integration/luprex/cpp/drv/readline.cpp

#include "readline.hpp"

#define MAXLINE 512

static CodepointString n_backspaces(int n) {
    CodepointString result(3 * n, 0);
    for (int i = 0; i < n; i++) {
        result[i*3 + 0] = '\b';
        result[i*3 + 1] = ' ';
        result[i*3 + 2] = '\b';
    }
    return result;
}

static int common_prefix_length(const CodepointString &a, const CodepointString &b) {
    int minlen = std::min(a.size(), b.size());
    for (int i = 0; i < minlen; i++) {
        if (a[i] != b[i]) return i;
    }
    return minlen;
}

static int buffer_codepoint_utf8(char32_t scp, char *buffer) {
    uint32_t cp = (uint32_t)scp;
    unsigned char *c = (unsigned char *)buffer;
    if (cp < 0) {
        return 0;
    }
    else if (cp <= 0x7F) {
        c[0] = cp;
        return 1;
    }
    else if (cp <= 0x7FF) {
        c[0] = (cp>>6)+192;
        c[1] = (cp&63)+128;
        return 2;
    }
    else if (cp <= 0xFFFF) {
        c[0] = (cp>>12)+224;
        c[1] = ((cp>>6)&63)+128;
        c[2] = (cp&63)+128;
        return 3;
    }
    else if (cp <= 0x10FFFF) {
        c[0] = (cp>>18)+240;
        c[1] = ((cp>>12)&63)+128;
        c[2] = ((cp>>6)&63)+128;
        c[3] = (cp&63)+128;
        return 4;
    } else {
        return 0;
    }
}

static int32_t read_codepoint_utf8(std::string_view &source) {
    size_t size = source.size();
    if (size == 0) return -1;

    const unsigned char *bytes = (const unsigned char *)source.data();
    int codepoint;
    size_t seqlen;
    if ((bytes[0] & 0x80) == 0x00) {
        // U+0000 to U+007F
        codepoint = (bytes[0] & 0x7F);
        seqlen = 1;
    } else if ((bytes[0] & 0xE0) == 0xC0) {
        // U+0080 to U+07FF
        codepoint = (bytes[0] & 0x1F);
        seqlen = 2;
    } else if ((bytes[0] & 0xF0) == 0xE0) {
        // U+0800 to U+FFFF
        codepoint = (bytes[0] & 0x0F);
        seqlen = 3;
    } else if ((bytes[0] & 0xF8) == 0xF0) {
        // U+10000 to U+10FFFF
        codepoint = (bytes[0] & 0x07);
        seqlen = 4;
    } else {
        // Bad character. Drop a byte and return invalid CP.
        source.remove_prefix(1);
        return -2;
    }

    if (seqlen > size) {
        return -1;
    }

    for (size_t i = 1; i < seqlen; ++i) {
        if ((bytes[i] & 0xC0) != 0x80) {
            // Bad character. Drop a byte and return invalid CP.
            source.remove_prefix(1);
            return -2;
        }
        codepoint = (codepoint << 6) | (bytes[i] & 0x3F);
    }

    if ((codepoint > 0x10FFFF) ||
        ((codepoint <= 0x007F) && (seqlen != 1)) ||
        ((codepoint >= 0x0080) && (codepoint <= 0x07FF) && (seqlen != 2)) ||
        ((codepoint >= 0x0800) && (codepoint <= 0xFFFF) && (seqlen != 3)) ||
        ((codepoint >= 0x10000) && (codepoint <= 0x1FFFFF) && (seqlen != 4))) {
        // Bad character. Drop a byte and return invalid CP.
        source.remove_prefix(1);
        return -2;
    }

    source.remove_prefix(seqlen);
    return codepoint;
}

ReadlineDevice::ReadlineDevice() {
    desired_prompt_ = CodepointString(1, '>');
}


void ReadlineDevice::set_print_callback(print_callback cb) {
    print_cb_ = cb;
}

void ReadlineDevice::erase_command() {
    int ccsize = current_prompt_.size() + current_command_.size();
    if (ccsize > 0) {
        print_cb_(n_backspaces(ccsize));
        current_prompt_.clear();
        current_command_.clear();
    }
}

void ReadlineDevice::echo_command() {
    // If the prompt has changed, erase everything and start over.
    if (desired_prompt_ != current_prompt_) {
        int ccsize = current_prompt_.size() + current_command_.size();
        print_cb_(n_backspaces(ccsize));
        print_cb_(desired_prompt_);
        current_command_.clear();
        current_prompt_ = desired_prompt_;
    }

    // Find out how much of the command matches.
    int match = common_prefix_length(current_command_, desired_command_);

    // Echo backspaces to remove the non-matching part.
    int remove = current_command_.size() - match;
    if (remove > 0) {
        print_cb_(n_backspaces(remove));
        current_command_ = current_command_.substr(0, match);
    }

    // Echo the new part.
    CodepointString newpart = desired_command_.substr(current_command_.size());
    if (!newpart.empty()) {
        print_cb_(newpart);
        current_command_ = desired_command_;
    }
}

CodepointString ReadlineDevice::putcode(char32_t c) {
    if ((c == '\n') && (readline_lastc_ == '\r')) {
        // Ignore newline immediately after carriage return.
        // Otherwise, crlf produces two newlines.
        return CodepointString();
    } else if ((c == '\r') || (c == '\n')) {
        CodepointString white(1, ' ');
        CodepointString newline(1, '\n');
        echo_command();
        print_cb_(white + newline);
        CodepointString result = desired_command_ + newline;
        desired_command_.clear();
        current_prompt_.clear();
        current_command_.clear();
        return result;
    } else if ((c == '\b') || (c == 127)) {
        int len = desired_command_.size();
        if (len > 0) {
            desired_command_ = desired_command_.substr(0, len-1);
        }
        echo_command();
        return CodepointString();
    } else if ((c >= 32)&&(c <= 0x10FFFF)) {
        int len = desired_command_.size();
        if (len < MAXLINE) {
            desired_command_ = desired_command_ + c;
        }
        echo_command();
        return CodepointString();
    }
    readline_lastc_ = c;
    return CodepointString();
}

void ReadlineDevice::print(const CodepointString &s) {
    if (!s.empty()) {
        erase_command();
        print_cb_(s);
        echo_command();
    }
}

std::string ReadlineDevice::to_utf8(const CodepointString &s) {
    std::string result(s.size() * 4, 0);
    char *buffer = &result[0];
    int len = 0;
    for (char32_t c : s) {
        int clen = buffer_codepoint_utf8(c, buffer + len);
        len += clen;
    }
    return result.substr(0, len);
}

CodepointString ReadlineDevice::from_utf8(std::string_view s, int *consumed) {
    std::string_view rest = s;
    CodepointString result(s.size(), 0);
    int len = 0;
    while (true) {
        int32_t c = read_codepoint_utf8(rest);
        if (c == -1) break; // EOF reached;
        if (c == -2) continue; // Filter out bad UTF8 but continue.
        result[len++] = (char32_t)c;
    }
    if (consumed != nullptr) {
        *consumed = s.size() - rest.size();
    }
    return result.substr(0, len);
}