fix(merkletree): reject non-positive segment sizes in ReadAll

[gophersg]

Description

merkletree.Tree.ReadAll accepts a public segmentSize int with no validation.
When segmentSize == 0, io.ReadFull(r, make([]byte, 0)) returns (0, nil), so the
for loop never hits the io.EOF break and keeps pushing empty leaves forever.
When segmentSize < 0, make([]byte, segmentSize) panics with
makeslice: len out of range.

Both ReaderRoot and BuildReaderProof call ReadAll, so a non-positive segment
size (e.g. a computed/zero length from a caller) can hang the process or panic.
This is the same input-validation hardening class as #893 and #878.

This PR adds an early segmentSize <= 0 check at the top of ReadAll and returns a
clear error, so all callers fail fast.

Type of change

• Bug fix (non-breaking change which fixes an issue)
• New feature
• Breaking change
• This change requires a documentation update

How has this been tested?

• go test ./accumulator/merkletree/ -run TestReadAll -v
  - TestReadAllRejectsNonPositiveSegmentSize: segmentSize 0 and -1 rejected by
  ReadAll, ReaderRoot, and BuildReaderProof
  - TestReadAllValidSegmentSize: positive segmentSize still builds the expected root
• gofmt -l clean and go vet ./accumulator/merkletree/ clean

How has this been benchmarked?

• N/A — pure early-return on invalid input; no hot-path change.

Checklist:

• I have performed a self-review of my code
• I have commented my code, particularly in hard-to-understand areas
• I have made corresponding changes to the documentation (N/A for this fix)
• I have added tests that prove my fix is effective or that my feature works
• golangci-lint does not output errors locally
• New and existing unit tests pass locally with my changes
• Any dependent changes have been merged and published in downstream modules

---

Note

Low Risk
Input validation only on an invalid edge case; valid positive segment sizes are unchanged per regression test.

Overview
Tree.ReadAll now rejects segmentSize <= 0 up front with merkletree: segment size must be positive, instead of infinite looping when size is 0 or panicking on negative make([]byte, segmentSize).

Because ReaderRoot and BuildReaderProof both call ReadAll, invalid segment sizes from callers now fail fast with that error.

New tests cover rejection for 0 and -1 on ReadAll, ReaderRoot, and BuildReaderProof, plus a regression check that a positive segment size still yields the same root as before.

^{Reviewed by Cursor Bugbot for commit 49b6086. Bugbot is set up for automated code reviews on this repo. Configure here.}