psyqo: make fixed-point iDiv/dDiv constexpr

[nicolasnoble]

Move the FixedPointInternals iDiv/dDiv division helpers out of fixed-point.cpp
into the header as constexpr, so fixed-point division can be evaluated at
compile time. The out-of-line definitions are removed; the bodies are
unchanged. printInt stays in the .cpp.

[coderabbitai]

📝 Walkthrough

Walkthrough

The iDiv and dDiv fixed-point division helper functions are relocated from src/mips/psyqo/src/fixed-point.cpp into src/mips/psyqo/fixed-point.hh as inline constexpr definitions, replacing their prior forward declarations and removing the corresponding implementations from the cpp file.

Changes

Fixed-point division helper relocation

Layer / File(s)	Summary
Inline constexpr division implementation src/mips/psyqo/fixed-point.hh, src/mips/psyqo/src/fixed-point.cpp	iDiv and dDiv are now defined inline as constexpr functions in the header (implementing shift/subtract division and signed division via absolute values), while their prior implementations are removed from the cpp file.

Estimated code review effort: 3 (Moderate) | ~20 minutes

🚥 Pre-merge checks | ✅ 5

✅ Passed checks (5 passed)

Check name	Status	Explanation
Title check	✅ Passed	The title accurately summarizes the main change: making the fixed-point division helpers constexpr in the header.
Description check	✅ Passed	The description matches the changeset by stating the helpers moved to the header as constexpr and printInt stayed in the .cpp.
Docstring Coverage	✅ Passed	No functions found in the changed files to evaluate docstring coverage. Skipping docstring coverage check.
Linked Issues check	✅ Passed	Check skipped because no linked issues were found for this pull request.
Out of Scope Changes check	✅ Passed	Check skipped because no linked issues were found for this pull request.

✨ Finishing Touches

🧪 Generate unit tests (beta)

• Create PR with unit tests

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}

[coderabbitai]

Actionable comments posted: 1

🤖 Prompt for all review comments with AI agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

Inline comments:
In `@src/mips/psyqo/fixed-point.hh`:
- Around line 72-83: The dDiv helper in fixed-point.hh negates signed inputs
directly, which still breaks when a or b is INT32_MIN. Update dDiv (used by
FixedPoint::operator/ and FixedPoint::operator/=) to guard the minimum 32-bit
value before negation, either by widening to a larger temporary type or by
special-casing INT32_MIN so constant evaluation remains valid.

🪄 Autofix (Beta)

Fix all unresolved CodeRabbit comments on this PR:

• Push a commit to this branch (recommended)
• Create a new PR with the fixes

---

ℹ️ Review info

⚙️ Run configuration

Configuration used: Path: .coderabbit.yaml

Review profile: CHILL

Plan: Pro

Run ID: d4eebe76-e534-43e3-a8dc-f94077b4a052

📥 Commits

Reviewing files that changed from the base of the PR and between e909cc3 and fa42b7b.

📒 Files selected for processing (2)

• src/mips/psyqo/fixed-point.hh
• src/mips/psyqo/src/fixed-point.cpp

💤 Files with no reviewable changes (1)

• src/mips/psyqo/src/fixed-point.cpp

[coderabbitai]

🎯 Functional Correctness | 🟡 Minor | ⚡ Quick win

🧩 Analysis chain

🏁 Script executed:

#!/bin/bash
# Find compile-time (constexpr/consteval) usages that could feed dDiv with signed division.
rg -nP -C3 '\bdDiv\s*\(' src/mips/psyqo
# Look for constexpr/consteval contexts invoking FixedPoint division on signed 32-bit types.
rg -nP -C3 'operator/|operator/=' src/mips/psyqo/fixed-point.hh

Repository: grumpycoders/pcsx-redux

Length of output: 3784

---

🏁 Script executed:

#!/bin/bash
set -euo pipefail

# Inspect the relevant parts of fixed-point.hh around dDiv and the signed operators.
sed -n '1,140p' src/mips/psyqo/fixed-point.hh
printf '\n---\n'
sed -n '300,430p' src/mips/psyqo/fixed-point.hh
printf '\n---\n'
sed -n '470,520p' src/mips/psyqo/fixed-point.hh

Repository: grumpycoders/pcsx-redux

Length of output: 10563

---

🏁 Script executed:

#!/bin/bash
set -euo pipefail

# Find constructors, raw accessors, and any direct uses of INT32_MIN / numeric_limits in the fixed-point implementation.
rg -n -C2 'raw\(|INT32_MIN|numeric_limits|FixedPoint\s*\(|constexpr\s+FixedPoint' src/mips/psyqo/fixed-point.hh

Repository: grumpycoders/pcsx-redux

Length of output: 6974

---

Guard INT32_MIN before negating in dDiv

FixedPoint::operator/ and operator/= for signed 32-bit values call this directly, so a raw INT32_MIN still makes constant evaluation ill-formed when a = -a; or b = -b;. Widen the temporary or special-case the minimum value.

🤖 Prompt for AI Agents

Verify each finding against current code. Fix only still-valid issues, skip the
rest with a brief reason, keep changes minimal, and validate.

In `@src/mips/psyqo/fixed-point.hh` around lines 72 - 83, The dDiv helper in
fixed-point.hh negates signed inputs directly, which still breaks when a or b is
INT32_MIN. Update dDiv (used by FixedPoint::operator/ and
FixedPoint::operator/=) to guard the minimum 32-bit value before negation,
either by widening to a larger temporary type or by special-casing INT32_MIN so
constant evaluation remains valid.