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

#include "table_stats.h"
#include "line_index.h"
#include <string.h>

int table_is_separator_row(const char *line, size_t len)
{
    int pipes = 0;
    int dashes = 0;
    for (size_t i = 0; i < len; i++) {
        char c = line[i];
        if (c == '|') pipes++;
        else if (c == '-' || c == ':' || c == ' ') dashes++;
        else if (c != '\r') return 0;
    }
    return pipes >= 1 && dashes > 0;
}

/* GFM: leading/trailing '|' are delimiters, not empty cells. */
int table_count_cols(const char *line, size_t len, int *has_pipe)
{
    if (has_pipe)
        *has_pipe = 0;

    while (len > 0 && line[len - 1] == '\r')
        len--;

    size_t start = 0;
    size_t end = len;
    while (start < end && (line[start] == ' ' || line[start] == '\t'))
        start++;
    if (start < end && line[start] == '|') {
        start++;
        if (has_pipe)
            *has_pipe = 1;
    }
    while (end > start && (line[end - 1] == ' ' || line[end - 1] == '\t'))
        end--;
    if (end > start && line[end - 1] == '|') {
        end--;
        if (has_pipe)
            *has_pipe = 1;
    }

    if (start >= end)
        return 1;

    int cols = 1;
    for (size_t i = start; i < end; i++) {
        if (line[i] == '|') {
            cols++;
            if (has_pipe)
                *has_pipe = 1;
        }
    }
    return cols > SASHIDOWN_MAX_COLS ? SASHIDOWN_MAX_COLS : cols;
}

static size_t te_line_end(sashidown_ctx_t *ctx, const char *input, size_t len, size_t pos)
{
    if (ctx && ctx->line_ix)
        return line_end_ix((line_index_t *)ctx->line_ix, input, len, pos);
    const char *hit = memchr(input + pos, '\n', len - pos);
    return hit ? (size_t)(hit - input) : len;
}

static size_t te_skip_line(sashidown_ctx_t *ctx, const char *input, size_t len, size_t pos)
{
    if (ctx && ctx->line_ix)
        return skip_line_ix((line_index_t *)ctx->line_ix, input, len, pos);
    pos = te_line_end(NULL, input, len, pos);
    if (pos < len)
        pos++;
    return pos;
}

static int line_has_pipe_raw(const char *line, size_t len)
{
    return memchr(line, '|', len) != NULL;
}

static int try_parse_table(sashidown_ctx_t *ctx, const char *input, size_t len,
                           size_t line_start, table_entry_t *e)
{
    size_t le = te_line_end(ctx, input, len, line_start);
    int has_pipe = 0;
    int cols = table_count_cols(input + line_start, le - line_start, &has_pipe);
    if (!has_pipe)
        return 0;

    size_t sep = te_skip_line(ctx, input, len, line_start);
    if (sep >= len) {
        if (ctx && ctx->line_ix)
            line_index_seek((line_index_t *)ctx->line_ix, line_start);
        return 0;
    }
    size_t sep_le = te_line_end(ctx, input, len, sep);
    if (!table_is_separator_row(input + sep, sep_le - sep)) {
        if (ctx && ctx->line_ix)
            line_index_seek((line_index_t *)ctx->line_ix, line_start);
        return 0;
    }

    size_t pos = te_skip_line(ctx, input, len, sep);

    e->start = line_start;
    e->body_start = pos;
    e->cols = cols; /* GFM: header determines column count */
    e->body_rows = 0;
    e->valid = 1;

    while (pos < len) {
        le = te_line_end(ctx, input, len, pos);
        size_t line_len = le - pos;
        if (!line_has_pipe_raw(input + pos, line_len))
            break;
        if (table_is_separator_row(input + pos, line_len))
            break;

        e->body_rows++;
        pos = te_skip_line(ctx, input, len, pos);
    }

    e->end = pos;
    return 1;
}

int table_try_parse_at(sashidown_ctx_t *ctx, const char *input, size_t len,
                       size_t pos, table_entry_t *e)
{
    e->start = 0;
    e->end = 0;
    e->body_start = 0;
    e->cols = 0;
    e->body_rows = 0;
    e->valid = 0;
    return try_parse_table(ctx, input, len, pos, e);
}