libregexis024/src/libregexis024sol/command_expression.cpp

#include <libregexis024sol/special_terminals.h>

#include <libregexis024sol/utils.h>
#include <libregexis024sol/sol_misc_base.h>
#include <assert.h>
#include <memory>

namespace regexis024 {
    struct ParseCall{
        virtual ~ParseCall() = default;
        virtual std::unique_ptr<ParseCall> afterReceive(REGEX_IS024_MeaningContext& ctx) { assert(false); }
        virtual std::unique_ptr<ParseCall> firstTime(REGEX_IS024_MeaningContext& ctx) { assert(false); }
    };

    struct Top_ParseCall: public ParseCall{
        Command& res;
        explicit Top_ParseCall(Command &res) : res(res) {}
        std::unique_ptr<ParseCall> firstTime(REGEX_IS024_MeaningContext &ctx) override;
        std::unique_ptr<ParseCall> afterReceive(REGEX_IS024_MeaningContext &ctx) override;
    };

    struct Bracker_ParseCall: public ParseCall{
        std::vector<CommandArgument>& res;
        bool closingBraceEnded = false;
        explicit Bracker_ParseCall(std::vector<CommandArgument> &res) : res(res) {}
        std::unique_ptr<ParseCall> argReadProc(REGEX_IS024_MeaningContext& ctx);
        std::unique_ptr<ParseCall> firstTime(REGEX_IS024_MeaningContext &ctx) override;
        std::unique_ptr<ParseCall> afterReceive(REGEX_IS024_MeaningContext &ctx) override;
    };

#define call_ERROR_CHECK do { if (ctx.error) { return NULL; } } while (0)
#define call_THROW(str) do { report(ctx, "command expression: " str); return NULL; } while (0)

    std::unique_ptr<ParseCall> Top_ParseCall::firstTime(REGEX_IS024_MeaningContext &ctx) {
        assert(readChar(ctx) == U'!');
        int32_t ch = peep(ctx); call_ERROR_CHECK;
        if (ch == U'~'){
            /* I assume during construction I received reference to newly initialized struct */
            res.tilda = true;
            return NULL;
        }
        res.name = tryRead_REGEX024_name(ctx); call_ERROR_CHECK;
        if (res.name.empty())
            call_THROW("top lvl: no command name specified");
        ch = peep(ctx); call_ERROR_CHECK;
        if (ch == U';'){
            readChar(ctx);
            return NULL;
        }
        if (ch == U'{'){
            return std::make_unique<Bracker_ParseCall>(res.arguments);
        }
        call_THROW("top lvl: command call should be ended with ';' or '{...}'");
    }

    std::unique_ptr<ParseCall> Top_ParseCall::afterReceive(REGEX_IS024_MeaningContext &ctx) {
        return NULL;
    }

    std::unique_ptr<ParseCall> Bracker_ParseCall::firstTime(REGEX_IS024_MeaningContext &ctx) {
        assert(readChar(ctx) == U'{');
        return argReadProc(ctx);
    }

    std::unique_ptr<ParseCall> Bracker_ParseCall::afterReceive(REGEX_IS024_MeaningContext &ctx) {
        closingBraceEnded = true;
        return argReadProc(ctx);
    }

    std::unique_ptr<ParseCall> Bracker_ParseCall::argReadProc(REGEX_IS024_MeaningContext &ctx) {
        repeat:
        int32_t ch = peep(ctx); call_ERROR_CHECK;
        if (ch == U';'){
            res.emplace_back();
            readChar(ctx);
            closingBraceEnded = false;
            goto repeat;
        } else if (ch == U'}'){
            readChar(ctx);
            if (!closingBraceEnded){
                res.emplace_back();
            }
            return NULL;
        } else if (is_REGEX024_nameConstituent(ch)){
            res.emplace_back();
            res.back().is_empty = false;
            res.back().name = tryRead_REGEX024_name(ctx);
            int32_t eCh = peep(ctx); call_ERROR_CHECK;
            if (eCh == U';'){
                readChar(ctx);
                closingBraceEnded = false;
                goto repeat;
            } else if (eCh == U'{'){
                return std::make_unique<Bracker_ParseCall>(res.back().arguments);
            } else if (eCh == U'}'){
                readChar(ctx);
                return NULL;
            }
            call_THROW("brace lvl: argument ends with ';' or {...}");
        }
        call_THROW("brace lvl: argument starts with ';' or it's name");
    }

    Command command_expr_parse(REGEX_IS024_MeaningContext &ctx) {
        std::vector<std::unique_ptr<ParseCall>> callStack;
        Command res;
        callStack.push_back(std::make_unique<Top_ParseCall>(res));
        bool first_time = true;
        while (!callStack.empty()){
            if (ctx.error)
                return {};
            auto nxt = first_time ? callStack.back()->firstTime(ctx) : callStack.back()->afterReceive(ctx);
            if (nxt){
                callStack.push_back(std::move(nxt));
                first_time = true;
            } else {
                callStack.pop_back();
                first_time = false;
            }
        }
        return res;
    }

    const char* commands_for_codesets[] = {"word", "space", "digit", "variable", "any", "A", NULL};

    bool is_command_for_charset(const Command &cmd) {
        return !cmd.tilda && cmd.arguments.empty() && is_string_in_stringset(cmd.name.c_str(), commands_for_codesets);
    }

    void interpret_command_as_charset_giving(const CommonCodesets& cc, const Command &cmd, codeset_t& ret)
    {
        if (cmd.name == "word")
            ret = cc.word_constituents;
        else if (cmd.name == "space")
            ret = cc.spaces;
        else if (cmd.name == "digit")
            ret = cc.digits;
        else if (cmd.name == "variable")
            ret = cc.variable_constituents;
        else if (cmd.name == "any" || cmd.name == "A")
            ret = codeset_of_all;
        else
            assert(false);
    }
}