fix ghost heap accounting

[arvidn]

We only have a heap limit in mempool mode, so this is not consensus critical.

Currently, we subtract 1 from the heap limit to account for the 1 byte we used to pre-allocate in the allocator (for Allocator::one()). Instead, just add that as a "ghost heap" counter.

Also, make sure every way we create a new atom honors the heap limit.

And give heap_size() the same treatment as atom_count() and pair_count() already received, and separate the allocated bytes from the logical heap size.

Note

^{Cursor Bugbot is generating a summary for commit 84715c6. Configure here.}

[coveralls-official]

Pull Request Test Coverage Report for Build 22622192615

Details

• 61 of 62 (98.39%) changed or added relevant lines in 2 files are covered.
• 2 unchanged lines in 1 file lost coverage.
• Overall coverage increased (+0.05%) to 88.176%

---

Changes Missing Coverage	Covered Lines	Changed/Added Lines	%
src/allocator.rs	60	61	98.36%

Files with Coverage Reduction	New Missed Lines	%
src/allocator.rs	2	96.93%

Totals
Change from base Build 22596628669:	0.05%
Covered Lines:	6883
Relevant Lines:	7806

---

💛 - Coveralls

[cursor]

Cursor Bugbot has reviewed your changes and found 1 potential issue.

^{Bugbot Autofix is OFF. To automatically fix reported issues with cloud agents, enable autofix in the Cursor dashboard.}