Merge remote-tracking branch 'XRPLF/ximinez/number-division-accuracy' into ximinez/number-maxint-range

* XRPLF/ximinez/number-division-accuracy:
  CI feedback: Add test cases covering other rounding modes
  Apply suggestions from code review
  Accept AI suggestion
  Update the testUpwardRoundsDown test to run under all scales
  Review feedback from Tapanito and AI
  Skip clang-tidy false positive: misc-include-cleaner
  ci: Run PR title and description checks on staging and release branches (7331)
  style: Run shfmt on workflows, actions and markdown bash code (7333)
  Remove TMax entirely from normalizeToRange; check type matching directly
  Tweak how the denominator is handled in division
  Minor fixes: missing include, variable init, typo
  Test optimization: Number_test::pow10
  Significant rewrite
  Use the local range instead of calling a function
  Make Number::operator/= significantly more accurate

.github/actions/build-deps/action.yml

@@ -37,12 +37,12 @@ runs:
       run: |
         echo 'Installing dependencies.'
         conan install \
-          --profile ci \
-          --build="${BUILD_OPTION}" \
-          --options:host='&:tests=True' \
-          --options:host='&:xrpld=True' \
-          --settings:all build_type="${BUILD_TYPE}" \
-          --conf:all tools.build:jobs=${BUILD_NPROC} \
-          --conf:all tools.build:verbosity="${LOG_VERBOSITY}" \
-          --conf:all tools.compilation:verbosity="${LOG_VERBOSITY}" \
-          .
+            --profile ci \
+            --build="${BUILD_OPTION}" \
+            --options:host='&:tests=True' \
+            --options:host='&:xrpld=True' \
+            --settings:all build_type="${BUILD_TYPE}" \
+            --conf:all tools.build:jobs=${BUILD_NPROC} \
+            --conf:all tools.build:verbosity="${LOG_VERBOSITY}" \
+            --conf:all tools.compilation:verbosity="${LOG_VERBOSITY}" \
+            .

.github/actions/generate-version/action.yml

@@ -15,7 +15,7 @@ runs:
       shell: bash
       env:
         VERSION: ${{ github.ref_name }}
-      run: echo "VERSION=${VERSION}" >> "${GITHUB_ENV}"
+      run: echo "VERSION=${VERSION}" >>"${GITHUB_ENV}"
 
     # When a tag is not pushed, then the version (e.g. 1.2.3-b0) is extracted
     # from the BuildInfo.cpp file and the shortened commit hash appended to it.
@@ -28,17 +28,17 @@ runs:
         echo 'Extracting version from BuildInfo.cpp.'
         VERSION="$(cat src/libxrpl/protocol/BuildInfo.cpp | grep "versionString =" | awk -F '"' '{print $2}')"
         if [[ -z "${VERSION}" ]]; then
-          echo 'Unable to extract version from BuildInfo.cpp.'
-          exit 1
+            echo 'Unable to extract version from BuildInfo.cpp.'
+            exit 1
         fi
 
         echo 'Appending shortened commit hash to version.'
         SHA='${{ github.sha }}'
         VERSION="${VERSION}+${SHA:0:7}"
 
-        echo "VERSION=${VERSION}" >> "${GITHUB_ENV}"
+        echo "VERSION=${VERSION}" >>"${GITHUB_ENV}"
 
     - name: Output version
       id: version
       shell: bash
-      run: echo "version=${VERSION}" >> "${GITHUB_OUTPUT}"
+      run: echo "version=${VERSION}" >>"${GITHUB_OUTPUT}"

.github/scripts/format-inline-bash.py

@@ -0,0 +1,403 @@
+#!/usr/bin/env python3
+
+"""
+Format embedded shell snippets using the shfmt hook configured in
+.pre-commit-config.yaml.
+
+Two shapes are recognised:
+
+* YAML workflow/action files: literal block-scalar runs (`run: |`) and
+  single-line runs (`run: some command`). A single-line run is upgraded to
+  a `run: |` block scalar if shfmt's output spans multiple lines.
+
+* Markdown files: ``` ```bash ``` fenced code blocks.
+
+Any block that shfmt cannot parse is skipped with a warning on stderr, so
+the file is left untouched and surrounding blocks still get formatted.
+
+For each occurrence the body is dedented, written to a temp .sh file,
+formatted via `pre-commit run shfmt --files <temp>` (falling back to
+`prek`), then re-indented and written back in place.
+
+When invoked without arguments, every .yml/.yaml under .github/ plus every
+.md file in the repo is scanned. When invoked with file arguments (the
+pre-commit case), only those files are processed.
+"""
+
+from __future__ import annotations
+
+import re
+import shutil
+import subprocess
+import sys
+import tempfile
+from dataclasses import dataclass
+from pathlib import Path
+from typing import Union
+
+REPO = Path(__file__).resolve().parents[2]
+
+_HOOK_RUNNER = next((cmd for cmd in ("pre-commit", "prek") if shutil.which(cmd)), None)
+if _HOOK_RUNNER is None:
+    sys.exit("error: neither `pre-commit` nor `prek` found on PATH")
+
+RUN_BLOCK_RE = re.compile(r"^(?P<prefix>[ \t]*(?:- )?)run:[ \t]*\|[+-]?[ \t]*$")
+RUN_INLINE_RE = re.compile(
+    r"^(?P<prefix>[ \t]*(?:- )?)run:[ \t]+" r"(?P<value>(?!\|[+-]?[ \t]*$)\S.*?)[ \t]*$"
+)
+MD_BASH_OPEN_RE = re.compile(r"^(?P<indent>[ ]{0,3})`{3}bash[ \t]*$")
+MD_FENCE_CLOSE_RE = re.compile(r"^[ ]{0,3}`{3,}[ \t]*$")
+
+
+@dataclass(frozen=True)
+class BlockRun:
+    """A `run: |` block scalar; `body_start:body_end` slices into `lines`."""
+
+    body_start: int
+    body_end: int
+    body_indent: int
+
+
+@dataclass(frozen=True)
+class InlineRun:
+    """A single-line `run: value` at `line_idx`."""
+
+    line_idx: int
+    prefix: str
+    value: str
+
+
+@dataclass(frozen=True)
+class MdBashBlock:
+    """A markdown ``` ```bash ``` fenced code block.
+
+    `body_start:body_end` slices into the file's lines; `open_line_idx`
+    points at the opening fence line.
+    """
+
+    open_line_idx: int
+    body_start: int
+    body_end: int
+    body_indent: int
+
+
+RunItem = Union[BlockRun, InlineRun]
+
+
+def _scan_block_body(
+    lines: list[str], body_start: int, run_col: int
+) -> tuple[int | None, int]:
+    """Locate the body of a `run: |` block scalar starting at `body_start`.
+
+    Returns `(body_indent, scan_end)`. `scan_end` is the line index where the
+    outer scanner should resume. `body_indent` is `None` when no body is
+    present (the scalar is empty, or the next non-blank line has indent
+    `<= run_col`).
+    """
+    body_indent: int | None = None
+    scan_end = len(lines)
+    for idx in range(body_start, len(lines)):
+        line = lines[idx]
+        if line.strip() == "":
+            continue
+        indent = len(line) - len(line.lstrip(" "))
+        if body_indent is None:
+            if indent > run_col:
+                body_indent = indent
+            else:
+                scan_end = idx
+                break
+        elif indent < body_indent:
+            scan_end = idx
+            break
+    if body_indent is not None:
+        while scan_end > body_start and lines[scan_end - 1].strip() == "":
+            scan_end -= 1
+        if scan_end <= body_start:
+            body_indent = None
+    return body_indent, scan_end
+
+
+def find_run_blocks(lines: list[str]) -> list[RunItem]:
+    """Return run items in document order."""
+    items: list[RunItem] = []
+    line_idx = 0
+    while line_idx < len(lines):
+        line = lines[line_idx]
+        if block_match := RUN_BLOCK_RE.match(line):
+            run_col = len(block_match.group("prefix"))
+            body_start = line_idx + 1
+            body_indent, scan_end = _scan_block_body(lines, body_start, run_col)
+            if body_indent is not None:
+                items.append(
+                    BlockRun(
+                        body_start=body_start,
+                        body_end=scan_end,
+                        body_indent=body_indent,
+                    )
+                )
+            line_idx = scan_end
+            continue
+        if inline_match := RUN_INLINE_RE.match(line):
+            items.append(
+                InlineRun(
+                    line_idx=line_idx,
+                    prefix=inline_match.group("prefix"),
+                    value=inline_match.group("value"),
+                )
+            )
+        line_idx += 1
+    return items
+
+
+def find_md_bash_blocks(lines: list[str]) -> list[MdBashBlock]:
+    """Return ``` ```bash ``` fenced code blocks in document order."""
+    blocks: list[MdBashBlock] = []
+    line_idx = 0
+    while line_idx < len(lines):
+        open_match = MD_BASH_OPEN_RE.match(lines[line_idx])
+        if not open_match:
+            line_idx += 1
+            continue
+        body_start = line_idx + 1
+        close_idx = next(
+            (
+                j
+                for j in range(body_start, len(lines))
+                if MD_FENCE_CLOSE_RE.match(lines[j])
+            ),
+            None,
+        )
+        if close_idx is None:
+            line_idx = body_start
+            continue
+        body = lines[body_start:close_idx]
+        non_blank = [b for b in body if b.strip()]
+        body_indent = (
+            min(len(b) - len(b.lstrip(" ")) for b in non_blank)
+            if non_blank
+            else len(open_match.group("indent"))
+        )
+        blocks.append(
+            MdBashBlock(
+                open_line_idx=line_idx,
+                body_start=body_start,
+                body_end=close_idx,
+                body_indent=body_indent,
+            )
+        )
+        line_idx = close_idx + 1
+    return blocks
+
+
+def dedent(lines: list[str], n: int) -> list[str]:
+    pad = " " * n
+    return [
+        (
+            ""
+            if line.strip() == ""
+            else (line[n:] if line.startswith(pad) else line.lstrip(" "))
+        )
+        for line in lines
+    ]
+
+
+def reindent(lines: list[str], n: int) -> list[str]:
+    pad = " " * n
+    return [pad + line if line else "" for line in lines]
+
+
+_SHFMT_ERR_RE = re.compile(r"\.sh:\d+:\d+:\s")
+_GHA_EXPR_RE = re.compile(r"\$\{\{.*?\}\}", re.DOTALL)
+_GHA_PLACEHOLDER_RE = re.compile(r"__GHA_EXPR_(\d+)__")
+
+
+def _encode_gha_exprs(text: str) -> tuple[str, list[str]]:
+    """Replace `${{ ... }}` expressions with bash-safe placeholder identifiers."""
+    exprs: list[str] = []
+
+    def repl(match: re.Match[str]) -> str:
+        exprs.append(match.group(0))
+        return f"__GHA_EXPR_{len(exprs) - 1}__"
+
+    return _GHA_EXPR_RE.sub(repl, text), exprs
+
+
+def _decode_gha_exprs(text: str, exprs: list[str]) -> str:
+    """Restore `${{ ... }}` expressions from placeholder identifiers."""
+    return _GHA_PLACEHOLDER_RE.sub(lambda m: exprs[int(m.group(1))], text)
+
+
+def shfmt_via_hook(tmp_path: Path) -> tuple[bool, str]:
+    # `${{ ... }}` is not valid shell, so swap it for a placeholder identifier
+    # that shfmt can parse, then restore it after formatting.
+    encoded, exprs = _encode_gha_exprs(tmp_path.read_text())
+    if exprs:
+        tmp_path.write_text(encoded)
+    res = subprocess.run(
+        [_HOOK_RUNNER, "run", "shfmt", "--files", str(tmp_path)],
+        cwd=REPO,
+        capture_output=True,
+        text=True,
+    )
+    output = res.stdout + res.stderr
+    # shfmt emits parse errors as "<path>:<line>:<col>: <message>".
+    parse_err = bool(_SHFMT_ERR_RE.search(output))
+    # A non-zero exit that is neither a parse error nor pre-commit's "I had
+    # to modify files" signal means the hook itself failed to run (missing
+    # binary, install failure, bad config, ...). Surface that loudly rather
+    # than silently treating it as a no-op.
+    if (
+        res.returncode != 0
+        and not parse_err
+        and "files were modified by this hook" not in output
+    ):
+        sys.exit(
+            f"error: `{_HOOK_RUNNER} run shfmt` failed with exit {res.returncode}:\n{output}"
+        )
+    if exprs and not parse_err:
+        tmp_path.write_text(_decode_gha_exprs(tmp_path.read_text(), exprs))
+    return not parse_err, output
+
+
+def _skip(path: Path, where: int, kind: str, output: str) -> None:
+    print(
+        f"  shfmt could not parse {kind} at {path}:{where + 1} — skipped",
+        file=sys.stderr,
+    )
+    print(f"    {output.strip()}", file=sys.stderr)
+
+
+def process_yaml_file(path: Path, tmp_path: Path) -> int:
+    text = path.read_text()
+    had_nl = text.endswith("\n")
+    lines = text.split("\n")
+    if had_nl:
+        lines = lines[:-1]
+    items = find_run_blocks(lines)
+    if not items:
+        return 0
+    changed = 0
+    # Process in reverse so earlier indices remain valid as we splice.
+    for item in reversed(items):
+        if isinstance(item, BlockRun):
+            body = lines[item.body_start : item.body_end]
+            tmp_path.write_text("\n".join(dedent(body, item.body_indent)) + "\n")
+            ok, output = shfmt_via_hook(tmp_path)
+            if not ok:
+                _skip(path, item.body_start, "block", output)
+                continue
+            formatted = tmp_path.read_text().rstrip("\n")
+            new_body = reindent(formatted.split("\n"), item.body_indent)
+            if new_body != body:
+                lines[item.body_start : item.body_end] = new_body
+                changed += 1
+        else:
+            tmp_path.write_text(item.value + "\n")
+            ok, output = shfmt_via_hook(tmp_path)
+            if not ok:
+                _skip(path, item.line_idx, "inline run", output)
+                continue
+            formatted = tmp_path.read_text().rstrip("\n")
+            if formatted == item.value:
+                continue
+            formatted_lines = formatted.split("\n")
+            if len(formatted_lines) == 1:
+                lines[item.line_idx] = f"{item.prefix}run: {formatted}"
+            else:
+                body_indent = len(item.prefix) + 2
+                lines[item.line_idx : item.line_idx + 1] = [
+                    f"{item.prefix}run: |",
+                    *reindent(formatted_lines, body_indent),
+                ]
+            changed += 1
+    new_text = "\n".join(lines) + ("\n" if had_nl else "")
+    if new_text != text:
+        path.write_text(new_text)
+    return changed
+
+
+def process_md_file(path: Path, tmp_path: Path) -> int:
+    text = path.read_text()
+    had_nl = text.endswith("\n")
+    lines = text.split("\n")
+    if had_nl:
+        lines = lines[:-1]
+    blocks = find_md_bash_blocks(lines)
+    if not blocks:
+        return 0
+    changed = 0
+    for block in reversed(blocks):
+        body = lines[block.body_start : block.body_end]
+        tmp_path.write_text("\n".join(dedent(body, block.body_indent)) + "\n")
+        ok, output = shfmt_via_hook(tmp_path)
+        if not ok:
+            _skip(path, block.open_line_idx, "```bash block", output)
+            continue
+        formatted = tmp_path.read_text().rstrip("\n")
+        formatted_lines = formatted.split("\n") if formatted else []
+        new_body = reindent(formatted_lines, block.body_indent)
+        if new_body != body:
+            lines[block.body_start : block.body_end] = new_body
+            changed += 1
+    new_text = "\n".join(lines) + ("\n" if had_nl else "")
+    if new_text != text:
+        path.write_text(new_text)
+    return changed
+
+
+def process_file(path: Path, tmp_path: Path) -> int:
+    if path.suffix in (".yml", ".yaml"):
+        return process_yaml_file(path, tmp_path)
+    if path.suffix == ".md":
+        return process_md_file(path, tmp_path)
+    return 0
+
+
+def gather_files(argv: list[str]) -> list[Path]:
+    """Return YAML workflow/action files and markdown files that we should
+    process — either the paths in `argv` or, when `argv` is empty, every
+    such file in the repo (skipping `external/`)."""
+    if argv:
+        candidates: list[Path] = [
+            (REPO / a).resolve() if not Path(a).is_absolute() else Path(a) for a in argv
+        ]
+    else:
+        gh = REPO / ".github"
+        candidates = [
+            *gh.rglob("*.yml"),
+            *gh.rglob("*.yaml"),
+            *(
+                p
+                for p in REPO.rglob("*.md")
+                if "external" not in p.relative_to(REPO).parts
+            ),
+        ]
+    return sorted(
+        p
+        for p in candidates
+        if p.exists()
+        and (
+            (p.suffix in (".yml", ".yaml") and ".github" in p.parts)
+            or p.suffix == ".md"
+        )
+    )
+
+
+def main(argv: list[str]) -> int:
+    files = gather_files(argv)
+    if not files:
+        return 0
+    with tempfile.TemporaryDirectory(prefix="format-inline-bash-") as tmpdir:
+        tmp_path = Path(tmpdir) / "shfmt.sh"
+        total = 0
+        for f in files:
+            n = process_file(f, tmp_path)
+            if n:
+                print(f"{f.relative_to(REPO)}: reformatted {n} block(s)")
+                total += n
+        return 1 if total else 0
+
+
+if __name__ == "__main__":
+    sys.exit(main(sys.argv[1:]))

.github/workflows/build-nix-image.yml

@@ -100,8 +100,8 @@ jobs:
 
       - name: Create multi-arch manifests
         run: |
-          for tag in $(jq -cr '.tags[]' <<< "$DOCKER_METADATA_OUTPUT_JSON"); do
-            docker buildx imagetools create -t "$tag" "${tag}-amd64" "${tag}-arm64"
+          for tag in $(jq -cr '.tags[]' <<<"$DOCKER_METADATA_OUTPUT_JSON"); do
+              docker buildx imagetools create -t "$tag" "${tag}-amd64" "${tag}-arm64"
           done
 
       - name: Inspect image

.github/workflows/check-pr-description.yml

@@ -5,8 +5,17 @@ on:
     types:
       - checks_requested
   pull_request:
-    types: [opened, edited, reopened, synchronize, ready_for_review]
-    branches: [develop]
+    types:
+      - opened
+      - edited
+      - reopened
+      - synchronize
+      - ready_for_review
+    branches:
+      - develop
+      - "release-*"
+      - "release/*"
+      - "staging/*"
 
 jobs:
   check_description:
@@ -20,11 +29,11 @@ jobs:
         env:
           PR_BODY: ${{ github.event.pull_request.body }}
         if: ${{ github.event_name == 'pull_request' }}
-        run: printenv PR_BODY > pr_body.md
+        run: printenv PR_BODY >pr_body.md
 
       - name: Check PR description differs from template
         if: ${{ github.event_name == 'pull_request' }}
-        run: >
-          python .github/scripts/check-pr-description.py
-          --template-file .github/pull_request_template.md
-          --pr-body-file pr_body.md
+        run: |
+          python .github/scripts/check-pr-description.py \
+              --template-file .github/pull_request_template.md \
+              --pr-body-file pr_body.md

.github/workflows/check-pr-title.yml

@@ -5,10 +5,19 @@ on:
     types:
       - checks_requested
   pull_request:
-    types: [opened, edited, reopened, synchronize, ready_for_review]
-    branches: [develop]
+    types:
+      - opened
+      - edited
+      - reopened
+      - synchronize
+      - ready_for_review
+    branches:
+      - develop
+      - "release-*"
+      - "release/*"
+      - "staging/*"
 
 jobs:
   check_title:
     if: ${{ github.event.pull_request.draft != true }}
-    uses: XRPLF/actions/.github/workflows/check-pr-title.yml@291206777251b4d493641b5afbdf7c23009d2988
+    uses: XRPLF/actions/.github/workflows/check-pr-title.yml@cba1f0891650baf1a9c88624dc2d72573be2eb81

.github/workflows/on-pr.yml

@@ -98,7 +98,7 @@ jobs:
           READY: ${{ contains(github.event.pull_request.labels.*.name, 'Ready to merge') }}
           MERGE: ${{ github.event_name == 'merge_group' }}
         run: |
-          echo "go=${{ (env.DRAFT != 'true' && env.READY == 'true') || env.FILES == 'true' || env.MERGE == 'true' }}" >> "${GITHUB_OUTPUT}"
+          echo "go=${{ (env.DRAFT != 'true' && env.READY == 'true') || env.FILES == 'true' || env.MERGE == 'true' }}" >>"${GITHUB_OUTPUT}"
           cat "${GITHUB_OUTPUT}"
     outputs:
       go: ${{ steps.go.outputs.go == 'true' }}
@@ -168,9 +168,9 @@ jobs:
           PR_URL: ${{ github.event.pull_request.html_url }}
         run: |
           gh api --method POST -H "Accept: application/vnd.github+json" -H "X-GitHub-Api-Version: 2022-11-28" \
-          /repos/xrplf/clio/dispatches -f "event_type=check_libxrpl" \
-          -F "client_payload[ref]=${{ needs.upload-recipe.outputs.recipe_ref }}" \
-          -F "client_payload[pr_url]=${PR_URL}"
+              /repos/xrplf/clio/dispatches -f "event_type=check_libxrpl" \
+              -F "client_payload[ref]=${{ needs.upload-recipe.outputs.recipe_ref }}" \
+              -F "client_payload[pr_url]=${PR_URL}"
 
   passed:
     if: failure() || cancelled()

.github/workflows/pre-commit.yml

@@ -14,7 +14,7 @@ on:
 jobs:
   # Call the workflow in the XRPLF/actions repo that runs the pre-commit hooks.
   run-hooks:
-    uses: XRPLF/actions/.github/workflows/pre-commit.yml@5e942d61bf32f7557a7c159cfac4712a687b3e3a
+    uses: XRPLF/actions/.github/workflows/pre-commit.yml@cba1f0891650baf1a9c88624dc2d72573be2eb81
     with:
       runs_on: ubuntu-latest
       container: '{ "image": "ghcr.io/xrplf/ci/tools-rippled-pre-commit:sha-41ec7c1" }'

.github/workflows/reusable-build-docker-image.yml

@@ -53,7 +53,7 @@ jobs:
         env:
           PLATFORM: ${{ inputs.platform }}
         run: |
-          echo "arch=${PLATFORM##*/}" >> $GITHUB_OUTPUT
+          echo "arch=${PLATFORM##*/}" >>$GITHUB_OUTPUT
 
       - name: Set up Docker Buildx
         uses: docker/setup-buildx-action@d7f5e7f509e45cec5c76c4d5afdd7de93d0b3df5 # v4.1.0

.github/workflows/reusable-build-test-config.yml

@@ -113,7 +113,7 @@ jobs:
 
       - name: Set ccache log file
         if: ${{ inputs.ccache_enabled && runner.debug == '1' }}
-        run: echo "CCACHE_LOGFILE=${{ runner.temp }}/ccache.log" >> "${GITHUB_ENV}"
+        run: echo "CCACHE_LOGFILE=${{ runner.temp }}/ccache.log" >>"${GITHUB_ENV}"
 
       - name: Print build environment
         uses: XRPLF/actions/print-build-env@59dec886e4afb05a1724443af08baccbc045b574
@@ -146,11 +146,11 @@ jobs:
           CMAKE_ARGS: ${{ inputs.cmake_args }}
         run: |
           cmake \
-            -G '${{ runner.os == 'Windows' && 'Visual Studio 17 2022' || 'Ninja' }}' \
-            -DCMAKE_TOOLCHAIN_FILE:FILEPATH=build/generators/conan_toolchain.cmake \
-            -DCMAKE_BUILD_TYPE="${BUILD_TYPE}" \
-            ${CMAKE_ARGS} \
-            ..
+              -G '${{ runner.os == 'Windows' && 'Visual Studio 17 2022' || 'Ninja' }}' \
+              -DCMAKE_TOOLCHAIN_FILE:FILEPATH=build/generators/conan_toolchain.cmake \
+              -DCMAKE_BUILD_TYPE="${BUILD_TYPE}" \
+              ${CMAKE_ARGS} \
+              ..
 
       - name: Check protocol autogen files are up-to-date
         working-directory: ${{ env.BUILD_DIR }}
@@ -172,10 +172,10 @@ jobs:
           cmake --build . --target code_gen
           DIFF=$(git -C .. status --porcelain -- include/xrpl/protocol_autogen src/tests/libxrpl/protocol_autogen)
           if [ -n "${DIFF}" ]; then
-            echo "::error::Generated protocol files are out of date"
-            git -C .. diff -- include/xrpl/protocol_autogen src/tests/libxrpl/protocol_autogen
-            echo "${MESSAGE}"
-            exit 1
+              echo "::error::Generated protocol files are out of date"
+              git -C .. diff -- include/xrpl/protocol_autogen src/tests/libxrpl/protocol_autogen
+              echo "${MESSAGE}"
+              exit 1
           fi
 
       - name: Build the binary
@@ -186,18 +186,18 @@ jobs:
           CMAKE_TARGET: ${{ inputs.cmake_target }}
         run: |
           cmake \
-            --build . \
-            --config "${BUILD_TYPE}" \
-            --parallel "${BUILD_NPROC}" \
-            --target "${CMAKE_TARGET}"
+              --build . \
+              --config "${BUILD_TYPE}" \
+              --parallel "${BUILD_NPROC}" \
+              --target "${CMAKE_TARGET}"
 
       - name: Show ccache statistics
         if: ${{ inputs.ccache_enabled }}
         run: |
           ccache --show-stats -vv
           if [ '${{ runner.debug }}' = '1' ]; then
-            cat "${CCACHE_LOGFILE}"
-            curl ${CCACHE_REMOTE_STORAGE%|*}/status || true
+              cat "${CCACHE_LOGFILE}"
+              curl ${CCACHE_REMOTE_STORAGE%|*}/status || true
           fi
 
       - name: Upload the binary (Linux)
@@ -214,7 +214,7 @@ jobs:
         working-directory: ${{ env.BUILD_DIR }}
         run: |
           set -o pipefail
-          ./xrpld --definitions | python3 -m json.tool > server_definitions.json
+          ./xrpld --definitions | python3 -m json.tool >server_definitions.json
 
       - name: Upload server definitions
         if: ${{ github.event.repository.visibility == 'public' && inputs.config_name == 'debian-bookworm-gcc-13-amd64-release' }}
@@ -231,10 +231,10 @@ jobs:
         run: |
           ldd ./xrpld
           if [ "$(ldd ./xrpld | grep -E '(libstdc\+\+|libgcc)' | wc -l)" -eq 0 ]; then
-            echo 'The binary is statically linked.'
+              echo 'The binary is statically linked.'
           else
-            echo 'The binary is dynamically linked.'
-            exit 1
+              echo 'The binary is dynamically linked.'
+              exit 1
           fi
 
       - name: Verify presence of instrumentation (Linux)
@@ -250,12 +250,12 @@ jobs:
         run: |
           ASAN_OPTS="include=${GITHUB_WORKSPACE}/sanitizers/suppressions/runtime-asan-options.txt:suppressions=${GITHUB_WORKSPACE}/sanitizers/suppressions/asan.supp"
           if [[ "${CONFIG_NAME}" == *gcc* ]]; then
-            ASAN_OPTS="${ASAN_OPTS}:alloc_dealloc_mismatch=0"
+              ASAN_OPTS="${ASAN_OPTS}:alloc_dealloc_mismatch=0"
           fi
-          echo "ASAN_OPTIONS=${ASAN_OPTS}" >> ${GITHUB_ENV}
-          echo "TSAN_OPTIONS=include=${GITHUB_WORKSPACE}/sanitizers/suppressions/runtime-tsan-options.txt:suppressions=${GITHUB_WORKSPACE}/sanitizers/suppressions/tsan.supp" >> ${GITHUB_ENV}
-          echo "UBSAN_OPTIONS=include=${GITHUB_WORKSPACE}/sanitizers/suppressions/runtime-ubsan-options.txt:suppressions=${GITHUB_WORKSPACE}/sanitizers/suppressions/ubsan.supp" >> ${GITHUB_ENV}
-          echo "LSAN_OPTIONS=include=${GITHUB_WORKSPACE}/sanitizers/suppressions/runtime-lsan-options.txt:suppressions=${GITHUB_WORKSPACE}/sanitizers/suppressions/lsan.supp" >> ${GITHUB_ENV}
+          echo "ASAN_OPTIONS=${ASAN_OPTS}" >>${GITHUB_ENV}
+          echo "TSAN_OPTIONS=include=${GITHUB_WORKSPACE}/sanitizers/suppressions/runtime-tsan-options.txt:suppressions=${GITHUB_WORKSPACE}/sanitizers/suppressions/tsan.supp" >>${GITHUB_ENV}
+          echo "UBSAN_OPTIONS=include=${GITHUB_WORKSPACE}/sanitizers/suppressions/runtime-ubsan-options.txt:suppressions=${GITHUB_WORKSPACE}/sanitizers/suppressions/ubsan.supp" >>${GITHUB_ENV}
+          echo "LSAN_OPTIONS=include=${GITHUB_WORKSPACE}/sanitizers/suppressions/runtime-lsan-options.txt:suppressions=${GITHUB_WORKSPACE}/sanitizers/suppressions/lsan.supp" >>${GITHUB_ENV}
 
       - name: Run the separate tests
         if: ${{ !inputs.build_only }}
@@ -266,9 +266,9 @@ jobs:
           PARALLELISM: ${{ runner.os == 'Windows' && '1' || steps.nproc.outputs.nproc }}
         run: |
           ctest \
-            --output-on-failure \
-            -C "${BUILD_TYPE}" \
-            -j "${PARALLELISM}"
+              --output-on-failure \
+              -C "${BUILD_TYPE}" \
+              -j "${PARALLELISM}"
 
       - name: Run the embedded tests
         if: ${{ !inputs.build_only }}
@@ -278,7 +278,7 @@ jobs:
         run: |
           set -o pipefail
           # Coverage builds are slower due to instrumentation; use fewer parallel jobs to avoid flakiness
-          [ "$COVERAGE_ENABLED" = "true" ] && BUILD_NPROC=$(( BUILD_NPROC - 2 ))
+          [ "$COVERAGE_ENABLED" = "true" ] && BUILD_NPROC=$((BUILD_NPROC - 2))
           ./xrpld --unittest --unittest-jobs "${BUILD_NPROC}" 2>&1 | tee unittest.log
 
       - name: Show test failure summary
@@ -287,19 +287,19 @@ jobs:
           WORKING_DIR: ${{ runner.os == 'Windows' && format('{0}\{1}', env.BUILD_DIR, inputs.build_type) || env.BUILD_DIR }}
         run: |
           if [ ! -d "${WORKING_DIR}" ]; then
-            echo "Working directory '${WORKING_DIR}' does not exist."
-            exit 0
+              echo "Working directory '${WORKING_DIR}' does not exist."
+              exit 0
           fi
 
           cd "${WORKING_DIR}"
 
           if [ ! -f unittest.log ]; then
-            echo "unittest.log not found; embedded tests may not have run."
-            exit 0
+              echo "unittest.log not found; embedded tests may not have run."
+              exit 0
           fi
 
           if ! grep -E "failed" unittest.log; then
-            echo "Log present but no failure lines found in unittest.log."
+              echo "Log present but no failure lines found in unittest.log."
           fi
       - name: Debug failure (Linux)
         if: ${{ failure() && runner.os == 'Linux' && !inputs.build_only }}
@@ -317,10 +317,10 @@ jobs:
           BUILD_TYPE: ${{ inputs.build_type }}
         run: |
           cmake \
-            --build . \
-            --config "${BUILD_TYPE}" \
-            --parallel "${BUILD_NPROC}" \
-            --target coverage
+              --build . \
+              --config "${BUILD_TYPE}" \
+              --parallel "${BUILD_NPROC}" \
+              --target coverage
 
       - name: Upload coverage report
         if: ${{ github.repository == 'XRPLF/rippled' && !inputs.build_only && env.COVERAGE_ENABLED == 'true' }}

.github/workflows/reusable-check-levelization.yml

@@ -38,9 +38,9 @@ jobs:
         run: |
           DIFF=$(git status --porcelain)
           if [ -n "${DIFF}" ]; then
-            # Print the differences to give the contributor a hint about what to
-            # expect when running levelization on their own machine.
-            git diff
-            echo "${MESSAGE}"
-            exit 1
+              # Print the differences to give the contributor a hint about what to
+              # expect when running levelization on their own machine.
+              git diff
+              echo "${MESSAGE}"
+              exit 1
           fi

.github/workflows/reusable-check-rename.yml

@@ -48,9 +48,9 @@ jobs:
         run: |
           DIFF=$(git status --porcelain)
           if [ -n "${DIFF}" ]; then
-            # Print the differences to give the contributor a hint about what to
-            # expect when running the renaming scripts on their own machine.
-            git diff
-            echo "${MESSAGE}"
-            exit 1
+              # Print the differences to give the contributor a hint about what to
+              # expect when running the renaming scripts on their own machine.
+              git diff
+              echo "${MESSAGE}"
+              exit 1
           fi

.github/workflows/reusable-clang-tidy.yml

@@ -70,13 +70,13 @@ jobs:
         working-directory: ${{ env.BUILD_DIR }}
         run: |
           cmake \
-            -G 'Ninja' \
-            -DCMAKE_TOOLCHAIN_FILE:FILEPATH=build/generators/conan_toolchain.cmake \
-            -DCMAKE_BUILD_TYPE="${BUILD_TYPE}" \
-            -Dtests=ON \
-            -Dwerr=ON \
-            -Dxrpld=ON \
-            ..
+              -G 'Ninja' \
+              -DCMAKE_TOOLCHAIN_FILE:FILEPATH=build/generators/conan_toolchain.cmake \
+              -DCMAKE_BUILD_TYPE="${BUILD_TYPE}" \
+              -Dtests=ON \
+              -Dwerr=ON \
+              -Dxrpld=ON \
+              ..
 
       # clang-tidy needs headers generated from proto files
       - name: Build libxrpl.libpb
@@ -133,7 +133,7 @@ jobs:
       - name: Write issue header
         if: ${{ steps.run_clang_tidy.outcome != 'success' }}
         run: |
-          cat > "${ISSUE_FILE}" <<EOF
+          cat >"${ISSUE_FILE}" <<EOF
           ## Clang-tidy Check Failed
 
           ### Clang-tidy Output:
@@ -144,30 +144,30 @@ jobs:
         if: ${{ steps.run_clang_tidy.outcome != 'success' }}
         run: |
           if [ -f "${OUTPUT_FILE}" ]; then
-            # Extract lines containing 'error:', 'warning:', or 'note:'
-            grep -E '(error:|warning:|note:)' "${OUTPUT_FILE}" > filtered-output.txt || true
+              # Extract lines containing 'error:', 'warning:', or 'note:'
+              grep -E '(error:|warning:|note:)' "${OUTPUT_FILE}" >filtered-output.txt || true
 
-            # If filtered output is empty, use original (might be a different error format)
-            if [ ! -s filtered-output.txt ]; then
-              cp "${OUTPUT_FILE}" filtered-output.txt
-            fi
+              # If filtered output is empty, use original (might be a different error format)
+              if [ ! -s filtered-output.txt ]; then
+                  cp "${OUTPUT_FILE}" filtered-output.txt
+              fi
 
-            # Truncate if too large
-            head -c 60000 filtered-output.txt >> "${ISSUE_FILE}"
-            if [ "$(wc -c < filtered-output.txt)" -gt 60000 ]; then
-              echo "" >> "${ISSUE_FILE}"
-              echo "... (output truncated, see artifacts for full output)" >> "${ISSUE_FILE}"
-            fi
+              # Truncate if too large
+              head -c 60000 filtered-output.txt >>"${ISSUE_FILE}"
+              if [ "$(wc -c <filtered-output.txt)" -gt 60000 ]; then
+                  echo "" >>"${ISSUE_FILE}"
+                  echo "... (output truncated, see artifacts for full output)" >>"${ISSUE_FILE}"
+              fi
 
-            rm filtered-output.txt
+              rm filtered-output.txt
           else
-            echo "No output file found" >> "${ISSUE_FILE}"
+              echo "No output file found" >>"${ISSUE_FILE}"
           fi
 
       - name: Append issue footer
         if: ${{ steps.run_clang_tidy.outcome != 'success' }}
         run: |
-          cat >> "${ISSUE_FILE}" <<EOF
+          cat >>"${ISSUE_FILE}" <<EOF
           \`\`\`
 
           ---
@@ -176,7 +176,7 @@ jobs:
 
       - name: Create issue
         if: ${{ steps.run_clang_tidy.outcome != 'success' && inputs.create_issue_on_failure }}
-        uses: XRPLF/actions/create-issue@36d450d12d301e8410c1b7936e5de70c291cbe36
+        uses: XRPLF/actions/create-issue@2b8bc36af85b88bca0dd7bfac2e2dc05f94ad712
         with:
           title: "Clang-tidy check failed"
           body_file: ${{ env.ISSUE_FILE }}

.github/workflows/reusable-package.yml

@@ -39,7 +39,7 @@ jobs:
         id: generate
         working-directory: .github/scripts/strategy-matrix
         run: |
-          ./generate.py --packaging --config=linux.json >> "${GITHUB_OUTPUT}"
+          ./generate.py --packaging --config=linux.json >>"${GITHUB_OUTPUT}"
 
   generate-version:
     runs-on: ubuntu-latest

.github/workflows/reusable-strategy-matrix.yml

@@ -42,4 +42,4 @@ jobs:
         env:
           GENERATE_CONFIG: ${{ inputs.os != '' && format('--config={0}.json', inputs.os) || '' }}
           GENERATE_OPTION: ${{ inputs.strategy_matrix == 'all' && '--all' || '' }}
-        run: ./generate.py ${GENERATE_OPTION} ${GENERATE_CONFIG} >> "${GITHUB_OUTPUT}"
+        run: ./generate.py ${GENERATE_OPTION} ${GENERATE_CONFIG} >>"${GITHUB_OUTPUT}"

.pre-commit-config.yaml

@@ -66,6 +66,19 @@ repos:
       - id: shfmt
         args: [--write, --indent=4, --case-indent=true]
 
+  - repo: local
+    hooks:
+      - id: format-inline-bash-workflows
+        name: "format `run:` blocks in workflows/actions"
+        entry: ./.github/scripts/format-inline-bash.py
+        language: python
+        files: ^\.github/(workflows|actions)/.*\.ya?ml$
+      - id: format-inline-bash-markdown
+        name: "format ```bash blocks in markdown"
+        entry: ./.github/scripts/format-inline-bash.py
+        language: python
+        files: \.md$
+
   - repo: https://github.com/streetsidesoftware/cspell-cli
     rev: 4643f154907327ee0a2c7038f0296e0dd77d9776 # frozen: v10.0.0
     hooks:

BUILD.md

@@ -151,8 +151,8 @@ git init
 git remote add origin git@github.com:XRPLF/conan-center-index.git
 git sparse-checkout init
 for recipe in "${recipes[@]}"; do
-  echo "Checking out recipe '${recipe}'..."
-  git sparse-checkout add recipes/${recipe}
+    echo "Checking out recipe '${recipe}'..."
+    git sparse-checkout add recipes/${recipe}
 done
 git fetch origin master
 git checkout master
@@ -180,7 +180,7 @@ the new recipe will be automatically pulled from the official Conan Center.
 
 If you see an error similar to the following after running `conan profile show`:
 
-```bash
+```text
 ERROR: Invalid setting '17' is not a valid 'settings.compiler.version' value.
 Possible values are ['5.0', '5.1', '6.0', '6.1', '7.0', '7.3', '8.0', '8.1',
 '9.0', '9.1', '10.0', '11.0', '12.0', '13', '13.0', '13.1', '14', '14.0', '15',

cspell.config.yaml

@@ -93,6 +93,7 @@ words:
   - daria
   - dcmake
   - dearmor
+  - dedented
   - deleteme
   - demultiplexer
   - deserializaton

include/xrpl/basics/Number.h

@@ -2,6 +2,7 @@
 
 #include <xrpl/beast/utility/instrumentation.h>
 
+#include <array>
 #include <concepts>
 #include <cstdint>
 #include <functional>
@@ -67,6 +68,47 @@ isPowerOfTen(T value)
     return logTen(value).has_value();
 }
 
+namespace detail {
+
+/** Builds a table of the powers of 10
+ *
+ * This function is marked consteval, so it can only be run in
+ * a constexpr context. This assures that it is and can only be run at
+ * compile time. Doing it at runtime would be pretty wasteful and
+ * inefficient.
+ */
+constexpr std::size_t kInt64Digits = 20;
+consteval std::array<std::uint64_t, kInt64Digits>
+buildPowersOfTen()
+{
+    std::array<std::uint64_t, kInt64Digits> result{};
+
+    std::uint64_t power = 1;
+    std::size_t exponent = 0;
+    // end the loop early so it doesn't overflow;
+    for (; exponent < result.size() - 1; ++exponent, power *= 10)
+    {
+        result[exponent] = power;
+        if (power > std::numeric_limits<std::uint64_t>::max() / 10)
+            throw std::logic_error("Power of 10 table is too big");
+    }
+    result[exponent] = power;
+    if (power < std::numeric_limits<std::uint64_t>::max() / 10)
+        throw std::logic_error("Power of 10 table is not big enough for the uint64_t type");
+
+    return result;
+}
+
+}  // namespace detail
+
+constexpr std::array<std::uint64_t, detail::kInt64Digits> kPowerOfTen = detail::buildPowersOfTen();
+
+static_assert(kPowerOfTen[0] == 1);
+static_assert(kPowerOfTen[1] == 10);
+static_assert(kPowerOfTen[10] == 10'000'000'000);
+static_assert(
+    isPowerOfTen(kPowerOfTen.back()) && *logTen(kPowerOfTen.back()) == detail::kInt64Digits - 1);
+
 /** MantissaRange defines a range for the mantissa of a normalized Number.
  *
  * The mantissa is in the range [min, max], where
@@ -102,6 +144,7 @@ isPowerOfTen(T value)
 struct MantissaRange final
 {
     using rep = std::uint64_t;
+
     enum class MantissaScale {
         Small,
         // LargeLegacy can be removed when fixCleanup3_2_0 is retired
@@ -115,11 +158,7 @@ struct MantissaRange final
         Enabled = true,
     };
 
-    explicit constexpr MantissaRange(MantissaScale scale)
-        : max(getMax(scale))
-        , internalMin(getInternalMin(scale, min))
-        , cuspRoundingFixEnabled(isCuspFixEnabled(scale))
-        , scale(scale)
+    explicit constexpr MantissaRange(MantissaScale sc) : scale(sc)
     {
         // Keep the error messages terse. Since this is constexpr, if any of these throw, it won't
         // compile, so there's no real need to worry about runtime exceptions here.
@@ -132,8 +171,8 @@ struct MantissaRange final
         // This is a little hacky
         if ((max + 10) / 10 < min)
             throw std::out_of_range("Invalid mantissa range: (max + 10) / 10 < min");
-        if (computeLog(internalMin) != log)
-            throw std::out_of_range("Invalid mantissa range: computeLog(internalMin) != log");
+        if (internalMin != kPowerOfTen[log])
+            throw std::out_of_range("Invalid mantissa range: internalMin != kPowersOfTen[log]");
     }
 
     // Explicitly delete copy and move operations
@@ -144,17 +183,17 @@ struct MantissaRange final
     MantissaRange&
     operator=(MantissaRange&&) = delete;
 
-    rep const max;
+    MantissaScale const scale;
+    int const log{getExponent(scale)};
+    rep const max{getMax(scale, log)};
     rep const min{computeMin(max)};
     /* Used to determine if mantissas are in range, but have fewer digits than max.
      *
      * Unlike min, internalMin is always an exact power of 10, so a mantissa in the internal
      * representation will always have a consistent number of digits.
      */
-    rep const internalMin;
-    int const log{computeLog(min)};
-    CuspRoundingFix const cuspRoundingFixEnabled;
-    MantissaScale const scale;
+    rep const internalMin{getInternalMin(scale, log)};
+    CuspRoundingFix const cuspRoundingFixEnabled{isCuspFixEnabled(scale)};
 
     static MantissaRange const&
     getMantissaRange(MantissaScale scale);
@@ -163,20 +202,40 @@ struct MantissaRange final
     getAllScales();
 
 private:
-    static constexpr rep
-    getMax(MantissaScale scale)
+    static constexpr int
+    getExponent(MantissaScale scale)
     {
         switch (scale)
         {
             case MantissaScale::Small:
-                return 9'999'999'999'999'999ULL;
+                return 15;
+            case MantissaScale::LargeLegacy:
+            case MantissaScale::Large:
+                return 18;
+            // LCOV_EXCL_START
+            default:
+                // If called in a constexpr context, this throw assures that the build fails if an
+                // invalid scale is used.
+                throw std::runtime_error("Unknown mantissa scale");
+                // LCOV_EXCL_STOP
+        }
+    }
+
+    static constexpr rep
+    getMax(MantissaScale scale, int log)
+    {
+        switch (scale)
+        {
+            case MantissaScale::Small:
+                return kPowerOfTen[log + 1] - 1;
             case MantissaScale::LargeLegacy:
             case MantissaScale::Large:
                 return std::numeric_limits<std::int64_t>::max();
             default:
                 // If called in a constexpr context, this throw assures that the build fails if an
                 // invalid scale is used.
-                throw std::runtime_error("Unknown mantissa scale");  // LCOV_EXCL_LINE
+                throw std::runtime_error("Unknown mantissa scale");
+                // LCOV_EXCL_STOP
         }
     }
 
@@ -187,25 +246,13 @@ private:
     }
 
     static constexpr rep
-    getInternalMin(MantissaScale scale, rep min)
+    getInternalMin(MantissaScale scale, int exponent)
     {
-        switch (scale)
-        {
-            case MantissaScale::LargeLegacy:
-            case MantissaScale::Large:
-                return 1'000'000'000'000'000'000ULL;
-            default:
-                if (isPowerOfTen(min))
-                    return min;
-                throw std::runtime_error("Unknown/bad mantissa scale");
-        }
-    }
-
-    static constexpr int
-    computeLog(rep min)
-    {
-        auto const estimate = logTenEstimate(min);
-        return estimate.first + (estimate.second == 1 ? 0 : 1);
+        if (exponent < 0 || exponent >= kPowerOfTen.size())
+            // If called in a constexpr context, this throw assures that the build fails if an
+            // invalid exponent is used.
+            throw std::runtime_error("Invalid exponent");  // LCOV_EXCL_LINE
+        return kPowerOfTen[exponent];
     }
 
     static constexpr CuspRoundingFix
@@ -575,8 +622,7 @@ public:
     template <
         auto MinMantissa,
         auto MaxMantissa,
-        Integral64 T = std::decay_t<decltype(MinMantissa)>,
-        Integral64 TMax = std::decay_t<decltype(MaxMantissa)>>
+        Integral64 T = std::decay_t<decltype(MinMantissa)>>
     [[nodiscard]]
     std::pair<T, int>
     normalizeToRange() const;
@@ -637,7 +683,8 @@ private:
         int& exponent,
         MantissaRange::rep const& minMantissa,
         MantissaRange::rep const& maxMantissa,
-        MantissaRange::CuspRoundingFix cuspRoundingFixEnabled);
+        MantissaRange::CuspRoundingFix cuspRoundingFixEnabled,
+        bool dropped);
 
     [[nodiscard]]
     bool
@@ -876,16 +923,18 @@ Number::isnormal() const noexcept
     return isnormal(kRange);
 }
 
-template <auto MinMantissa, auto MaxMantissa, Integral64 T, Integral64 TMax>
+template <auto MinMantissa, auto MaxMantissa, Integral64 T>
 std::pair<T, int>
 Number::normalizeToRange() const
 {
     static_assert(std::is_same_v<T, std::uint64_t> || std::is_same_v<T, std::int64_t>);
-    static_assert(std::is_same_v<T, TMax>);
+    static_assert(std::is_same_v<T, std::decay_t<decltype(MinMantissa)>>);
+    static_assert(std::is_same_v<T, std::decay_t<decltype(MaxMantissa)>>);
     auto constexpr kMIN = static_cast<T>(MinMantissa);
     auto constexpr kMAX = static_cast<T>(MaxMantissa);
     static_assert(kMIN > 0);
     static_assert(kMIN % 10 == 0);
+    static_assert(isPowerOfTen(static_cast<std::make_unsigned_t<T>>(kMIN)));
     static_assert(kMAX % 10 == 9);
     static_assert((kMAX + 1) / 10 == kMIN);
 

include/xrpl/protocol_autogen/README.md

@@ -15,8 +15,8 @@ Generation requires a one-time setup step to create a virtual environment
 and install Python dependencies, followed by running the generation target:
 
 ```bash
-cmake --build . --target setup_code_gen  # create venv and install dependencies (once)
-cmake --build . --target code_gen        # generate code
+cmake --build . --target setup_code_gen # create venv and install dependencies (once)
+cmake --build . --target code_gen       # generate code
 ```
 
 By default, `CODEGEN_VENV_DIR` points to `.venv` in the project root. The

package/README.md

@@ -74,10 +74,10 @@ VERSION=2.4.0-local
 PKG_RELEASE=1
 
 docker run --rm \
-  -v "$(pwd):/src" \
-  -w /src \
-  "$IMAGE" \
-  ./package/build_pkg.sh --pkg-version "$VERSION" --pkg-release "$PKG_RELEASE"
+    -v "$(pwd):/src" \
+    -w /src \
+    "$IMAGE" \
+    ./package/build_pkg.sh --pkg-version "$VERSION" --pkg-release "$PKG_RELEASE"
 
 # Output:
 #   build/debbuild/*.deb         (DEB + dbgsym .ddeb)
@@ -92,12 +92,12 @@ needed, but the host toolchain replaces the pinned CI image:
 
 ```bash
 cmake \
-  -Dxrpld=ON \
-  -Dxrpld_version=2.4.0-local \
-  -Dtests=OFF \
-  ..
+    -Dxrpld=ON \
+    -Dxrpld_version=2.4.0-local \
+    -Dtests=OFF \
+    ..
 
-cmake --build . --target package       # deb on Debian/Ubuntu, rpm on RHEL
+cmake --build . --target package # deb on Debian/Ubuntu, rpm on RHEL
 ```
 
 The `cmake/XrplPackaging.cmake` module defines the target only if at least one

src/libxrpl/basics/Number.cpp

@@ -179,6 +179,10 @@ public:
     setPositive() noexcept;
     void
     setNegative() noexcept;
+    // Should only be called by doNormalize, and then only for division
+    // operations with remainders.
+    void
+    setDropped() noexcept;
     [[nodiscard]] bool
     isNegative() const noexcept;
 
@@ -251,6 +255,12 @@ Number::Guard::setNegative() noexcept
     sbit_ = 1;
 }
 
+inline void
+Number::Guard::setDropped() noexcept
+{
+    xbit_ = 1;
+}
+
 inline bool
 Number::Guard::isNegative() const noexcept
 {
@@ -397,7 +407,7 @@ Number::Guard::doRoundUp(
                 // _don't_ increment the mantissa. Instead, divide and round recursively. It should
                 // be impossible to recurse more than once, because once the mantissa is divided by
                 // 10, it will be _well_ under maxMantissa and kLargestMantissa, so adding 1 will
-                // have no change of bringing it back over.
+                // have no chance of bringing it back over.
                 doDropDigit(mantissa, exponent);
                 XRPL_ASSERT_PARTS(
                     safeToIncrement(mantissa),
@@ -630,7 +640,6 @@ Number::one()
     return one(kRange);
 }
 
-// Use the member names in this static function for now so the diff is cleaner
 template <class T>
 void
 doNormalize(
@@ -639,7 +648,8 @@ doNormalize(
     int& exponent,
     MantissaRange::rep const& minMantissa,
     MantissaRange::rep const& maxMantissa,
-    MantissaRange::CuspRoundingFix cuspRoundingFixEnabled)
+    MantissaRange::CuspRoundingFix cuspRoundingFixEnabled,
+    bool dropped)
 {
     auto constexpr kMinExponent = Number::kMinExponent;
     auto constexpr kMaxExponent = Number::kMaxExponent;
@@ -663,6 +673,8 @@ doNormalize(
     Guard g;
     if (negative)
         g.setNegative();
+    if (dropped)
+        g.setDropped();
     while (m > maxMantissa)
     {
         if (exponent >= kMaxExponent)
@@ -707,7 +719,12 @@ Number::normalize<uint128_t>(
     internalrep const& maxMantissa,
     MantissaRange::CuspRoundingFix cuspRoundingFixEnabled)
 {
-    doNormalize(negative, mantissa, exponent, minMantissa, maxMantissa, cuspRoundingFixEnabled);
+    // Not used by every compiler version, and thus not necessarily
+    // counted by coverage build
+    // LCOV_EXCL_START
+    doNormalize(
+        negative, mantissa, exponent, minMantissa, maxMantissa, cuspRoundingFixEnabled, false);
+    // LCOV_EXCL_STOP
 }
 
 template <>
@@ -720,7 +737,12 @@ Number::normalize<unsigned long long>(
     internalrep const& maxMantissa,
     MantissaRange::CuspRoundingFix cuspRoundingFixEnabled)
 {
-    doNormalize(negative, mantissa, exponent, minMantissa, maxMantissa, cuspRoundingFixEnabled);
+    // Not used by every compiler version, and thus not necessarily
+    // counted by coverage build
+    // LCOV_EXCL_START
+    doNormalize(
+        negative, mantissa, exponent, minMantissa, maxMantissa, cuspRoundingFixEnabled, false);
+    // LCOV_EXCL_STOP
 }
 
 template <>
@@ -733,7 +755,8 @@ Number::normalize<unsigned long>(
     internalrep const& maxMantissa,
     MantissaRange::CuspRoundingFix cuspRoundingFixEnabled)
 {
-    doNormalize(negative, mantissa, exponent, minMantissa, maxMantissa, cuspRoundingFixEnabled);
+    doNormalize(
+        negative, mantissa, exponent, minMantissa, maxMantissa, cuspRoundingFixEnabled, false);
 }
 
 void
@@ -942,78 +965,164 @@ Number::operator/=(Number const& y)
         return *this;
     // n* = numerator
     // d* = denominator
-    // *p = negative (positive?)
+    // z* = result (quotient)
+    // *p = negative (p for positive, even though the value means not
+    //      positive?)
     // *s = sign
     // *m = mantissa
     // *e = exponent
 
-    auto [np, nm, ne] = toInternal(range);
+    // Create the mantissas as 128-bit unsigned, since that's what we
+    // need to work with.
+    auto const [np, nm, ne] = toInternal<uint128_t>(range);
     int const ns = (np ? -1 : 1);
 
-    auto [dp, dm, de] = y.toInternal(range);
+    auto const [dp, dm, de] = y.toInternal<uint128_t>(range);
     int const ds = (dp ? -1 : 1);
 
     auto const& minMantissa = range.min;
     auto const& maxMantissa = range.max;
     auto const cuspRoundingFixEnabled = range.cuspRoundingFixEnabled;
 
-    // Shift by 10^17 gives greatest precision while not overflowing
-    // uint128_t or the cast back to int64_t
-    // TODO: Can/should this be made bigger for kLARGE_RANGE?
-    // log(2^128,10) ~ 38.5
-    // kLARGE_RANGE.log = 18, fits in 10^19
-    // f can be up to 10^(38-19) = 10^19 safely
-    bool const small = range.scale == MantissaRange::MantissaScale::Small;
-    uint128_t const f = small ? 100'000'000'000'000'000 : 10'000'000'000'000'000'000ULL;
-    XRPL_ASSERT_PARTS(f >= minMantissa * 10, "Number::operator/=", "factor expected size");
+    // Division operates on two large integers (16-digit for small
+    // mantissas, 19-digit for large) using integer math. If the values
+    // were just divided directly, the result would be only ever be one
+    // digit or zero - not very useful.
+    // e.g. 9'876'543'210'987'654 / 1'234'567'890'123'456 = 8
+    //      1'234'567'890'123'456 / 9'876'543'210'987'654 = 0
+    // Introduce a power-of-ten multiplication factor for the numerator
+    // which will ensure the result has a meaningful number of digits.
+    //
+    // Consider numbers with a 2-digit mantissa:
+    // * Assume both numbers have an exponent of 0, using "ToNearest" rounding
+    // * 23 / 67 = 0
+    // * Use a factor of 10^4
+    // * 230'000 / 67 = 3432 with an exponent of -4
+    // * The normalized result will be 34, exponent -2, or 0.34
+    //
+    // The most extreme results are 10/99 and 99/10
+    // * 100'000 / 99 =  1'010e-4 = 10e-2 or 0.10
+    // * 990'000 / 10 = 99'000e-4 = 99e-1 or 9.9
+    //
+    // Note that the computations give 2 or 3 digits after the
+    // decimal point to determine which way to round for most scenarios.
+    //
+    // For small mantissas (where the MantissaRange.log == 15), shifting by 10^17 gives sufficient
+    // precision while not overflowing uint128_t or the cast back to int64_t. (This is legacy
+    // behavior, which must not be changed.)
+    //
+    // For large mantissas (where the MantissaRange.log == 18), a shift by 10^20 would be optimal
+    // for most scenarios. However, larger mantissa values would overflow 2^128.
+    //
+    // * log(2^128,10) ~ 38.5
+    // * largeRange.log = 18, fits in 10^19
+    // * The expanded numerator must fit in 10^38
+    // * f not be more than 10^(38-19) = 10^19 safely
+    //
+    // So, we do the division into stages:
+    //
+    // Stage 1: Use the same factor of 10^17, for the initial division. This
+    // will frequently not result in a whole number quotient.
+    //
+    // Stage 2: If there is a remainder from the first step, repeat the
+    // process with a "correction" factor of 10^5. Shift the
+    // result of Stage 1 over by 5 places, and add the second result to it.
+    // This is equivalent to if we had used an initial factor of 10^22,
+    // a couple digits more than we actually need.
+    //
+    // Stage 3: If there is still a remainder, and the CuspRoundingFix
+    // is enabled, pass a flag indicating such to doNormalize. The Guard
+    // in doNormalize will treat that flag as if non-zero digits had
+    // been dropped from the mantissa when shrinking it into range.
+    // This is only relevant when rounding away from zero (Upward for
+    // positive numbers, Downward for negative), or if the "regular"
+    // remainder is exactly 0.5 for "ToNearest". This will give the
+    // rounding the most accurate result possible, as if infinite
+    // precision was used in the initial calculation.
 
-    // unsigned denominator
-    auto const dmu = static_cast<uint128_t>(dm);
-    // correctionFactor can be anything between 10 and f, depending on how much
-    // extra precision we want to only use for rounding with the
-    // kLARGE_RANGE. Three digits seems like plenty, and is more than
-    // the kSMALL_RANGE uses.
-    uint128_t const correctionFactor = 1'000;
+    // Stage 1: Do the initial division with a factor of 10^17.
+    auto constexpr factorExponent = 17;
+
+    uint128_t constexpr f = kPowerOfTen[factorExponent];
 
     auto const numerator = uint128_t(nm) * f;
 
-    auto zm = numerator / dmu;
-    auto ze = ne - de - (small ? 17 : 19);
-    bool zn = (ns * ds) < 0;
-    if (!small)
+    auto zm = numerator / dm;
+    auto ze = ne - de - factorExponent;
+    bool zp = (ns * ds) < 0;
+    // dropped is used in the same way as Guard::xbit_. In the case of
+    // division, it indicates if there's any remainder left over after
+    // we have been as precise as reasonable. If there is, it would be as
+    // if we were using infinite precision math, and a non-zero digit
+    // had been shifted off the end of the result when normalizing.
+    bool dropped = false;
+
+    if (range.scale != MantissaRange::MantissaScale::Small)
     {
-        // Virtually multiply numerator by correctionFactor. Since that would
-        // overflow in the existing uint128_t, we'll do that part separately.
+        // Stage 2
+        //
+        // If there is a remainder, treat it as a secondary numerator.
+        // Multiply by correctionFactor separately from stage 1.
         // The math for this would work for small mantissas, but we need to
-        // preserve existing behavior.
+        // preserve legacy behavior.
         //
         // Consider:
-        // ((numerator * correctionFactor) / dmu) / correctionFactor
-        // = ((numerator / dmu) * correctionFactor) / correctionFactor)
+        // ((numerator * correctionFactor) / dm) / correctionFactor
+        // = ((numerator / dm) * correctionFactor) / correctionFactor)
         //
         // But that assumes infinite precision. With integer math, this is
         // equivalent to
         //
-        // = ((numerator / dmu * correctionFactor)
-        //   + ((numerator % dmu) * correctionFactor) / dmu) / correctionFactor
+        // = ((numerator / dm * correctionFactor)
+        //   + ((numerator % dm) * correctionFactor) / dm) / correctionFactor
+        // = ((zm * correctionFactor)
+        //   + (remainder * correctionFactor) / dm) / correctionFactor
         //
-        // We have already set `mantissa_ = numerator / dmu`. Now we
-        // compute `remainder = numerator % dmu`, and if it is
-        // nonzero, we do the rest of the arithmetic. If it's zero, we can skip
-        // it.
-        auto const remainder = (numerator % dmu);
+        // The trick is that multiplication by correctionFactor is done on the mantissa, but
+        // division by correctionFactor is done by modifying the exponent, so no precision is lost
+        // until we normalize.
+        //
+        // If remainder is zero, we can skip this stage entirely because
+        // the first stage gave an exact answer.
+        auto constexpr correctionExponent = 5;
+        uint128_t constexpr correctionFactor = kPowerOfTen[correctionExponent];
+        static_assert(factorExponent + correctionExponent == 22);
+
+        auto const remainder = (numerator % dm);
         if (remainder != 0)
         {
-            zm *= correctionFactor;
-            auto const correction = remainder * correctionFactor / dmu;
-            zm += correction;
-            // divide by 1000 by moving the exponent, so we don't lose the
-            // integer value we just computed
-            ze -= 3;
+            auto const partialNumerator = remainder * correctionFactor;
+            auto const correction = partialNumerator / dm;
+
+            // If the correction is zero, we do not have to make any
+            // modifications to z*, because it will not have any
+            // effect on the final result. (We'd be adding a bunch of
+            // zeros to the end of zm that would just be removed in
+            // normalize.) However, if that is the case, then Stage 3 is
+            // even more important for accuracy.
+            if (correction != 0)
+            {
+                zm *= correctionFactor;
+                // divide by the correctionFactor by moving the exponent, so we don't lose the
+                // integer value we just computed
+                ze -= correctionExponent;
+
+                zm += correction;
+            }
+
+            // Stage 3: If there's still anything left, and the cusp
+            // rounding fix is enabled, flag if there is still
+            // a remainder from stage 2.
+            bool const useTrailingRemainder =
+                cuspRoundingFixEnabled == MantissaRange::CuspRoundingFix::Enabled;
+            if (useTrailingRemainder)
+            {
+                dropped = partialNumerator % dm != 0;
+            }
         }
     }
-    normalize(zn, zm, ze, minMantissa, maxMantissa, cuspRoundingFixEnabled);
-    fromInternal(zn, zm, ze, &range);
+    doNormalize(zp, zm, ze, minMantissa, maxMantissa, cuspRoundingFixEnabled, dropped);
+    fromInternal(zp, zm, ze, &range);
     XRPL_ASSERT_PARTS(isnormal(range), "xrpl::Number::operator/=", "result is normalized");
 
     return *this;

src/test/basics/Number_test.cpp

@@ -6,12 +6,15 @@
 #include <xrpl/protocol/SystemParameters.h>
 #include <xrpl/protocol/XRPAmount.h>
 
+// NOLINTNEXTLINE(misc-include-cleaner)
+#include <boost/multiprecision/cpp_dec_float.hpp>
 #include <boost/multiprecision/number.hpp>
 
 #include <array>
 #include <cctype>
 #include <chrono>
 #include <cstdint>
+#include <iomanip>
 #include <limits>
 #include <map>
 #include <sstream>
@@ -41,6 +44,40 @@ class Number_test : public beast::unit_test::Suite
         return out;
     }
 
+    using dec = boost::multiprecision::cpp_dec_float_50;
+
+    template <class T = dec>
+    static T
+    pow10(int n)
+    {
+        if (n == 0)
+            return 1;
+        if (n == 1)
+            return 10;
+
+        if (n > 1)
+        {
+            auto r = pow10<T>(n / 2);
+            r *= r;
+            if (n % 2 != 0)
+                r *= 10;
+            return r;
+        }
+
+        // n < 0
+        T p = 1;
+        p /= pow10<T>(-n);
+        return p;
+    }
+
+    static std::string
+    fmt(dec const& v)
+    {
+        std::ostringstream os;
+        os << std::setprecision(40) << v;
+        return os.str();
+    }
+
 public:
     void
     testZero()
@@ -1949,39 +1986,249 @@ public:
     void
     testUpwardRoundsDown()
     {
-        testcase << "upward rounding produces a value below exact at kLargestMantissa cusp";
+        auto const scale = Number::getMantissaScale();
+        {
+            testcase << "upward rounding produces a value below exact at kLargestMantissa cusp"
+                     << to_string(scale);
 
-        NumberMantissaScaleGuard const mg{MantissaRange::MantissaScale::Large};
-        NumberRoundModeGuard const rg{Number::RoundingMode::Upward};
+            NumberRoundModeGuard const rg{Number::RoundingMode::Upward};
 
-        constexpr std::int64_t kAValue = 1'000'000'000'000'049'863LL;
-        constexpr std::int64_t kBValue = 9'223'372'036'854'315'903LL;
+            constexpr std::int64_t kAValue = 1'000'000'000'000'049'863LL;
+            constexpr std::int64_t kBValue = 9'223'372'036'854'315'903LL;
 
-        Number const a = kAValue;
-        Number const b = kBValue;
-        Number const product = a * b;
+            Number const a = kAValue;
+            Number const b = kBValue;
+            Number const product = a * b;
 
-        // Exact reference in BigInt.
-        BigInt const exactProduct = BigInt(kAValue) * BigInt(kBValue);
+            // Exact reference in BigInt.
+            BigInt const exactProduct = BigInt(kAValue) * BigInt(kBValue);
 
-        // What Number actually stored.
-        BigInt storedValue = BigInt(product.mantissa());
-        for (int i = 0; i < product.exponent(); ++i)
-            storedValue *= 10;
+            // What Number actually stored.
+            BigInt storedValue = BigInt(product.mantissa());
+            for (int i = 0; i < product.exponent(); ++i)
+                storedValue *= 10;
 
-        BigInt const signedDifference = storedValue - exactProduct;
+            BigInt const signedDifference = storedValue - exactProduct;
 
-        log << "\n"
-            << "  a              = " << fmt(BigInt(kAValue)) << "\n"
-            << "  b              = " << fmt(BigInt(kBValue)) << "\n"
-            << "  exact a*b      = " << fmt(exactProduct) << "\n"
-            << "  stored         = " << fmt(storedValue) << "\n"
-            << "  stored - exact = " << fmt(signedDifference) << "\n"
-            << "  upward         = " << (signedDifference >= 0 ? "held" : "VIOLATED") << "\n";
+            log << "\n"
+                << "  a              = " << fmt(BigInt(kAValue)) << "\n"
+                << "  b              = " << fmt(BigInt(kBValue)) << "\n"
+                << "  exact a*b      = " << fmt(exactProduct) << "\n"
+                << "  stored         = " << fmt(storedValue) << "\n"
+                << "  stored - exact = " << fmt(signedDifference) << "\n"
+                << "  upward         = " << (signedDifference >= 0 ? "held" : "VIOLATED") << "\n"
+                << " stored.mantissa = " << product.mantissa() << "\n"
+                << " stored.exponent = " << product.exponent() << "\n";
+            log.flush();
 
-        BEAST_EXPECT(signedDifference >= 0);
-        BEAST_EXPECT(product.mantissa() == (std::numeric_limits<std::int64_t>::max() / 10) + 1);
-        BEAST_EXPECT(product.exponent() == 19);
+            switch (scale)
+            {
+                case MantissaRange::MantissaScale::Large:
+                    BEAST_EXPECT(signedDifference >= 0);
+                    BEAST_EXPECT(signedDifference < pow10<BigInt>(product.exponent()));
+                    BEAST_EXPECT(
+                        product.mantissa() == (std::numeric_limits<std::int64_t>::max() / 10) + 1);
+                    BEAST_EXPECT(product.exponent() == 19);
+                    break;
+
+                case MantissaRange::MantissaScale::LargeLegacy:
+                    BEAST_EXPECT(signedDifference < 0);
+                    BEAST_EXPECT(
+                        product.mantissa() ==
+                        (std::numeric_limits<std::int64_t>::max() / 100) * 100);
+                    BEAST_EXPECT(product.exponent() == 18);
+                    break;
+
+                case MantissaRange::MantissaScale::Small:
+                    // The seemingly weird rounding here is because
+                    // a & b are both normalized, and both round up when
+                    // being converted to Number, so you're really
+                    // getting 1_000_000_000_000_050 * 9_223_372_036_854_316
+                    BEAST_EXPECT(signedDifference >= 0);
+                    BEAST_EXPECT(
+                        product.mantissa() ==
+                        (std::numeric_limits<std::int64_t>::max() / 1000) + 3);
+                    BEAST_EXPECT(product.exponent() == 21);
+                    break;
+            }
+        }
+
+        {
+            /* Companion regression for the kMaxRep cusp behavior, but for
+             * `operator/=` on the cusp-fix-ENABLED `Large` scale.
+             *
+             * Before the dropped-remainder fix, `operator/=` with Upward
+             * rounding could return a value STRICTLY LESS than the exact quotient,
+             * violating Upward's directional invariant.
+             *
+             * Mechanism (fix-enabled path):
+             *   1. `operator/=` computes `numerator = nm * 10^17` and
+             *      `zm = numerator / dm` (integer division, truncates remainder).
+             *   2. If `remainder != 0`, the correction block runs:
+             *        zm *= 100000
+             *        correction = (remainder * 100000) / dm  // also truncates
+             *        zm += correction
+             *        ze -= 5
+             *      The truncation in `correction` discards a sub-1/100000 residual.
+             *   3. `normalize`'s shift loop reduces zm to fit, but the discarded
+             *      residual is BELOW the Guard's visibility, so the Guard sees fraction = 0.
+             *   4. Under Upward + positive, `round()` returns -1 (no round-up), and
+             *      the algorithm returns the truncated zm
+             */
+            testcase << "operator/= Upward on Large returns value < truth " << to_string(scale);
+
+            NumberRoundModeGuard const roundGuard{Number::RoundingMode::Upward};
+
+            constexpr std::int64_t aValue = 2LL;
+            constexpr std::int64_t bValue = 1'000'000'000'000'000'007LL;
+            // bValue = 10^18 + 7 (prime, in [minMantissa, kMaxRep]).
+
+            Number const a{aValue, 0};
+            Number const b{bValue, 0};
+            Number const quotient = a / b;
+
+            dec const exact = dec(aValue) / dec(bValue);
+            dec const stored = dec(quotient.mantissa()) * pow10(quotient.exponent());
+            dec const diff = stored - exact;
+
+            log << "\n"
+                << "  a                 = " << aValue << "\n"
+                << "  b                 = " << bValue << "\n"
+                << "  exact a/b         = " << fmt(exact) << "\n"
+                << "  stored a/b        = " << fmt(stored) << "\n"
+                << "  stored - exact    = " << fmt(diff)
+                << "    (negative => Upward gave value BELOW truth)\n"
+                << "  quotient.mantissa = " << quotient.mantissa() << "\n"
+                << "  quotient.exponent = " << quotient.exponent() << "\n";
+            log.flush();
+
+            // Upward invariant: stored >= exact. Bug: stored < exact.
+            switch (scale)
+            {
+                case MantissaRange::MantissaScale::Large:
+                    BEAST_EXPECT(stored >= exact);
+                    BEAST_EXPECT(diff < pow10(quotient.exponent()));
+                    break;
+
+                case MantissaRange::MantissaScale::LargeLegacy:
+                    BEAST_EXPECT(stored < exact);
+                    BEAST_EXPECT(diff >= -pow10(quotient.exponent()));
+                    break;
+
+                case MantissaRange::MantissaScale::Small:
+                    // Small mantissa doesn't have the correction for
+                    // dropped remainders
+                    BEAST_EXPECT(stored < exact);
+                    break;
+            }
+        }
+        {
+            /* Companion test case for Upward positive operator/=: Downward negative
+             */
+            testcase << "operator/= Downward on Large returns value < truth " << to_string(scale);
+
+            NumberRoundModeGuard const roundGuard{Number::RoundingMode::Downward};
+
+            constexpr std::int64_t aValue = -2LL;
+            constexpr std::int64_t bValue = 1'000'000'000'000'000'007LL;
+            // bValue = 10^18 + 7 (prime, in [minMantissa, kMaxRep]).
+
+            Number const a{aValue, 0};
+            Number const b{bValue, 0};
+            Number const quotient = a / b;
+
+            dec const exact = dec(aValue) / dec(bValue);
+            dec const stored = dec(quotient.mantissa()) * pow10(quotient.exponent());
+            dec const diff = stored - exact;
+
+            log << "\n"
+                << "  a                 = " << aValue << "\n"
+                << "  b                 = " << bValue << "\n"
+                << "  exact a/b         = " << fmt(exact) << "\n"
+                << "  stored a/b        = " << fmt(stored) << "\n"
+                << "  stored - exact    = " << fmt(diff)
+                << "    (positive => Downward gave value ABOVE truth)\n"
+                << "  quotient.mantissa = " << quotient.mantissa() << "\n"
+                << "  quotient.exponent = " << quotient.exponent() << "\n";
+            log.flush();
+
+            // invariant: stored <= exact. Bug: stored > exact.
+            switch (scale)
+            {
+                case MantissaRange::MantissaScale::Large:
+                    BEAST_EXPECT(stored <= exact);
+                    BEAST_EXPECT(diff > -pow10(quotient.exponent()));
+                    break;
+
+                case MantissaRange::MantissaScale::LargeLegacy:
+                    BEAST_EXPECT(stored > exact);
+                    BEAST_EXPECT(diff <= pow10(quotient.exponent()));
+                    break;
+
+                case MantissaRange::MantissaScale::Small:
+                    // Small mantissa doesn't have the correction for
+                    // dropped remainders
+                    BEAST_EXPECT(stored < exact);
+                    break;
+            }
+        }
+        {
+            /* Companion test case for Upward positive operator/=: ToNearest
+             *
+             * With ToNearest, if the dropped digits are exactly "5", then the mantissa will be
+             * rounded to even. The numbers below result in a value where the unrounded mantissa
+             * ends in an even digit, and "infinite precision" would drop
+             * "500000000000000000145...", but doNormalize only sees "5". Without the rounding fix,
+             * doNormalize rounds down to the even value. With the rounding fix, doNormalize knows
+             * there are more digits beyond "5", and so rounds _up_ to the odd value.
+             */
+            testcase << "operator/= ToNearest on Large returns value < truth " << to_string(scale);
+
+            NumberRoundModeGuard const roundGuard{Number::RoundingMode::ToNearest};
+
+            constexpr std::int64_t aValue = 1'269'917'268'816'087'809LL;
+            constexpr std::int64_t bValue = 3'458'525'013'821'685'511LL;
+            // bValue = 10^18 + 7 (prime, in [minMantissa, kMaxRep]).
+
+            Number const a{aValue, 0};
+            Number const b{bValue, 0};
+            Number const quotient = a / b;
+
+            dec const exact = dec(aValue) / dec(bValue);
+            dec const stored = dec(quotient.mantissa()) * pow10(quotient.exponent());
+            dec const diff = stored - exact;
+
+            log << "\n"
+                << "  a                 = " << aValue << "\n"
+                << "  b                 = " << bValue << "\n"
+                << "  exact a/b         = " << fmt(exact) << "\n"
+                << "  stored a/b        = " << fmt(stored) << "\n"
+                << "  stored - exact    = " << fmt(diff)
+                << "    (negative => ToNearest gave value BELOW truth)\n"
+                << "  quotient.mantissa = " << quotient.mantissa() << "\n"
+                << "  quotient.exponent = " << quotient.exponent() << "\n";
+            log.flush();
+
+            // invariant: stored >= exact. Bug: stored < exact.
+            switch (scale)
+            {
+                case MantissaRange::MantissaScale::Large:
+                    BEAST_EXPECT(stored >= exact);
+                    BEAST_EXPECT(diff < pow10(quotient.exponent()));
+                    break;
+
+                case MantissaRange::MantissaScale::LargeLegacy:
+                    BEAST_EXPECT(stored < exact);
+                    BEAST_EXPECT(diff >= -pow10(quotient.exponent()));
+                    break;
+
+                case MantissaRange::MantissaScale::Small:
+                    // Small mantissa doesn't have the correction for
+                    // dropped remainders
+                    BEAST_EXPECT(stored < exact);
+                    break;
+            }
+        }
     }
 
     void
@@ -2012,9 +2259,9 @@ public:
             testRounding();
             testInt64();
             testNormalizeToRange();
+
+            testUpwardRoundsDown();
         }
-        // This test sets its own number range
-        testUpwardRoundsDown();
     }
 };