netsurf/content/handlers/html/search.c

/*
 * Copyright 2004 John M Bell <jmb202@ecs.soton.ac.uk>
 * Copyright 2005 Adrian Lees <adrianl@users.sourceforge.net>
 * Copyright 2009 Mark Benjamin <netsurf-browser.org.MarkBenjamin@dfgh.net>
 *
 * This file is part of NetSurf, http://www.netsurf-browser.org/
 *
 * NetSurf 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; version 2 of the License.
 *
 * NetSurf 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 program.  If not, see <http://www.gnu.org/licenses/>.
 */

/**
 * \file
 * Free text search (core)
 */

#include <ctype.h>
#include <string.h>
#include <dom/dom.h>

#include "utils/config.h"
#include "utils/log.h"
#include "utils/messages.h"
#include "utils/utils.h"
#include "content/content.h"
#include "content/hlcache.h"
#include "desktop/selection.h"
#include "netsurf/search.h"
#include "netsurf/misc.h"
#include "desktop/gui_internal.h"

#include "text/textplain.h"
#include "html/box.h"
#include "html/box_inspect.h"
#include "html/html.h"
#include "html/private.h"
#include "html/search.h"

#ifndef NOF_ELEMENTS
#define NOF_ELEMENTS(array) (sizeof(array)/sizeof(*(array)))
#endif


struct list_entry {
	unsigned start_idx;	/* start position of match */
	unsigned end_idx;	/* end of match */

	struct box *start_box;	/* used only for html contents */
	struct box *end_box;

	struct selection *sel;

	struct list_entry *prev;
	struct list_entry *next;
};

struct search_context {
	void *gui_p;
	struct content *c;
	struct list_entry *found;
	struct list_entry *current; /* first for select all */
	char *string;
	bool prev_case_sens;
	bool newsearch;
	bool is_html;
};


/* Exported function documented in search.h */
struct search_context *
search_create_context(struct content *c, content_type type, void *gui_data)
{
	struct search_context *context;
	struct list_entry *search_head;

	if (type != CONTENT_HTML && type != CONTENT_TEXTPLAIN) {
		return NULL;
	}

	context = malloc(sizeof(struct search_context));
	if (context == NULL) {
		return NULL;
	}

	search_head = malloc(sizeof(struct list_entry));
	if (search_head == NULL) {
		free(context);
		return NULL;
	}

	search_head->start_idx = 0;
	search_head->end_idx = 0;
	search_head->start_box = NULL;
	search_head->end_box = NULL;
	search_head->sel = NULL;
	search_head->prev = NULL;
	search_head->next = NULL;

	context->found = search_head;
	context->current = NULL;
	context->string = NULL;
	context->prev_case_sens = false;
	context->newsearch = true;
	context->c = c;
	context->is_html = (type == CONTENT_HTML) ? true : false;
	context->gui_p = gui_data;

	return context;
}


/**
 * Release the memory used by the list of matches,
 * deleting selection objects too
 */

static void free_matches(struct search_context *context)
{
	struct list_entry *a;
	struct list_entry *b;

	a = context->found->next;

	/* empty the list before clearing and deleting the
	 * selections because the the clearing updates the
	 * screen immediately, causing nested accesses to the list */

	context->found->prev = NULL;
	context->found->next = NULL;

	for (; a; a = b) {
		b = a->next;
		if (a->sel) {
			selection_clear(a->sel, true);
			selection_destroy(a->sel);
		}
		free(a);
	}
}


/**
 * Find the first occurrence of 'match' in 'string' and return its index
 *
 * \param  string     the string to be searched (unterminated)
 * \param  s_len      length of the string to be searched
 * \param  pattern    the pattern for which we are searching (unterminated)
 * \param  p_len      length of pattern
 * \param  case_sens  true iff case sensitive match required
 * \param  m_len      accepts length of match in bytes
 * \return pointer to first match, NULL if none
 */

static const char *find_pattern(const char *string, int s_len,
		const char *pattern, int p_len, bool case_sens,
		unsigned int *m_len)
{
	struct { const char *ss, *s, *p; bool first; } context[16];
	const char *ep = pattern + p_len;
	const char *es = string  + s_len;
	const char *p = pattern - 1;  /* a virtual '*' before the pattern */
	const char *ss = string;
	const char *s = string;
	bool first = true;
	int top = 0;

	while (p < ep) {
		bool matches;
		if (p < pattern || *p == '*') {
			char ch;

			/* skip any further asterisks; one is the same as many
			*/
			do p++; while (p < ep && *p == '*');

			/* if we're at the end of the pattern, yes, it matches
			*/
			if (p >= ep) break;

			/* anything matches a # so continue matching from
			   here, and stack a context that will try to match
			   the wildcard against the next character */

			ch = *p;
			if (ch != '#') {
				/* scan forwards until we find a match for
				   this char */
				if (!case_sens) ch = toupper(ch);
				while (s < es) {
					if (case_sens) {
						if (*s == ch) break;
					} else if (toupper(*s) == ch)
						break;
					s++;
				}
			}

			if (s < es) {
				/* remember where we are in case the match
				   fails; we may then resume */
				if (top < (int)NOF_ELEMENTS(context)) {
					context[top].ss = ss;
					context[top].s  = s + 1;
					context[top].p  = p - 1;
					/* ptr to last asterisk */
					context[top].first = first;
					top++;
				}

				if (first) {
					ss = s;
					/* remember first non-'*' char */
					first = false;
				}

				matches = true;
			} else {
				matches = false;
			}

		} else if (s < es) {
			char ch = *p;
			if (ch == '#')
				matches = true;
			else {
				if (case_sens)
					matches = (*s == ch);
				else
					matches = (toupper(*s) == toupper(ch));
			}
			if (matches && first) {
				ss = s;  /* remember first non-'*' char */
				first = false;
			}
		} else {
			matches = false;
		}

		if (matches) {
			p++; s++;
		} else {
			/* doesn't match,
			 * resume with stacked context if we have one */
			if (--top < 0)
				return NULL;  /* no match, give up */

			ss = context[top].ss;
			s  = context[top].s;
			p  = context[top].p;
			first = context[top].first;
		}
	}

	/* end of pattern reached */
	*m_len = max(s - ss, 1);
	return ss;
}


/**
 * Add a new entry to the list of matches
 *
 * \param start_idx Offset of match start within textual representation
 * \param end_idx Offset of match end
 * \param context The search context to add the entry to.
 * \return Pointer to added entry, NULL iff failed.
 */
static struct list_entry *
add_entry(unsigned start_idx, unsigned end_idx,	struct search_context *context)
{
	struct list_entry *entry;

	/* found string in box => add to list */
	entry = calloc(1, sizeof(*entry));
	if (!entry) {
		return NULL;
	}

	entry->start_idx = start_idx;
	entry->end_idx = end_idx;
	entry->sel = NULL;

	entry->next = 0;
	entry->prev = context->found->prev;

	if (context->found->prev == NULL) {
		context->found->next = entry;
	} else {
		context->found->prev->next = entry;
	}

	context->found->prev = entry;

	return entry;
}


/**
 * Finds all occurrences of a given string in the html box tree
 *
 * \param pattern   the string pattern to search for
 * \param p_len     pattern length
 * \param cur       pointer to the current box
 * \param case_sens whether to perform a case sensitive search
 * \param context   The search context to add the entry to.
 * \return true on success, false on memory allocation failure
 */
static bool find_occurrences_html(const char *pattern, int p_len,
		struct box *cur, bool case_sens,
		struct search_context *context)
{
	struct box *a;

	/* ignore this box, if there's no visible text */
	if (!cur->object && cur->text) {
		const char *text = cur->text;
		unsigned length = cur->length;

		while (length > 0) {
			struct list_entry *entry;
			unsigned match_length;
			unsigned match_offset;
			const char *new_text;
			const char *pos = find_pattern(text, length,
					pattern, p_len, case_sens,
					&match_length);
			if (!pos)
				break;

			/* found string in box => add to list */
			match_offset = pos - cur->text;

			entry = add_entry(cur->byte_offset + match_offset,
						cur->byte_offset +
							match_offset +
							match_length, context);
			if (!entry)
				return false;

			entry->start_box = cur;
			entry->end_box = cur;

			new_text = pos + match_length;
			length -= (new_text - text);
			text = new_text;
		}
	}

	/* and recurse */
	for (a = cur->children; a; a = a->next) {
		if (!find_occurrences_html(pattern, p_len, a, case_sens,
				context))
			return false;
	}

	return true;
}


/**
 * Finds all occurrences of a given string in a textplain content
 *
 * \param pattern   the string pattern to search for
 * \param p_len     pattern length
 * \param c         the content to be searched
 * \param case_sens whether to perform a case sensitive search
 * \param context   The search context to add the entry to.
 * \return true on success, false on memory allocation failure
 */

static bool find_occurrences_text(const char *pattern, int p_len,
		struct content *c, bool case_sens,
		struct search_context *context)
{
	int nlines = textplain_line_count(c);
	int line;

	for(line = 0; line < nlines; line++) {
		size_t offset, length;
		const char *text = textplain_get_line(c, line,
				&offset, &length);
		if (text) {
			while (length > 0) {
				struct list_entry *entry;
				unsigned match_length;
				size_t start_idx;
				const char *new_text;
				const char *pos = find_pattern(text, length,
						pattern, p_len, case_sens,
						&match_length);
				if (!pos)
					break;

				/* found string in line => add to list */
				start_idx = offset + (pos - text);
				entry = add_entry(start_idx, start_idx +
						match_length, context);
				if (!entry)
					return false;

				new_text = pos + match_length;
				offset += (new_text - text);
				length -= (new_text - text);
				text = new_text;
			}
		}
	}

	return true;
}


/**
 * Specifies whether all matches or just the current match should
 * be highlighted in the search text.
 */
static void search_show_all(bool all, struct search_context *context)
{
	struct list_entry *a;

	for (a = context->found->next; a; a = a->next) {
		bool add = true;
		if (!all && a != context->current) {
			add = false;
			if (a->sel) {
				selection_clear(a->sel, true);
				selection_destroy(a->sel);
				a->sel = NULL;
			}
		}
		if (add && !a->sel) {

			if (context->is_html == true) {
				html_content *html = (html_content *)context->c;
				a->sel = selection_create(context->c, true);
				if (!a->sel)
					continue;

				selection_init(a->sel, html->layout,
						&html->len_ctx);
			} else {
				a->sel = selection_create(context->c, false);
				if (!a->sel)
					continue;

				selection_init(a->sel, NULL, NULL);
			}

			selection_set_start(a->sel, a->start_idx);
			selection_set_end(a->sel, a->end_idx);
		}
	}
}


/**
 * Search for a string in the box tree
 *
 * \param string the string to search for
 * \param string_len length of search string
 * \param context The search context to add the entry to.
 * \param flags flags to control the search.
 */
static void
search_text(const char *string,
	    int string_len,
	    struct search_context *context,
	    search_flags_t flags)
{
	struct rect bounds;
	struct box *box = NULL;
	union content_msg_data msg_data;
	bool case_sensitive, forwards, showall;

	case_sensitive = ((flags & SEARCH_FLAG_CASE_SENSITIVE) != 0) ?
			true : false;
	forwards = ((flags & SEARCH_FLAG_FORWARDS) != 0) ? true : false;
	showall = ((flags & SEARCH_FLAG_SHOWALL) != 0) ? true : false;

	if (context->c == NULL)
		return;

	if (context->is_html == true) {
		html_content *html = (html_content *)context->c;

		box = html->layout;

		if (!box)
			return;
	}


	/* check if we need to start a new search or continue an old one */
	if ((context->newsearch) ||
	    (context->prev_case_sens != case_sensitive)) {
		bool res;

		if (context->string != NULL)
			free(context->string);

		context->current = NULL;
		free_matches(context);

		context->string = malloc(string_len + 1);
		if (context->string != NULL) {
			memcpy(context->string, string, string_len);
			context->string[string_len] = '\0';
		}

		guit->search->hourglass(true, context->gui_p);

		if (context->is_html == true) {
			res = find_occurrences_html(string, string_len,
					box, case_sensitive, context);
		} else {
			res = find_occurrences_text(string, string_len,
					context->c, case_sensitive, context);
		}

		if (!res) {
			free_matches(context);
			guit->search->hourglass(false, context->gui_p);
			return;
		}
		guit->search->hourglass(false, context->gui_p);

		context->prev_case_sens = case_sensitive;

		/* new search, beginning at the top of the page */
		context->current = context->found->next;
		context->newsearch = false;

	} else if (context->current != NULL) {
		/* continued search in the direction specified */
		if (forwards) {
			if (context->current->next)
				context->current = context->current->next;
		} else {
			if (context->current->prev)
				context->current = context->current->prev;
		}
	}

	guit->search->status((context->current != NULL), context->gui_p);

	search_show_all(showall, context);

	guit->search->back_state((context->current != NULL) &&
				(context->current->prev != NULL),
				context->gui_p);
	guit->search->forward_state((context->current != NULL) &&
				(context->current->next != NULL),
				context->gui_p);

	if (context->current == NULL)
		return;

	if (context->is_html == true) {
		/* get box position and jump to it */
		box_coords(context->current->start_box, &bounds.x0, &bounds.y0);
		/* \todo: move x0 in by correct idx */
		box_coords(context->current->end_box, &bounds.x1, &bounds.y1);
		/* \todo: move x1 in by correct idx */
		bounds.x1 += context->current->end_box->width;
		bounds.y1 += context->current->end_box->height;
	} else {
		textplain_coords_from_range(context->c,
				context->current->start_idx,
				context->current->end_idx, &bounds);
	}

	msg_data.scroll.area = true;
	msg_data.scroll.x0 = bounds.x0;
	msg_data.scroll.y0 = bounds.y0;
	msg_data.scroll.x1 = bounds.x1;
	msg_data.scroll.y1 = bounds.y1;
	content_broadcast(context->c, CONTENT_MSG_SCROLL, &msg_data);
}


/* Exported function documented in search.h */
void
search_step(struct search_context *context,
	    search_flags_t flags,
	    const char *string)
{
	int string_len;
	int i = 0;

	assert(context != NULL);

	guit->search->add_recent(string, context->gui_p);

	string_len = strlen(string);
	for (i = 0; i < string_len; i++)
		if (string[i] != '#' && string[i] != '*')
			break;
	if (i >= string_len) {
		union content_msg_data msg_data;
		free_matches(context);

		guit->search->status(true, context->gui_p);
		guit->search->back_state(false, context->gui_p);
		guit->search->forward_state(false, context->gui_p);

		msg_data.scroll.area = false;
		msg_data.scroll.x0 = 0;
		msg_data.scroll.y0 = 0;
		content_broadcast(context->c, CONTENT_MSG_SCROLL, &msg_data);
		return;
	}
	search_text(string, string_len, context, flags);
}


/* Exported function documented in search.h */
bool search_term_highlighted(struct content *c,
		unsigned start_offset, unsigned end_offset,
		unsigned *start_idx, unsigned *end_idx,
		struct search_context *context)
{
	if (c == context->c) {
		struct list_entry *a;
		for (a = context->found->next; a; a = a->next)
			if (a->sel && selection_defined(a->sel) &&
					selection_highlighted(a->sel,
						start_offset, end_offset,
						start_idx, end_idx))
				return true;
	}

	return false;
}




/* Exported function documented in search.h */
void search_destroy_context(struct search_context *context)
{
	assert(context != NULL);

	if (context->string != NULL) {
		guit->search->add_recent(context->string, context->gui_p);
		free(context->string);
	}

	guit->search->forward_state(true, context->gui_p);
	guit->search->back_state(true, context->gui_p);

	free_matches(context);
	free(context);
}