CST-agent/vendor/sashidown/src/block_scan.c

#include "block_scan.h"
#include <string.h>

int block_count_heading(const char *input, size_t len, size_t pos, size_t *content_start)
{
    size_t i = pos;
    int level = 0;
    while (i < len && input[i] == '#') {
        level++;
        i++;
    }
    if (level == 0 || level > 6) return 0;
    if (i < len && input[i] != ' ' && input[i] != '\n') return 0;
    while (i < len && input[i] == ' ') i++;
    *content_start = i;
    return level;
}

int block_is_fence(const char *input, size_t len, size_t pos,
                   int *fence_len, char *fence_char)
{
    size_t i = pos;
    while (i < len && (input[i] == ' ' || input[i] == '\t')) i++;
    if (i + 2 >= len) return 0;
    char ch = input[i];
    if (ch != '`' && ch != '~') return 0;
    if (input[i + 1] != ch || input[i + 2] != ch) return 0;
    int fl = 3;
    i += 3;
    while (i < len && input[i] == ch) { fl++; i++; }
    *fence_len = fl;
    if (fence_char)
        *fence_char = ch;
    return 1;
}

void block_fence_info_range(const char *input, size_t len, size_t pos,
                            int fence_len, char fence_char,
                            size_t *info_start, size_t *info_end)
{
    size_t i = pos;
    while (i < len && (input[i] == ' ' || input[i] == '\t')) i++;
    i += (size_t)fence_len;
    (void)fence_char;
    while (i < len && (input[i] == ' ' || input[i] == '\t')) i++;
    *info_start = i;
    size_t end = i;
    while (end < len && input[end] != '\n' && input[end] != '\r')
        end++;
    while (end > *info_start &&
           (input[end - 1] == ' ' || input[end - 1] == '\t'))
        end--;
    *info_end = end;
}

int block_is_thematic_break(const char *input, size_t pos, size_t le)
{
    size_t i = pos;
    while (i < le && (input[i] == ' ' || input[i] == '\t')) i++;
    if (i >= le) return 0;

    char marker = input[i];
    if (marker != '*' && marker != '-' && marker != '_') return 0;

    int count = 0;
    for (; i < le; i++) {
        char c = input[i];
        if (c == marker) {
            count++;
        } else if (c == ' ' || c == '\t') {
            continue;
        } else if (c == '\r') {
            continue;
        } else {
            return 0;
        }
    }
    return count >= 3;
}

int block_is_setext_underline(const char *input, size_t pos, size_t le, int *level)
{
    size_t i = pos;
    while (i < le && (input[i] == ' ' || input[i] == '\t')) i++;
    if (i >= le) return 0;

    char marker = input[i];
    if (marker != '=' && marker != '-') return 0;

    size_t start = i;
    while (i < le && input[i] == marker) i++;
    if (i == start) return 0;
    while (i < le && (input[i] == ' ' || input[i] == '\t' || input[i] == '\r'))
        i++;
    if (i < le) return 0;

    /* Thematic-break-looking '---' is still a valid setext underline when
     * paired with a preceding text line (caller decides context). */
    if (level)
        *level = (marker == '=') ? 1 : 2;
    return 1;
}

int block_is_blockquote(const char *input, size_t len, size_t pos, size_t *content_start)
{
    size_t i = pos;
    while (i < len && (input[i] == ' ' || input[i] == '\t')) i++;
    if (i >= len || input[i] != '>') return 0;
    i++;
    if (i < len && input[i] == ' ') i++;
    *content_start = i;
    return 1;
}

int block_is_ul(const char *input, size_t len, size_t pos, size_t *content_start)
{
    size_t i = pos;
    while (i < len && (input[i] == ' ' || input[i] == '\t')) i++;
    if (i >= len) return 0;
    if (input[i] != '-' && input[i] != '*' && input[i] != '+') return 0;
    if (i + 1 < len && input[i + 1] != ' ') return 0;
    i += 2;
    *content_start = i;
    return 1;
}

int block_parse_task_marker(const char *input, size_t len, size_t content_start,
                            int *task, size_t *text_start)
{
    size_t i = content_start;
    if (i + 3 > len) return 0;
    if (input[i] != '[') return 0;
    char mark = input[i + 1];
    if (input[i + 2] != ']') return 0;
    if (mark != ' ' && mark != 'x' && mark != 'X') return 0;
    i += 3;
    if (i < len && input[i] == ' ') i++;
    *task = (mark == ' ') ? 1 : 2;
    *text_start = i;
    return 1;
}

int block_is_ol(const char *input, size_t len, size_t pos, size_t *content_start, int *num)
{
    size_t i = pos;
    while (i < len && (input[i] == ' ' || input[i] == '\t')) i++;
    size_t num_start = i;
    { unsigned char c = (unsigned char)input[i];
      if (i >= len || c < '0' || c > '9') return 0; }
    while (i < len) {
        unsigned char c = (unsigned char)input[i];
        if (c < '0' || c > '9') break;
        i++;
    }
    if (i >= len || input[i] != '.') return 0;
    if (i + 1 < len && input[i + 1] != ' ') return 0;
    *num = 0;
    for (size_t j = num_start; j < i; j++)
        *num = *num * 10 + (input[j] - '0');
    i += 2;
    *content_start = i;
    return 1;
}

int block_line_starts_table(const char *input, size_t len, size_t pos)
{
    size_t i = pos;
    while (i < len && (input[i] == ' ' || input[i] == '\t')) i++;
    if (i >= len) return 0;
    /* GFM: leading '|' OR a pipe somewhere on the line with a separator below.
     * Cheap reject: must contain '|' before newline. */
    const char *nl = memchr(input + i, '\n', len - i);
    size_t line_len = nl ? (size_t)(nl - (input + i)) : (len - i);
    return memchr(input + i, '|', line_len) != NULL;
}

int block_is_latex_line(sashidown_ctx_t *ctx, const char *input, size_t len,
                        size_t pos, size_t *content_start, size_t *content_end)
{
    size_t i = pos;
    while (i < len && (input[i] == ' ' || input[i] == '\t')) i++;
    if (i + 1 >= len || input[i] != '$' || input[i + 1] != '$') return 0;
    i += 2;
    *content_start = i;
    size_t le = block_line_end(ctx, input, len, pos);
    if (i + 1 < le && input[le - 2] == '$' && input[le - 1] == '$') {
        *content_end = le - 2;
        return 1;
    }
    return 0;
}

int block_try_setext(sashidown_ctx_t *ctx, const char *input, size_t len,
                     size_t pos, size_t le, size_t stop, int is_final,
                     line_info_t *info, size_t *end_pos)
{
    if (pos == le || (pos + 1 == le && input[pos] == '\r'))
        return 0;

    /* Next line starts after the newline at `le` (line_end points at '\n'). */
    if (le >= len || input[le] != '\n')
        return is_final ? 0 : -1;

    size_t next = le + 1;
    if (next >= len)
        return is_final ? 0 : -1;
    if (next >= stop && !is_final)
        return -1;

    /* Cheap reject before full underline scan. */
    size_t u = next;
    while (u < len && (input[u] == ' ' || input[u] == '\t'))
        u++;
    if (u >= len || (input[u] != '=' && input[u] != '-'))
        return 0;

    size_t nle = block_line_end(ctx, input, len, next);
    int level = 0;
    if (!block_is_setext_underline(input, next, nle, &level))
        return 0;

    info->kind = LK_HEADING;
    info->heading_level = level;
    info->content_start = pos;
    while (info->content_start < le &&
           (input[info->content_start] == ' ' || input[info->content_start] == '\t'))
        info->content_start++;
    info->content_end = le;
    while (info->content_end > info->content_start &&
           (input[info->content_end - 1] == ' ' ||
            input[info->content_end - 1] == '\t' ||
            input[info->content_end - 1] == '\r'))
        info->content_end--;
    info->fence_len = 0;
    info->fence_char = 0;
    info->info_start = 0;
    info->info_end = 0;
    info->task = 0;
    *end_pos = block_skip_line(ctx, input, len, next);
    return 1;
}

line_info_t block_classify_line(sashidown_ctx_t *ctx, const char *input, size_t len,
                                size_t pos, size_t le)
{
    line_info_t info;
    memset(&info, 0, sizeof(info));
    info.kind = LK_PARA;
    info.content_start = pos;
    info.content_end = le;

    if (pos == le || (pos + 1 == le && input[pos] == '\r')) {
        info.kind = LK_EMPTY;
        return info;
    }

    size_t i = pos;
    while (i < len && (input[i] == ' ' || input[i] == '\t')) i++;
    unsigned char first = (i < len) ? (unsigned char)input[i] : 0;

    /* Thematic breaks only possible for - / * / _ lines. */
    if ((first == '*' || first == '-' || first == '_') &&
        block_is_thematic_break(input, pos, le)) {
        info.kind = LK_HR;
        return info;
    }

    switch (first) {
    case '#':
        info.heading_level = block_count_heading(input, len, pos, &info.content_start);
        if (info.heading_level > 0) {
            info.kind = LK_HEADING;
            info.content_end = le;
            return info;
        }
        break;
    case '`':
    case '~':
        if (block_is_fence(input, len, pos, &info.fence_len, &info.fence_char)) {
            info.kind = LK_FENCE;
            block_fence_info_range(input, len, pos, info.fence_len, info.fence_char,
                                   &info.info_start, &info.info_end);
            return info;
        }
        break;
    case '>':
        if (block_is_blockquote(input, len, pos, &info.content_start)) {
            info.kind = LK_BQ;
            return info;
        }
        break;
    case '-': case '*': case '+':
        if (block_is_ul(input, len, pos, &info.content_start)) {
            info.kind = LK_UL;
            size_t text = info.content_start;
            int task = 0;
            if (block_parse_task_marker(input, le, info.content_start, &task, &text)) {
                info.task = task;
                info.content_start = text;
            }
            return info;
        }
        break;
    case '0': case '1': case '2': case '3': case '4':
    case '5': case '6': case '7': case '8': case '9':
    {
        int ol_num = 0;
        if (block_is_ol(input, len, pos, &info.content_start, &ol_num)) {
            info.kind = LK_OL;
            return info;
        }
        break;
    }
    case '$':
    {
        size_t cs, ce;
        if (i + 1 < len && input[i] == '$' && input[i + 1] == '$') {
            if (i == pos) {
                info.kind = LK_LATEX_BLOCK;
                return info;
            }
            if (block_is_latex_line(ctx, input, len, pos, &cs, &ce)) {
                info.kind = LK_LATEX_LINE;
                info.content_start = cs;
                info.content_end = ce;
                return info;
            }
        }
        break;
    }
    default:
        break;
    }

    info.kind = LK_PARA;
    return info;
}