l3afpad/src/gtksourceiter.c

/* -*- Mode: C; tab-width: 8; indent-tabs-mode: t; c-basic-offset: 8 -*-
 *  gtksourceiter.h
 *
 *  Copyright (C) 2000, 2002 Paolo Maggi
 *  Copyright (C) 2002, 2003 Jeroen Zwartepoorte
 *
 *  This library is free software; you can redistribute it and/or
 *  modify it under the terms of the GNU Lesser General Public
 *  License as published by the Free Software Foundation; either
 *  version 2.1 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
 *  Lesser General Public License for more details.
 *
 *  You should have received a copy of the GNU Lesser General Public
 *  License along with this library; if not, write to the Free Software
 *  Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA  02110-1301  USA
 */

/*
 * Parts of this file are copied from the gedit and glimmer project.
 */

#include <string.h>
#include "gtksourceiter.h"
#include "l3afpad.h"

#define GTK_TEXT_UNKNOWN_CHAR 0xFFFC

/* this function acts like g_utf8_offset_to_pointer() except that if it finds a
 * decomposable character it consumes the decomposition length from the given
 * offset.  So it's useful when the offset was calculated for the normalized
 * version of str, but we need a pointer to str itself. */
static const gchar *
pointer_from_offset_skipping_decomp (const gchar *str, gint offset)
{
	gchar *casefold, *normal;
	const gchar *p, *q;

	p = str;
	while (offset > 0)
	{
		q = g_utf8_next_char (p);
		casefold = g_utf8_casefold (p, q - p);
		normal = g_utf8_normalize (casefold, -1, G_NORMALIZE_NFD);
		offset -= g_utf8_strlen (normal, -1);
		g_free (casefold);
		g_free (normal);
		p = q;
	}
	return p;
}

static const gchar *
g_utf8_strcasestr (const gchar *haystack, const gchar *needle)
{
	gsize needle_len;
	gsize haystack_len;
	const gchar *ret = NULL;
	gchar *p;
	gchar *casefold;
	gchar *caseless_haystack;
	gint i;

	g_return_val_if_fail (haystack != NULL, NULL);
	g_return_val_if_fail (needle != NULL, NULL);

	casefold = g_utf8_casefold (haystack, -1);
	caseless_haystack = g_utf8_normalize (casefold, -1, G_NORMALIZE_DEFAULT);
	g_free (casefold);

	needle_len = g_utf8_strlen (needle, -1);
	haystack_len = g_utf8_strlen (caseless_haystack, -1);

	if (needle_len == 0)
	{
		ret = (gchar *)haystack;
		goto finally_1;
	}

	if (haystack_len < needle_len)
	{
		ret = NULL;
		goto finally_1;
	}

	p = (gchar*)caseless_haystack;
	needle_len = strlen (needle);
	i = 0;

	while (*p)
	{
		if ((strncmp (p, needle, needle_len) == 0))
		{
			ret = pointer_from_offset_skipping_decomp (haystack, i);
			goto finally_1;
		}

		p = g_utf8_next_char (p);
		i++;
	}

finally_1:
	g_free (caseless_haystack);

	return ret;
}

static const gchar *
g_utf8_strrcasestr (const gchar *haystack, const gchar *needle)
{
	gsize needle_len;
	gsize haystack_len;
	const gchar *ret = NULL;
	gchar *p;
	gchar *casefold;
	gchar *caseless_haystack;
	gint i;

	g_return_val_if_fail (haystack != NULL, NULL);
	g_return_val_if_fail (needle != NULL, NULL);

	casefold = g_utf8_casefold (haystack, -1);
	caseless_haystack = g_utf8_normalize (casefold, -1, G_NORMALIZE_ALL);
	g_free (casefold);

	needle_len = g_utf8_strlen (needle, -1);
	haystack_len = g_utf8_strlen (caseless_haystack, -1);

	if (needle_len == 0)
	{
		ret = (gchar *)haystack;
		goto finally_1;
	}

	if (haystack_len < needle_len)
	{
		ret = NULL;
		goto finally_1;
	}

	i = haystack_len - needle_len;
	p = g_utf8_offset_to_pointer (caseless_haystack, i);
	needle_len = strlen (needle);

	while (p >= caseless_haystack)
	{
		if (strncmp (p, needle, needle_len) == 0)
		{
			ret = pointer_from_offset_skipping_decomp (haystack, i);
			goto finally_1;
		}

		p = g_utf8_prev_char (p);
		i--;
	}

finally_1:
	g_free (caseless_haystack);

	return ret;
}

static gboolean
g_utf8_caselessnmatch (const char *s1, const char *s2,
		       gssize n1, gssize n2)
{
	gchar *casefold;
	gchar *normalized_s1;
	gchar *normalized_s2;
	gint len_s1;
	gint len_s2;
	gboolean ret = FALSE;

	g_return_val_if_fail (s1 != NULL, FALSE);
	g_return_val_if_fail (s2 != NULL, FALSE);
	g_return_val_if_fail (n1 > 0, FALSE);
	g_return_val_if_fail (n2 > 0, FALSE);

	casefold = g_utf8_casefold (s1, n1);
	normalized_s1 = g_utf8_normalize (casefold, -1, G_NORMALIZE_ALL);
	g_free (casefold);

	casefold = g_utf8_casefold (s2, n2);
	normalized_s2 = g_utf8_normalize (casefold, -1, G_NORMALIZE_ALL);
	g_free (casefold);

	len_s1 = strlen (normalized_s1);
	len_s2 = strlen (normalized_s2);

	if (len_s1 < len_s2)
		goto finally_2;

	ret = (strncmp (normalized_s1, normalized_s2, len_s2) == 0);

finally_2:
	g_free (normalized_s1);
	g_free (normalized_s2);

	return ret;
}

static void
forward_chars_with_skipping(GtkTextIter * iter,
			     gint         count,
			     gboolean     skip_invisible,
                            gboolean skip_nontext, gboolean skip_decomp)
{
	gint i;

    g_return_if_fail(count >= 0);

	i = count;

	while (i > 0)
	{
		gboolean ignored = FALSE;

        /* minimal workaround to avoid the infinite loop of bug #168247. It
         * doesn't fix the problemjust the symptom... */
        if (gtk_text_iter_is_end(iter))
            return;

        if (skip_nontext
            && gtk_text_iter_get_char(iter) == GTK_TEXT_UNKNOWN_CHAR)
			ignored = TRUE;

		if (!ignored && skip_invisible &&
            /* _gtk_text_btree_char_is_invisible (iter) */ FALSE)
			ignored = TRUE;

		if (!ignored && skip_decomp)
		{
            /* being UTF8 correct sucks; this accounts for extra offsets
             * coming from canonical decompositions of UTF8 characters (e.g.
             * accented characters) which g_utf8_normalize() performs */
            gchar *normal;
            gchar buffer[6];
            gint buffer_len;

            buffer_len =
                g_unichar_to_utf8(gtk_text_iter_get_char(iter), buffer);
            normal = g_utf8_normalize(buffer, buffer_len, G_NORMALIZE_NFD);
            i -= (g_utf8_strlen(normal, -1) - 1);
            g_free(normal);
		}

        gtk_text_iter_forward_char(iter);

		if (!ignored)
			--i;
	}
}

static gboolean
lines_match(const GtkTextIter * start,
            const gchar ** lines,
	     gboolean           visible_only,
	     gboolean           slice,
            GtkTextIter * match_start, GtkTextIter * match_end)
{
	GtkTextIter next;
	gchar *line_text;
	const gchar *found;
	gint offset;

	if (*lines == NULL || **lines == '\0')
	{
		if (match_start)
			*match_start = *start;
		if (match_end)
			*match_end = *start;
		return TRUE;
	}

	next = *start;
    gtk_text_iter_forward_line(&next);

	/* No more text in buffer, but *lines is nonempty */
    if (gtk_text_iter_equal(start, &next))
		return FALSE;

	if (slice)
	{
		if (visible_only)
            line_text = gtk_text_iter_get_visible_slice(start, &next);
		else
            line_text = gtk_text_iter_get_slice(start, &next);
	}
	else
	{
		if (visible_only)
            line_text = gtk_text_iter_get_visible_text(start, &next);
		else
            line_text = gtk_text_iter_get_text(start, &next);
	}

	if (match_start) /* if this is the first line we're matching */
	{
        found = g_utf8_strcasestr(line_text, *lines);
	}
	else
	{
        /* If it's not the first line, we have to match from the start of the
         * line. */
        if (g_utf8_caselessnmatch(line_text, *lines, strlen(line_text),
                                  strlen(*lines)))
			found = line_text;
		else
			found = NULL;
	}

	if (found == NULL)
	{
        g_free(line_text);
		return FALSE;
	}

	/* Get offset to start of search string */
    offset = g_utf8_strlen(line_text, found - line_text);

	next = *start;

    /* If match start needs to be returned, set it to the start of the search
     * string. */
    forward_chars_with_skipping(&next, offset, visible_only, !slice, FALSE);
	if (match_start)
	{
		*match_start = next;
	}

	/* Go to end of search string */
    forward_chars_with_skipping(&next, g_utf8_strlen(*lines, -1),
                                visible_only, !slice, TRUE);

    g_free(line_text);

	++lines;

	if (match_end)
		*match_end = next;

    /* pass NULL for match_start, since we don't need to find the start
     * again. */
    return lines_match(&next, lines, visible_only, slice, NULL, match_end);
}

static gboolean
backward_lines_match(const GtkTextIter * start,
                     const gchar ** lines,
		      gboolean           visible_only,
		      gboolean           slice,
                     GtkTextIter * match_start, GtkTextIter * match_end)
{
	GtkTextIter line, next;
	gchar *line_text;
	const gchar *found;
	gint offset;

	if (*lines == NULL || **lines == '\0')
	{
		if (match_start)
			*match_start = *start;
		if (match_end)
			*match_end = *start;
		return TRUE;
	}

	line = next = *start;
    if (gtk_text_iter_get_line_offset(&next) == 0)
	{
        if (!gtk_text_iter_backward_line(&next))
			return FALSE;
	}
	else
        gtk_text_iter_set_line_offset(&next, 0);

	if (slice)
	{
		if (visible_only)
            line_text = gtk_text_iter_get_visible_slice(&next, &line);
		else
            line_text = gtk_text_iter_get_slice(&next, &line);
	}
	else
	{
		if (visible_only)
            line_text = gtk_text_iter_get_visible_text(&next, &line);
		else
            line_text = gtk_text_iter_get_text(&next, &line);
	}

	if (match_start) /* if this is the first line we're matching */
	{
        found = g_utf8_strrcasestr(line_text, *lines);
	}
	else
	{
        /* If it's not the first line, we have to match from the start of the
         * line. */
        if (g_utf8_caselessnmatch(line_text, *lines, strlen(line_text),
                                  strlen(*lines)))
			found = line_text;
		else
			found = NULL;
	}

	if (found == NULL)
	{
        g_free(line_text);
		return FALSE;
	}

	/* Get offset to start of search string */
    offset = g_utf8_strlen(line_text, found - line_text);

    forward_chars_with_skipping(&next, offset, visible_only, !slice, FALSE);

    /* If match start needs to be returned, set it to the start of the search
     * string. */
	if (match_start)
	{
		*match_start = next;
	}

	/* Go to end of search string */
    forward_chars_with_skipping(&next, g_utf8_strlen(*lines, -1),
                                visible_only, !slice, TRUE);

    g_free(line_text);

	++lines;

	if (match_end)
		*match_end = next;

    /* try to match the rest of the lines forward, passing NULL for
     * match_start so lines_match will try to match the entire line */
    return lines_match(&next, lines, visible_only, slice, NULL, match_end);
}

/* strsplit () that retains the delimiter as part of the string. */
static gchar **
strbreakup (const char *string,
	    const char *delimiter,
	    gint        max_tokens)
{
	GSList *string_list = NULL, *slist;
	gchar **str_array, *s, *casefold, *new_string;
	guint i, n = 1;

	g_return_val_if_fail (string != NULL, NULL);
	g_return_val_if_fail (delimiter != NULL, NULL);

	if (max_tokens < 1)
		max_tokens = G_MAXINT;

	s = strstr (string, delimiter);
	if (s)
	{
		guint delimiter_len = strlen (delimiter);

		do
		{
			guint len;

			len = s - string + delimiter_len;
			new_string = g_new (gchar, len + 1);
			strncpy (new_string, string, len);
			new_string[len] = 0;
			casefold = g_utf8_casefold (new_string, -1);
			g_free (new_string);
			new_string = g_utf8_normalize (casefold, -1, G_NORMALIZE_ALL);
			g_free (casefold);
			string_list = g_slist_prepend (string_list, new_string);
			n++;
			string = s + delimiter_len;
			s = strstr (string, delimiter);
		} while (--max_tokens && s);
	}

	if (*string)
	{
		n++;
		casefold = g_utf8_casefold (string, -1);
		new_string = g_utf8_normalize (casefold, -1, G_NORMALIZE_ALL);
		g_free (casefold);
		string_list = g_slist_prepend (string_list, new_string);
	}

	str_array = g_new (gchar*, n);

	i = n - 1;

	str_array[i--] = NULL;
	for (slist = string_list; slist; slist = slist->next)
		str_array[i--] = slist->data;

	g_slist_free (string_list);

	return str_array;
}

/**
 * gtk_source_iter_forward_search:
 * @iter: start of search
 * @str: a search string
 * @flags: flags affecting how the search is done
 * @match_start: return location for start of match, or %NULL
 * @match_end: return location for end of match, or %NULL
 * @limit: bound for the search, or %NULL for the end of the buffer
 *
 * Searches forward for @str. Any match is returned by setting
 * @match_start to the first character of the match and @match_end to the
 * first character after the match. The search will not continue past
 * @limit. Note that a search is a linear or O(n) operation, so you
 * may wish to use @limit to avoid locking up your UI on large
 * buffers.
 *
 * If the #GTK_SOURCE_SEARCH_VISIBLE_ONLY flag is present, the match may
 * have invisible text interspersed in @str. i.e. @str will be a
 * possibly-noncontiguous subsequence of the matched range. similarly,
 * if you specify #GTK_SOURCE_SEARCH_TEXT_ONLY, the match may have
 * pixbufs or child widgets mixed inside the matched range. If these
 * flags are not given, the match must be exact; the special 0xFFFC
 * character in @str will match embedded pixbufs or child widgets.
 * If you specify the #GTK_SOURCE_SEARCH_CASE_INSENSITIVE flag, the text will
 * be matched regardless of what case it is in.
 *
 * Same as gtk_text_iter_forward_search(), but supports case insensitive
 * searching.
 *
 * Return value: whether a match was found
 **/
gboolean
gtk_source_iter_forward_search(const GtkTextIter * iter,
                               const gchar * str,
				GtkSourceSearchFlags flags,
                               GtkTextIter * match_start,
                               GtkTextIter * match_end,
                               const GtkTextIter * limit)
{
	gchar **lines = NULL;
	GtkTextIter match;
	gboolean retval = FALSE;
	GtkTextIter search;
	gboolean visible_only;
	gboolean slice;

    g_return_val_if_fail(iter != NULL, FALSE);
    g_return_val_if_fail(str != NULL, FALSE);

	if ((flags & GTK_SOURCE_SEARCH_CASE_INSENSITIVE) == 0)
        return gtk_text_iter_forward_search(iter, str, flags,
                                            match_start, match_end, limit);

    if (limit && gtk_text_iter_compare(iter, limit) >= 0)
		return FALSE;

	if (*str == '\0')
	{
		/* If we can move one char, return the empty string there */
		match = *iter;

        if (gtk_text_iter_forward_char(&match))
		{
            if (limit && gtk_text_iter_equal(&match, limit))
				return FALSE;

			if (match_start)
				*match_start = match;
			if (match_end)
				*match_end = match;
			return TRUE;
		}
		else
		{
			return FALSE;
		}
	}

	visible_only = (flags & GTK_SOURCE_SEARCH_VISIBLE_ONLY) != 0;
	slice = (flags & GTK_SOURCE_SEARCH_TEXT_ONLY) == 0;

	/* locate all lines */
    lines = strbreakup(str, "\n", -1);

	search = *iter;

	do
	{
		/* This loop has an inefficient worst-case, where
         * gtk_text_iter_get_text () is called repeatedly on a single line. */
		GtkTextIter end;

        if (limit && gtk_text_iter_compare(&search, limit) >= 0)
			break;

        if (lines_match(&search, (const gchar **) lines,
				 visible_only, slice, &match, &end))
		{
			if (limit == NULL || (limit &&
                                  gtk_text_iter_compare(&end, limit) < 0))
			{
				retval = TRUE;

				if (match_start)
					*match_start = match;
				if (match_end)
					*match_end = end;
			}
			break;
		}
    } while (gtk_text_iter_forward_line(&search));

    g_strfreev((gchar **) lines);

	return retval;
}

/**
 * gtk_source_iter_backward_search:
 * @iter: a #GtkTextIter where the search begins
 * @str: search string
 * @flags: bitmask of flags affecting the search
 * @match_start: return location for start of match, or %NULL
 * @match_end: return location for end of match, or %NULL
 * @limit: location of last possible @match_start, or %NULL for start of buffer
 *
 * Same as gtk_text_iter_backward_search(), but supports case insensitive
 * searching.
 *
 * Return value: whether a match was found
 **/
gboolean
gtk_source_iter_backward_search(const GtkTextIter * iter,
                                const gchar * str,
				 GtkSourceSearchFlags flags,
                                GtkTextIter * match_start,
                                GtkTextIter * match_end,
                                const GtkTextIter * limit)
{
	gchar **lines = NULL;
	GtkTextIter match;
	gboolean retval = FALSE;
	GtkTextIter search;
	gboolean visible_only;
	gboolean slice;

    g_return_val_if_fail(iter != NULL, FALSE);
    g_return_val_if_fail(str != NULL, FALSE);

	if ((flags & GTK_SOURCE_SEARCH_CASE_INSENSITIVE) == 0)
        return gtk_text_iter_backward_search(iter, str, flags,
                                             match_start, match_end, limit);

    if (limit && gtk_text_iter_compare(iter, limit) <= 0)
		return FALSE;

	if (*str == '\0')
	{
		/* If we can move one char, return the empty string there */
		match = *iter;

        if (gtk_text_iter_backward_char(&match))
		{
            if (limit && gtk_text_iter_equal(&match, limit))
				return FALSE;

			if (match_start)
				*match_start = match;
			if (match_end)
				*match_end = match;
			return TRUE;
		}
		else
		{
			return FALSE;
		}
	}

	visible_only = (flags & GTK_SOURCE_SEARCH_VISIBLE_ONLY) != 0;
	slice = (flags & GTK_SOURCE_SEARCH_TEXT_ONLY) == 0;

	/* locate all lines */
    lines = strbreakup(str, "\n", -1);

	search = *iter;

	while (TRUE)
	{
		/* This loop has an inefficient worst-case, where
         * gtk_text_iter_get_text () is called repeatedly on a single line. */
		GtkTextIter end;

        if (limit && gtk_text_iter_compare(&search, limit) <= 0)
			break;

        if (backward_lines_match(&search, (const gchar **) lines,
					  visible_only, slice, &match, &end))
		{
			if (limit == NULL || (limit &&
                                  gtk_text_iter_compare(&end, limit) > 0))
			{
				retval = TRUE;

				if (match_start)
					*match_start = match;
				if (match_end)
					*match_end = end;
			}
			break;
		}

        if (gtk_text_iter_get_line_offset(&search) == 0)
		{
            if (!gtk_text_iter_backward_line(&search))
				break;
		}
		else
		{
            gtk_text_iter_set_line_offset(&search, 0);
		}
	}

    g_strfreev((gchar **) lines);

	return retval;
}

/*
 * gtk_source_iter_find_matching_bracket is implemented in gtksourcebuffer.c
 */