Custom Atom type to replace unsound API

[Rigidity]

This PR changes the way a.atom(ptr) works to return an Atom type that can convert small integers to byte slices. This prevents a memory safety issue on the edge case where you exceed the temp vec length and references are left dangling.

Overall, the performance is actually improved slightly for the majority of benchmarks.

group                                                          after                                before
-----                                                          -------                                ------
deserialize/node_from_bytes                                    1.03    119.7±2.17µs        ? ?/sec    1.00    116.6±0.82µs        ? ?/sec
deserialize/node_from_bytes_backrefs                           1.00    141.8±4.51µs        ? ?/sec    1.19    168.8±1.96µs        ? ?/sec
deserialize/node_from_bytes_backrefs-compressed                1.00    144.6±4.76µs        ? ?/sec    1.00    144.5±7.68µs        ? ?/sec
deserialize/serialized_length_from_bytes                       1.03     98.8±2.57µs        ? ?/sec    1.00     95.9±1.70µs        ? ?/sec
deserialize/serialized_length_from_bytes-compressed            1.00    112.2±2.75µs        ? ?/sec    1.00    112.3±0.85µs        ? ?/sec
deserialize/serialized_length_from_bytes_trusted               1.03     27.2±0.99µs        ? ?/sec    1.00     26.5±0.05µs        ? ?/sec
deserialize/serialized_length_from_bytes_trusted-compressed    1.14     25.4±5.93µs        ? ?/sec    1.00     22.3±0.51µs        ? ?/sec
deserialize/tree_hash_from_stream                              1.02      2.0±0.02ms        ? ?/sec    1.00   1975.8±4.71µs        ? ?/sec
run_program/block-2000                                         1.01     99.1±1.26ns        ? ?/sec    1.00     98.2±0.39ns        ? ?/sec
run_program/compressed-2000                                    1.00    353.8±4.83ms        ? ?/sec    1.08    381.5±2.64ms        ? ?/sec
run_program/concat                                             1.00    262.1±2.85ms        ? ?/sec    1.00    261.1±1.98ms        ? ?/sec
run_program/count-even                                         1.00      5.6±0.04ms        ? ?/sec    1.05      5.9±0.01ms        ? ?/sec
run_program/factorial                                          1.00    106.5±0.05ms        ? ?/sec    1.01    107.2±0.31ms        ? ?/sec
run_program/hash-string                                        1.00    215.9±0.21ns        ? ?/sec    1.00    215.6±0.88ns        ? ?/sec
run_program/hash-tree                                          1.00     16.8±0.06ms        ? ?/sec    1.05     17.7±0.87ms        ? ?/sec
run_program/large-block                                        1.00     87.1±0.41ms        ? ?/sec    1.00     86.8±0.03ms        ? ?/sec
run_program/loop_add                                           1.00  1311.2±37.06ms        ? ?/sec    1.05   1380.9±3.58ms        ? ?/sec
run_program/loop_ior                                           1.00  1254.7±10.61ms        ? ?/sec    1.18   1482.9±3.57ms        ? ?/sec
run_program/loop_not                                           1.00  1283.7±10.03ms        ? ?/sec    1.07  1372.0±11.66ms        ? ?/sec
run_program/loop_sub                                           1.00  1403.8±29.64ms        ? ?/sec    1.03   1450.4±6.99ms        ? ?/sec
run_program/loop_xor                                           1.00  1372.0±18.20ms        ? ?/sec    1.06  1459.6±15.68ms        ? ?/sec
run_program/matrix-multiply                                    1.00     96.2±3.67ms        ? ?/sec    1.00     96.6±0.32ms        ? ?/sec
run_program/point-pow                                          1.00    135.9±0.56ms        ? ?/sec    1.00    136.3±1.38ms        ? ?/sec
run_program/pubkey-tree                                        1.00     73.5±0.47ms        ? ?/sec    1.01     74.1±0.88ms        ? ?/sec
run_program/shift-left                                         1.02   1067.5±6.91ms        ? ?/sec    1.00   1051.2±4.25ms        ? ?/sec
run_program/substr                                             1.00      2.9±0.03ms        ? ?/sec    1.09      3.2±0.02ms        ? ?/sec
run_program/substr-tree                                        1.00     64.3±1.46ms        ? ?/sec    1.05     67.4±0.42ms        ? ?/sec
run_program/sum-tree                                           1.00    142.7±1.79ms        ? ?/sec    1.05    150.1±1.01ms        ? ?/sec

[coveralls-official]

Pull Request Test Coverage Report for Build 7847070939

Warning: This coverage report may be inaccurate.

We've detected an issue with your CI configuration that might affect the accuracy of this pull request's coverage report.
To ensure accuracy in future PRs, please see these guidelines.
A quick fix for this PR: rebase it; your next report should be accurate.

• -11 of 222 (95.05%) changed or added relevant lines in 13 files are covered.
• 2 unchanged lines in 1 file lost coverage.
• Overall coverage decreased (-0.1%) to 94.361%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
wasm/src/lazy_node.rs	0	1	0.0%
src/runtime_dialect.rs	0	3	0.0%
src/allocator.rs	135	142	95.07%

Files with Coverage Reduction	New Missed Lines	%
src/allocator.rs	2	98.37%

Totals
Change from base Build 7837886158:	-0.1%
Covered Lines:	5756
Relevant Lines:	6100

---

💛 - Coveralls

[arvidn]

looks good!

[arvidn]

I have one comment that I think might be important to address