snac2/format.c

/* snac - A simple, minimalistic ActivityPub instance */
/* copyright (c) 2022 grunfink - MIT license */

#include "xs.h"
#include "xs_regex.h"

#include "snac.h"

/* emoticons, people laughing and such */

struct {
    const char *key;
    const char *value;
} smileys[] = {
    { ":-)",        "🙂" },
    { ":-D",        "😀" },
    { "X-D",        "😆" },
    { ";-)",        "😉" },
    { "B-)",        "😎" },
    { ":-(",        "😞" },
    { ":-*",        "😘" },
    { ":-/",        "😕" },
    { "8-o",        "😲" },
    { "%-)",        "🤪" },
    { ":_(",        "😢" },
    { ":-|",        "😐" },
    { "<3",         "&#128147;" },
    { ":facepalm:", "&#129318;" },
    { ":shrug:",    "&#129335;" },
    { ":eyeroll:",  "&#128580;" },
    { ":beer:",     "&#127866;" },
    { ":beers:",    "&#127867;" },
    { ":munch:",    "&#128561;" },
    { ":thumb:",    "&#128077;" },
    { NULL,         NULL }
};


static d_char *format_line(const char *line)
/* formats a line */
{
    d_char *s = xs_str_new(NULL);
    char *p, *v;

    /* split by markup */
    xs *sm = xs_regex_split(line,
        "(`[^`]+`|\\*\\*?[^\\*]+\\*?\\*|https?:/" "/[^[:space:]]+)");
    int n = 0;

    p = sm;
    while (xs_list_iter(&p, &v)) {
        if ((n & 0x1)) {
            /* markup */
            if (xs_startswith(v, "`")) {
                xs *s1 = xs_crop(xs_dup(v), 1, -1);
                xs *s2 = xs_fmt("<code>%s</code>", s1);
                s = xs_str_cat(s, s2);
            }
            else
            if (xs_startswith(v, "**")) {
                xs *s1 = xs_crop(xs_dup(v), 2, -2);
                xs *s2 = xs_fmt("<b>%s</b>", s1);
                s = xs_str_cat(s, s2);
            }
            else
            if (xs_startswith(v, "*")) {
                xs *s1 = xs_crop(xs_dup(v), 1, -1);
                xs *s2 = xs_fmt("<i>%s</i>", s1);
                s = xs_str_cat(s, s2);
            }
            else
            if (xs_startswith(v, "http")) {
                xs *s1 = xs_fmt("<a href=\"%s\" target=\"_blank\">%s</a>", v, v);
                s = xs_str_cat(s, s1);
            }
            else
                s = xs_str_cat(s, v);
        }
        else
            /* surrounded text, copy directly */
            s = xs_str_cat(s, v);

        n++;
    }

    return s;
}


d_char *not_really_markdown(const char *content)
/* formats a content using some Markdown rules */
{
    d_char *s = xs_str_new(NULL);
    int in_pre = 0;
    int in_blq = 0;
    xs *list;
    char *p, *v;

    /* work by lines */
    p = list = xs_split(content, "\n");

    while (xs_list_iter(&p, &v)) {
        xs *ss = NULL;

        if (strcmp(v, "```") == 0) {
            if (!in_pre)
                s = xs_str_cat(s, "<pre>");
            else
                s = xs_str_cat(s, "</pre>");

            in_pre = !in_pre;
            continue;
        }

        if (in_pre)
            ss = xs_dup(v);
        else
            ss = xs_strip(format_line(v));

        if (xs_startswith(ss, ">")) {
            /* delete the > and subsequent spaces */
            ss = xs_strip(xs_crop(ss, 1, 0));

            if (!in_blq) {
                s = xs_str_cat(s, "<blockquote>");
                in_blq = 1;
            }

            s = xs_str_cat(s, ss);
            s = xs_str_cat(s, "<br>");

            continue;
        }

        if (in_blq) {
            s = xs_str_cat(s, "</blockquote>");
            in_blq = 0;
        }

        s = xs_str_cat(s, ss);
        s = xs_str_cat(s, "<br>");
    }

    if (in_blq)
        s = xs_str_cat(s, "</blockquote>");
    if (in_pre)
        s = xs_str_cat(s, "</pre>");

    /* some beauty fixes */
    s = xs_replace_i(s, "<br><br><blockquote>", "<br><blockquote>");
    s = xs_replace_i(s, "</blockquote><br>", "</blockquote>");
    s = xs_replace_i(s, "</pre><br>", "</pre>");

    {
        /* traditional emoticons */
        int n;

        for (n = 0; smileys[n].key; n++)
            s = xs_replace_i(s, smileys[n].key, smileys[n].value);
    }

    return s;
}


const char *valid_tags[] = {
    "a", "p", "br", "br/", "blockquote", "ul", "li",
    "span", "i", "b", "pre", "code", "em", "strong", NULL
};

d_char *sanitize(const char *content)
/* cleans dangerous HTML output */
{
    d_char *s = xs_str_new(NULL);
    xs *sl;
    int n = 0;
    char *p, *v;

    sl = xs_regex_split(content, "</?[^>]+>");

    p = sl;

    while (xs_list_iter(&p, &v)) {
        if (n & 0x1) {
            xs *s1  = xs_strip(xs_crop(xs_dup(v), v[1] == '/' ? 2 : 1, -1));
            xs *l1  = xs_split_n(s1, " ", 1);
            xs *tag = xs_tolower(xs_dup(xs_list_get(l1, 0)));
            xs *s2  = NULL;
            int i;

            /* check if it's one of the valid tags */
            for (i = 0; valid_tags[i]; i++) {
                if (strcmp(tag, valid_tags[i]) == 0)
                    break;
            }

            if (valid_tags[i]) {
                /* accepted tag: rebuild it with only the accepted elements */
                xs *el = xs_regex_match(v, "(href|rel|class|target)=\"[^\"]*\"");
                xs *s3 = xs_join(el, " ");

                s2 = xs_fmt("<%s%s %s>", v[1] == '/' ? "/" : "", tag, s3);
            }
            else {
                /* bad tag: escape it */
                s2 = xs_replace(v, "<", "&lt;");
            }

            s = xs_str_cat(s, s2);
        }
        else {
            /* non-tag */
            s = xs_str_cat(s, v);
        }

        n++;
    }

    return s;
}