

It will assume the whole object is being used in some unknown way, which includes the field.

For mutable object, this implies the object address may be used (though there’s no requirement for that address to be allocated in the heap). For immutable object this only implies that the fields are all being used in some unknown way.