Crafter.Debug/implementations/Crafter.Debug.Core-Value.cpp

// SPDX-License-Identifier: LGPL-3.0-only
// SPDX-FileCopyrightText: Copyright (C) 2026 Catcrafts®

module Crafter.Debug.Core:Value_impl;
import std;
import :Value;

using namespace Crafter::Debug;

namespace {
    // Nesting deeper than this is refused rather than recursed into: the
    // parser is recursive, and its text comes from another process.
    constexpr std::uint32_t maxDepth = 256;

    bool IsDigit(char c) {
        return c >= '0' && c <= '9';
    }

    void AppendUtf8(std::string& out, std::uint32_t code) {
        if (code < 0x80) {
            out += static_cast<char>(code);
        } else if (code < 0x800) {
            out += static_cast<char>(0xC0 | (code >> 6));
            out += static_cast<char>(0x80 | (code & 0x3F));
        } else if (code < 0x10000) {
            out += static_cast<char>(0xE0 | (code >> 12));
            out += static_cast<char>(0x80 | ((code >> 6) & 0x3F));
            out += static_cast<char>(0x80 | (code & 0x3F));
        } else {
            out += static_cast<char>(0xF0 | (code >> 18));
            out += static_cast<char>(0x80 | ((code >> 12) & 0x3F));
            out += static_cast<char>(0x80 | ((code >> 6) & 0x3F));
            out += static_cast<char>(0x80 | (code & 0x3F));
        }
    }

    // The length of the valid UTF-8 sequence that starts at text[at], or 0
    // when it is not one: a stray continuation byte, a truncated or overlong
    // sequence, a surrogate, or a code point above U+10FFFF.
    std::size_t Utf8SequenceLength(std::string_view text, std::size_t at) {
        const auto byte = [&](std::size_t i) { return static_cast<std::uint8_t>(text[at + i]); };
        const std::uint8_t lead = byte(0);
        std::size_t length = 0;
        std::uint8_t low = 0x80;
        std::uint8_t high = 0xBF;
        if (lead >= 0xC2 && lead <= 0xDF) {
            length = 2;
        } else if (lead >= 0xE0 && lead <= 0xEF) {
            length = 3;
            if (lead == 0xE0) low = 0xA0;
            if (lead == 0xED) high = 0x9F;
        } else if (lead >= 0xF0 && lead <= 0xF4) {
            length = 4;
            if (lead == 0xF0) low = 0x90;
            if (lead == 0xF4) high = 0x8F;
        } else {
            return 0;
        }
        if (at + length > text.size()) return 0;
        if (byte(1) < low || byte(1) > high) return 0;
        for (std::size_t i = 2; i < length; ++i) {
            if (byte(i) < 0x80 || byte(i) > 0xBF) return 0;
        }
        return length;
    }

    // `text` as a JSON string. Invalid UTF-8 becomes U+FFFD, so the output
    // always parses, whatever bytes an app put in a name.
    void AppendEscaped(std::string& out, std::string_view text) {
        out += '"';
        std::size_t i = 0;
        while (i < text.size()) {
            const std::uint8_t c = static_cast<std::uint8_t>(text[i]);
            if (c < 0x80) {
                switch (c) {
                case '"': out += "\\\""; break;
                case '\\': out += "\\\\"; break;
                case '\b': out += "\\b"; break;
                case '\f': out += "\\f"; break;
                case '\n': out += "\\n"; break;
                case '\r': out += "\\r"; break;
                case '\t': out += "\\t"; break;
                default:
                    if (c < 0x20 || c == 0x7F) {
                        out += std::format("\\u{:04x}", c);
                    } else {
                        out += static_cast<char>(c);
                    }
                }
                ++i;
                continue;
            }
            const std::size_t length = Utf8SequenceLength(text, i);
            if (length == 0) {
                out += "\xEF\xBF\xBD";
                ++i;
                continue;
            }
            out.append(text.substr(i, length));
            i += length;
        }
        out += '"';
    }

    void AppendNumber(std::string& out, double number) {
        if (std::isnan(number)) {
            out += "\"NaN\"";
            return;
        }
        if (std::isinf(number)) {
            out += number > 0 ? "\"Infinity\"" : "\"-Infinity\"";
            return;
        }
        std::array<char, 32> buffer{};
        const auto [end, error] = std::to_chars(buffer.data(), buffer.data() + buffer.size(), number);
        const std::string_view digits(buffer.data(), static_cast<std::size_t>(end - buffer.data()));
        out += digits;
        // A whole double keeps a fraction, so it reads back as a Number.
        if (digits.find_first_of(".eE") == std::string_view::npos) out += ".0";
    }

    class Parser {
    public:
        explicit Parser(std::string_view text) : text(text) {}

        Value Document() {
            SkipWhitespace();
            Value value = Any(0);
            SkipWhitespace();
            if (position != text.size()) Fail("unexpected text after the value");
            return value;
        }

    private:
        std::string_view text;
        std::size_t position = 0;

        [[noreturn]] void Fail(std::string_view what) const {
            throw ParseError(position, what);
        }

        bool AtEnd() const {
            return position >= text.size();
        }

        bool Peek(char c) const {
            return !AtEnd() && text[position] == c;
        }

        bool Consume(char c) {
            if (!Peek(c)) return false;
            ++position;
            return true;
        }

        void SkipWhitespace() {
            while (!AtEnd() && (text[position] == ' ' || text[position] == '\t' || text[position] == '\n' || text[position] == '\r')) ++position;
        }

        void Literal(std::string_view word) {
            if (text.substr(position, word.size()) != word) Fail("expected a value");
            position += word.size();
        }

        Value Any(std::uint32_t depth) {
            if (depth > maxDepth) Fail("nested too deeply");
            if (AtEnd()) Fail("unexpected end of text");
            switch (text[position]) {
            case '{': return Object(depth);
            case '[': return Array(depth);
            case '"': return Value(String());
            case 't': Literal("true"); return Value(true);
            case 'f': Literal("false"); return Value(false);
            case 'n': Literal("null"); return Value();
            default: return Number();
            }
        }

        Value Object(std::uint32_t depth) {
            ++position;
            Value object = Value::MakeObject();
            SkipWhitespace();
            if (Consume('}')) return object;
            while (true) {
                SkipWhitespace();
                if (!Peek('"')) Fail("expected a member name in quotes");
                const std::string key = String();
                SkipWhitespace();
                if (!Consume(':')) Fail("expected ':' after a member name");
                SkipWhitespace();
                object.Set(key, Any(depth + 1));
                SkipWhitespace();
                if (Consume('}')) return object;
                if (!Consume(',')) Fail("expected ',' or '}'");
            }
        }

        Value Array(std::uint32_t depth) {
            ++position;
            Value array = Value::MakeArray();
            SkipWhitespace();
            if (Consume(']')) return array;
            while (true) {
                SkipWhitespace();
                array.Push(Any(depth + 1));
                SkipWhitespace();
                if (Consume(']')) return array;
                if (!Consume(',')) Fail("expected ',' or ']'");
            }
        }

        std::uint32_t Hex4() {
            if (position + 4 > text.size()) Fail("truncated \\u escape");
            std::uint32_t code = 0;
            const auto [end, error] = std::from_chars(text.data() + position, text.data() + position + 4, code, 16);
            if (error != std::errc() || end != text.data() + position + 4) Fail("malformed \\u escape");
            position += 4;
            return code;
        }

        std::string String() {
            ++position;
            std::string out;
            while (true) {
                if (AtEnd()) Fail("unterminated string");
                const char c = text[position];
                if (c == '"') {
                    ++position;
                    return out;
                }
                if (static_cast<std::uint8_t>(c) < 0x20) Fail("control character in a string");
                ++position;
                if (c != '\\') {
                    out += c;
                    continue;
                }
                if (AtEnd()) Fail("unterminated escape");
                const char escape = text[position++];
                switch (escape) {
                case '"': out += '"'; break;
                case '\\': out += '\\'; break;
                case '/': out += '/'; break;
                case 'b': out += '\b'; break;
                case 'f': out += '\f'; break;
                case 'n': out += '\n'; break;
                case 'r': out += '\r'; break;
                case 't': out += '\t'; break;
                case 'u': {
                    std::uint32_t code = Hex4();
                    if (code >= 0xD800 && code <= 0xDBFF) {
                        if (text.substr(position, 2) != "\\u") Fail("unpaired surrogate");
                        position += 2;
                        const std::uint32_t low = Hex4();
                        if (low < 0xDC00 || low > 0xDFFF) Fail("unpaired surrogate");
                        code = 0x10000 + ((code - 0xD800) << 10) + (low - 0xDC00);
                    } else if (code >= 0xDC00 && code <= 0xDFFF) {
                        Fail("unpaired surrogate");
                    }
                    AppendUtf8(out, code);
                    break;
                }
                default:
                    --position;
                    Fail("unknown escape");
                }
            }
        }

        Value Number() {
            const std::size_t start = position;
            Consume('-');
            if (AtEnd() || !IsDigit(text[position])) Fail("expected a value");
            if (!Consume('0')) {
                while (!AtEnd() && IsDigit(text[position])) ++position;
            }
            bool whole = true;
            if (Consume('.')) {
                whole = false;
                if (AtEnd() || !IsDigit(text[position])) Fail("expected a digit after the decimal point");
                while (!AtEnd() && IsDigit(text[position])) ++position;
            }
            if (Consume('e') || Consume('E')) {
                whole = false;
                if (!Consume('+')) Consume('-');
                if (AtEnd() || !IsDigit(text[position])) Fail("expected a digit in the exponent");
                while (!AtEnd() && IsDigit(text[position])) ++position;
            }
            const char* first = text.data() + start;
            const char* last = text.data() + position;
            if (whole) {
                std::int64_t integer = 0;
                const auto [end, error] = std::from_chars(first, last, integer);
                if (error == std::errc() && end == last) return Value(integer);
                // Beyond int64: read as a double below.
            }
            double number = 0;
            const auto [end, error] = std::from_chars(first, last, number);
            if (error == std::errc::result_out_of_range) {
                position = start;
                Fail("number out of range");
            }
            if (error != std::errc() || end != last) {
                position = start;
                Fail("malformed number");
            }
            return Value(number);
        }
    };

    std::logic_error WrongKind(ValueKind actual, std::string_view wanted) {
        return std::logic_error(std::format("Value is {}, not {}", KindName(actual), wanted));
    }
}

std::string_view Crafter::Debug::KindName(ValueKind kind) {
    switch (kind) {
    case ValueKind::Null: return "null";
    case ValueKind::Bool: return "boolean";
    case ValueKind::Integer: return "integer";
    case ValueKind::Number: return "number";
    case ValueKind::String: return "string";
    case ValueKind::Array: return "array";
    case ValueKind::Object: return "object";
    }
    return "unknown";
}

ParseError::ParseError(std::size_t offset, std::string_view what)
    : std::runtime_error(std::format("{} at offset {}", what, offset)), offset(offset) {}

Value Value::MakeArray(std::vector<Value> items) {
    Value value;
    value.kind = ValueKind::Array;
    value.items = std::move(items);
    return value;
}

Value Value::MakeObject() {
    Value value;
    value.kind = ValueKind::Object;
    return value;
}

bool Value::AsBool() const {
    if (kind != ValueKind::Bool) throw WrongKind(kind, "a boolean");
    return boolean;
}

std::int64_t Value::AsInteger() const {
    if (kind != ValueKind::Integer) throw WrongKind(kind, "an integer");
    return integer;
}

double Value::AsNumber() const {
    if (kind == ValueKind::Integer) return static_cast<double>(integer);
    if (kind != ValueKind::Number) throw WrongKind(kind, "a number");
    return number;
}

const std::string& Value::AsString() const {
    if (kind != ValueKind::String) throw WrongKind(kind, "a string");
    return text;
}

std::span<const Value> Value::Items() const {
    if (kind != ValueKind::Array) throw WrongKind(kind, "an array");
    return items;
}

std::span<Value> Value::Items() {
    if (kind != ValueKind::Array) throw WrongKind(kind, "an array");
    return items;
}

void Value::Push(Value item) {
    if (kind == ValueKind::Null) kind = ValueKind::Array;
    if (kind != ValueKind::Array) throw WrongKind(kind, "an array");
    items.push_back(std::move(item));
}

const Value* Value::Find(std::string_view key) const {
    if (kind != ValueKind::Object) return nullptr;
    const auto found = std::ranges::find(keys, key);
    return found == keys.end() ? nullptr : &values[static_cast<std::size_t>(found - keys.begin())];
}

Value* Value::Find(std::string_view key) {
    return const_cast<Value*>(std::as_const(*this).Find(key));
}

Value& Value::Set(std::string_view key, Value value) {
    if (kind == ValueKind::Null) kind = ValueKind::Object;
    if (kind != ValueKind::Object) throw WrongKind(kind, "an object");
    if (Value* existing = Find(key)) {
        *existing = std::move(value);
        return *existing;
    }
    keys.emplace_back(key);
    values.push_back(std::move(value));
    return values.back();
}

Value& Value::operator[](std::string_view key) {
    if (Value* existing = Find(key)) return *existing;
    return Set(key, Value());
}

bool Value::Erase(std::string_view key) {
    if (kind != ValueKind::Object) return false;
    const auto found = std::ranges::find(keys, key);
    if (found == keys.end()) return false;
    const auto index = found - keys.begin();
    keys.erase(found);
    values.erase(values.begin() + index);
    return true;
}

std::size_t Value::Size() const {
    if (kind == ValueKind::Array) return items.size();
    if (kind == ValueKind::Object) return keys.size();
    return 0;
}

Value Value::Parse(std::string_view json) {
    return Parser(json).Document();
}

std::string Value::Dump() const {
    std::string out;
    DumpTo(out, 0, 0);
    return out;
}

std::string Value::DumpPretty(std::uint32_t indent) const {
    std::string out;
    DumpTo(out, indent == 0 ? 1 : indent, 0);
    return out;
}

void Value::DumpTo(std::string& out, std::uint32_t indent, std::uint32_t depth) const {
    const auto newline = [&](std::uint32_t level) {
        if (indent == 0) return;
        out += '\n';
        out.append(static_cast<std::size_t>(indent) * level, ' ');
    };
    switch (kind) {
    case ValueKind::Null: out += "null"; return;
    case ValueKind::Bool: out += boolean ? "true" : "false"; return;
    case ValueKind::Integer: out += std::to_string(integer); return;
    case ValueKind::Number: AppendNumber(out, number); return;
    case ValueKind::String: AppendEscaped(out, text); return;
    case ValueKind::Array:
        if (items.empty()) {
            out += "[]";
            return;
        }
        out += '[';
        for (std::size_t i = 0; i < items.size(); ++i) {
            if (i > 0) out += ',';
            newline(depth + 1);
            items[i].DumpTo(out, indent, depth + 1);
        }
        newline(depth);
        out += ']';
        return;
    case ValueKind::Object:
        if (keys.empty()) {
            out += "{}";
            return;
        }
        out += '{';
        for (std::size_t i = 0; i < keys.size(); ++i) {
            if (i > 0) out += ',';
            newline(depth + 1);
            AppendEscaped(out, keys[i]);
            out += indent == 0 ? ":" : ": ";
            values[i].DumpTo(out, indent, depth + 1);
        }
        newline(depth);
        out += '}';
        return;
    }
}

namespace Crafter::Debug {
    bool operator==(const Value& a, const Value& b) {
        if (a.IsNumber() && b.IsNumber()) {
            if (a.kind == ValueKind::Integer && b.kind == ValueKind::Integer) return a.integer == b.integer;
            return a.AsNumber() == b.AsNumber();
        }
        if (a.kind != b.kind) return false;
        switch (a.kind) {
        case ValueKind::Null: return true;
        case ValueKind::Bool: return a.boolean == b.boolean;
        case ValueKind::String: return a.text == b.text;
        case ValueKind::Array: return a.items == b.items;
        case ValueKind::Object:
            if (a.keys.size() != b.keys.size()) return false;
            for (std::size_t i = 0; i < a.keys.size(); ++i) {
                const Value* other = b.Find(a.keys[i]);
                if (other == nullptr || !(a.values[i] == *other)) return false;
            }
            return true;
        default: return false;
        }
    }
}