libpurple/purplemarkup.c

/*
 * Purple - Internet Messaging Library
 * Copyright (C) Pidgin Developers <devel@pidgin.im>
 *
 * Purple is the legal property of its developers, whose names are too numerous
 * to list here. Please refer to the COPYRIGHT file distributed with this
 * source distribution.
 *
 * This library is free software; you can redistribute it and/or modify it
 * under the terms of the GNU General Public License as published by the Free
 * Software Foundation; either version 2 of the License, or (at your option)
 * any later version.
 *
 * This library is distributed in the hope that it will be useful, but WITHOUT
 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
 * FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License for
 * more details.
 *
 * You should have received a copy of the GNU General Public License along with
 * this library; if not, see <https://www.gnu.org/licenses/>.
 */

#include "purplemarkup.h"

#include "util.h"

static const char *
purple_markup_unescape_entity(const char *text, int *length)
{
	const char *pln;
	int len;

	if (!text || *text != '&')
		return NULL;

#define IS_ENTITY(s)  (!g_ascii_strncasecmp(text, s, (len = sizeof(s) - 1)))

	if(IS_ENTITY("&amp;"))
		pln = "&";
	else if(IS_ENTITY("&lt;"))
		pln = "<";
	else if(IS_ENTITY("&gt;"))
		pln = ">";
	else if(IS_ENTITY("&nbsp;"))
		pln = " ";
	else if(IS_ENTITY("&copy;"))
		pln = "\302\251";      /* or use g_unichar_to_utf8(0xa9); */
	else if(IS_ENTITY("&quot;"))
		pln = "\"";
	else if(IS_ENTITY("&reg;"))
		pln = "\302\256";      /* or use g_unichar_to_utf8(0xae); */
	else if(IS_ENTITY("&apos;"))
		pln = "\'";
	else if(text[1] == '#' && (g_ascii_isxdigit(text[2]) || text[2] == 'x')) {
		static char buf[7];
		const char *start = text + 2;
		char *end;
		guint64 pound;
		int base = 10;
		int buflen;

		if (*start == 'x') {
			base = 16;
			start++;
		}

		pound = g_ascii_strtoull(start, &end, base);
		if (pound == 0 || pound > INT_MAX || *end != ';') {
			return NULL;
		}

		len = (end - text) + 1;

		buflen = g_unichar_to_utf8((gunichar)pound, buf);
		buf[buflen] = '\0';
		pln = buf;
	}
	else
		return NULL;

	if (length)
		*length = len;
	return pln;
}

/* The following are probably reasonable changes:
 * - \n should be converted to a normal space
 * - in addition to <br>, <p> and <div> etc. should also be converted into \n
 * - We want to turn </td>#whitespace<td> sequences into a single tab
 * - We want to turn </tr>#whitespace<tr> sequences into a single \n
 * - <script>...</script> and <style>...</style> should be completely removed
 */

char *
purple_markup_strip_html(const char *str)
{
	int i, j, k, entlen;
	gboolean visible = TRUE;
	gboolean closing_td_p = FALSE;
	gchar *str2;
	const gchar *cdata_close_tag = NULL, *ent;
	gchar *href = NULL;
	int href_st = 0;

	if(!str)
		return NULL;

	str2 = g_strdup(str);

	for (i = 0, j = 0; str2[i]; i++)
	{
		if (str2[i] == '<')
		{
			if (cdata_close_tag)
			{
				/* Note: Don't even assume any other tag is a tag in CDATA */
				if (g_ascii_strncasecmp(str2 + i, cdata_close_tag,
						strlen(cdata_close_tag)) == 0)
				{
					i += strlen(cdata_close_tag) - 1;
					cdata_close_tag = NULL;
				}
				continue;
			}
			else if (g_ascii_strncasecmp(str2 + i, "<td", 3) == 0 && closing_td_p)
			{
				str2[j++] = '\t';
				visible = TRUE;
			}
			else if (g_ascii_strncasecmp(str2 + i, "</td>", 5) == 0)
			{
				closing_td_p = TRUE;
				visible = FALSE;
			}
			else
			{
				closing_td_p = FALSE;
				visible = TRUE;
			}

			k = i + 1;

			if(g_ascii_isspace(str2[k]))
				visible = TRUE;
			else if (str2[k])
			{
				/* Scan until we end the tag either implicitly (closed start
				 * tag) or explicitly, using a sloppy method (i.e., < or >
				 * inside quoted attributes will screw us up)
				 */
				while (str2[k] && str2[k] != '<' && str2[k] != '>')
				{
					k++;
				}

				/* If we've got an <a> tag with an href, save the address
				 * to print later. */
				if (g_ascii_strncasecmp(str2 + i, "<a", 2) == 0 &&
				    g_ascii_isspace(str2[i+2]))
				{
					int st; /* start of href, inclusive [ */
					int end; /* end of href, exclusive ) */
					char delim = ' ';
					/* Find start of href */
					for (st = i + 3; st < k; st++)
					{
						if (g_ascii_strncasecmp(str2+st, "href=", 5) == 0)
						{
							st += 5;
							if (str2[st] == '"' || str2[st] == '\'')
							{
								delim = str2[st];
								st++;
							}
							break;
						}
					}
					/* find end of address */
					for (end = st; end < k && str2[end] != delim; end++)
					{
						/* All the work is done in the loop construct above. */
					}

					/* If there's an address, save it.  If there was
					 * already one saved, kill it. */
					if (st < k)
					{
						char *tmp;
						g_free(href);
						tmp = g_strndup(str2 + st, end - st);
						href = purple_unescape_html(tmp);
						g_free(tmp);
						href_st = j;
					}
				}

				/* Replace </a> with an ascii representation of the
				 * address the link was pointing to. */
				else if (href != NULL && g_ascii_strncasecmp(str2 + i, "</a>", 4) == 0)
				{
					size_t hrlen = strlen(href);

					/* Only insert the href if it's different from the CDATA. */
					if ((hrlen != (gsize)(j - href_st) ||
					     strncmp(str2 + href_st, href, hrlen)) &&
					    (hrlen != (gsize)(j - href_st + 7) || /* 7 == strlen("http://") */
					     strncmp(str2 + href_st, href + 7, hrlen - 7)))
					{
						str2[j++] = ' ';
						str2[j++] = '(';
						memmove(str2 + j, href, hrlen);
						j += hrlen;
						str2[j++] = ')';
						g_free(href);
						href = NULL;
					}
				}

				/* Check for tags which should be mapped to newline (but ignore some of
				 * the tags at the beginning of the text) */
				else if ((j && (g_ascii_strncasecmp(str2 + i, "<p>", 3) == 0
				              || g_ascii_strncasecmp(str2 + i, "<tr", 3) == 0
				              || g_ascii_strncasecmp(str2 + i, "<hr", 3) == 0
				              || g_ascii_strncasecmp(str2 + i, "<li", 3) == 0
				              || g_ascii_strncasecmp(str2 + i, "<div", 4) == 0))
				 || g_ascii_strncasecmp(str2 + i, "<br", 3) == 0
				 || g_ascii_strncasecmp(str2 + i, "</table>", 8) == 0)
				{
					str2[j++] = '\n';
				}
				/* Check for tags which begin CDATA and need to be closed */
				else if (g_ascii_strncasecmp(str2 + i, "<script", 7) == 0)
				{
					cdata_close_tag = "</script>";
				}
				else if (g_ascii_strncasecmp(str2 + i, "<style", 6) == 0)
				{
					cdata_close_tag = "</style>";
				}
				/* Update the index and continue checking after the tag */
				i = (str2[k] == '<' || str2[k] == '\0')? k - 1: k;
				continue;
			}
		}
		else if (cdata_close_tag)
		{
			continue;
		}
		else if (!g_ascii_isspace(str2[i]))
		{
			visible = TRUE;
		}

		if (str2[i] == '&' && (ent = purple_markup_unescape_entity(str2 + i, &entlen)) != NULL)
		{
			while (*ent)
				str2[j++] = *ent++;
			i += entlen - 1;
			continue;
		}

		if (visible)
			str2[j++] = g_ascii_isspace(str2[i])? ' ': str2[i];
	}

	g_free(href);

	str2[j] = '\0';

	return str2;
}

static gboolean
badchar(char c)
{
	switch (c) {
	case ' ':
	case ',':
	case '\0':
	case '\n':
	case '\r':
	case '<':
	case '>':
	case '"':
		return TRUE;
	default:
		return FALSE;
	}
}

static gboolean
badentity(const char *c)
{
	if (!g_ascii_strncasecmp(c, "&lt;", 4) ||
		!g_ascii_strncasecmp(c, "&gt;", 4) ||
		!g_ascii_strncasecmp(c, "&quot;", 6)) {
		return TRUE;
	}
	return FALSE;
}

static const char *
process_link(GString *ret,
		const char *start, const char *c,
		int matchlen,
		const char *urlprefix,
		int inside_paren)
{
	char *url_buf = NULL;
	const char *t;

	for (t = c;; t++) {
		if (!badchar(*t) && !badentity(t))
			continue;

		if (t - c == matchlen)
			break;

		if (*t == ',' && *(t + 1) != ' ') {
			continue;
		}

		if (t > start && *(t - 1) == '.')
			t--;
		if (t > start && *(t - 1) == ')' && inside_paren > 0)
			t--;

		url_buf = g_strndup(c, t - c);
		g_string_append_printf(ret, "<a href=\"%s%s\">%s</a>",
				urlprefix,
				url_buf, url_buf);
		g_free(url_buf);
		return t;
	}

	return c;
}

char *
purple_markup_linkify(const char *text)
{
	const char *c, *t, *q = NULL;
	char *tmpurlbuf, *url_buf;
	gunichar g;
	gboolean inside_html = FALSE;
	int inside_paren = 0;
	GString *ret;

	if (text == NULL)
		return NULL;

	ret = g_string_new("");

	c = text;
	while (*c) {

		if(*c == '(' && !inside_html) {
			inside_paren++;
			ret = g_string_append_c(ret, *c);
			c++;
		}

		if(inside_html) {
			if(*c == '>') {
				inside_html = FALSE;
			} else if(!q && (*c == '\"' || *c == '\'')) {
				q = c;
			} else if(q) {
				if(*c == *q)
					q = NULL;
			}
		} else if(*c == '<') {
			inside_html = TRUE;
			if (!g_ascii_strncasecmp(c, "<a", 2)) {
				while (1) {
					if (!g_ascii_strncasecmp(c, "/a>", 3)) {
						inside_html = FALSE;
						break;
					}
					ret = g_string_append_c(ret, *c);
					c++;
					if (!(*c))
						break;
				}
			}
		} else if (!g_ascii_strncasecmp(c, "http://", 7)) {
			c = process_link(ret, text, c, 7, "", inside_paren);
		} else if (!g_ascii_strncasecmp(c, "https://", 8)) {
			c = process_link(ret, text, c, 8, "", inside_paren);
		} else if (!g_ascii_strncasecmp(c, "ftp://", 6)) {
			c = process_link(ret, text, c, 6, "", inside_paren);
		} else if (!g_ascii_strncasecmp(c, "sftp://", 7)) {
			c = process_link(ret, text, c, 7, "", inside_paren);
		} else if (!g_ascii_strncasecmp(c, "file://", 7)) {
			c = process_link(ret, text, c, 7, "", inside_paren);
		} else if (!g_ascii_strncasecmp(c, "www.", 4) && c[4] != '.' && (c == text || badchar(c[-1]) || badentity(c-1))) {
			c = process_link(ret, text, c, 4, "http://", inside_paren);
		} else if (!g_ascii_strncasecmp(c, "ftp.", 4) && c[4] != '.' && (c == text || badchar(c[-1]) || badentity(c-1))) {
			c = process_link(ret, text, c, 4, "ftp://", inside_paren);
		} else if (!g_ascii_strncasecmp(c, "xmpp:", 5) && (c == text || badchar(c[-1]) || badentity(c-1))) {
			c = process_link(ret, text, c, 5, "", inside_paren);
		} else if (!g_ascii_strncasecmp(c, "mailto:", 7)) {
			t = c;
			while (1) {
				if (badchar(*t) || badentity(t)) {
					char *d;
					if (t - c == 7) {
						break;
					}
					if (t > text && *(t - 1) == '.')
						t--;
					if ((d = strstr(c + 7, "?")) != NULL && d < t)
						url_buf = g_strndup(c + 7, d - c - 7);
					else
						url_buf = g_strndup(c + 7, t - c - 7);
					if (!purple_email_is_valid(url_buf)) {
						g_free(url_buf);
						break;
					}
					g_free(url_buf);
					url_buf = g_strndup(c, t - c);
					tmpurlbuf = purple_unescape_html(url_buf);
					g_string_append_printf(ret, "<a href=\"%s\">%s</a>",
							  tmpurlbuf, url_buf);
					g_free(url_buf);
					g_free(tmpurlbuf);
					c = t;
					break;
				}
				t++;
			}
		} else if (c != text && (*c == '@')) {
			int flag;
			GString *gurl_buf = NULL;
			const char illegal_chars[] = "!@#$%^&*()[]{}/|\\<>\":;\r\n \0";

			if (strchr(illegal_chars,*(c - 1)) || strchr(illegal_chars, *(c + 1)))
				flag = 0;
			else {
				flag = 1;
				gurl_buf = g_string_new("");
			}

			t = c;
			while (flag) {
				/* iterate backwards grabbing the local part of an email address */
				g = g_utf8_get_char(t);
				if (badchar(*t) || (g >= 127) || (*t == '(') ||
					((*t == ';') && ((t > (text+2) && (!g_ascii_strncasecmp(t - 3, "&lt;", 4) ||
				                                       !g_ascii_strncasecmp(t - 3, "&gt;", 4))) ||
				                     (t > (text+4) && (!g_ascii_strncasecmp(t - 5, "&quot;", 6)))))) {
					/* local part will already be part of ret, strip it out */
					ret = g_string_truncate(ret, ret->len - (c - t));
					ret = g_string_append_unichar(ret, g);
					break;
				} else {
					g_string_prepend_unichar(gurl_buf, g);
					t = g_utf8_find_prev_char(text, t);
					if (t < text) {
						ret = g_string_assign(ret, "");
						break;
					}
				}
			}

			t = g_utf8_find_next_char(c, NULL);

			while (flag) {
				/* iterate forwards grabbing the domain part of an email address */
				g = g_utf8_get_char(t);
				if (badchar(*t) || (g >= 127) || (*t == ')') || badentity(t)) {
					char *d;

					url_buf = g_string_free_and_steal(gurl_buf);
					gurl_buf = NULL;

					/* strip off trailing periods */
					if (*url_buf) {
						for (d = url_buf + strlen(url_buf) - 1; *d == '.'; d--, t--)
							*d = '\0';
					}

					tmpurlbuf = purple_unescape_html(url_buf);
					if (purple_email_is_valid(tmpurlbuf)) {
						g_string_append_printf(ret, "<a href=\"mailto:%s\">%s</a>",
								tmpurlbuf, url_buf);
					} else {
						g_string_append(ret, url_buf);
					}
					g_free(url_buf);
					g_free(tmpurlbuf);
					c = t;

					break;
				} else {
					g_string_append_unichar(gurl_buf, g);
					t = g_utf8_find_next_char(t, NULL);
				}
			}

			if (gurl_buf) {
				g_string_free(gurl_buf, TRUE);
			}
		}

		if(*c == ')' && !inside_html) {
			inside_paren--;
			ret = g_string_append_c(ret, *c);
			c++;
		}

		if (*c == 0)
			break;

		ret = g_string_append_c(ret, *c);
		c++;

	}
	return g_string_free_and_steal(ret);
}

char *
purple_unescape_text(const char *in)
{
	GString *ret;
	const char *c = in;

	if(in == NULL) {
		return NULL;
	}

	ret = g_string_new("");
	while(*c) {
		int len;
		const char *ent;

		if((ent = purple_markup_unescape_entity(c, &len)) != NULL) {
			g_string_append(ret, ent);
			c += len;
		} else {
			g_string_append_c(ret, *c);
			c++;
		}
	}

	return g_string_free_and_steal(ret);
}

char *purple_unescape_html(const char *html)
{
	GString *ret;
	const char *c = html;

	if (html == NULL)
		return NULL;

	ret = g_string_new("");
	while (*c) {
		int len;
		const char *ent;

		if ((ent = purple_markup_unescape_entity(c, &len)) != NULL) {
			g_string_append(ret, ent);
			c += len;
		} else if (!strncmp(c, "<br>", 4)) {
			g_string_append_c(ret, '\n');
			c += 4;
		} else {
			g_string_append_c(ret, *c);
			c++;
		}
	}

	return g_string_free_and_steal(ret);
}