Added case matching
This commit is contained in:
Jesse van den Kieboom
parent
5a63c731f0
commit
f578ba68f5
|
|
@ -1,5 +1,8 @@
|
|||
#include <string.h>
|
||||
#include "gm-searchable.h"
|
||||
|
||||
#define GTK_TEXT_UNKNOWN_CHAR 0xFFFC
|
||||
|
||||
GType
|
||||
gm_searchable_get_type() {
|
||||
static GType searchable_type = 0;
|
||||
|
|
@ -25,6 +28,493 @@ gm_searchable_get_type() {
|
|||
}
|
||||
|
||||
/* Searching implementations */
|
||||
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_NFD);
|
||||
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_NFD);
|
||||
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_NFD);
|
||||
g_free (casefold);
|
||||
|
||||
casefold = g_utf8_casefold (s2, n2);
|
||||
normalized_s2 = g_utf8_normalize (casefold, -1, G_NORMALIZE_NFD);
|
||||
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_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 (gtk_text_iter_get_char (iter) == GTK_TEXT_UNKNOWN_CHAR)
|
||||
ignored = TRUE;
|
||||
|
||||
if (!ignored &&
|
||||
/* _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,
|
||||
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;
|
||||
|
||||
line_text = gtk_text_iter_get_visible_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, FALSE);
|
||||
if (match_start) {
|
||||
*match_start = next;
|
||||
}
|
||||
|
||||
/* Go to end of search string */
|
||||
forward_chars_with_skipping(&next, g_utf8_strlen (*lines, -1), 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, NULL, match_end);
|
||||
}
|
||||
|
||||
static gboolean
|
||||
backward_lines_match (const GtkTextIter *start,
|
||||
const gchar **lines,
|
||||
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);
|
||||
|
||||
line_text = gtk_text_iter_get_visible_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, 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), 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, NULL, match_end);
|
||||
}
|
||||
|
||||
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_NFD);
|
||||
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_NFD);
|
||||
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;
|
||||
}
|
||||
|
||||
typedef gboolean (* search_fun)(const GtkTextIter *iter);
|
||||
|
||||
gboolean
|
||||
gm_searchable_text_view_search_case_insensitive(GtkTextView *text_view, gchar
|
||||
const *str, GmSearchableSearchFlags flags, GtkTextIter *start,
|
||||
GtkTextIter *end) {
|
||||
gchar **lines;
|
||||
lines = strbreakup(str, "\n", -1);
|
||||
GtkTextIter search;
|
||||
gboolean retval = FALSE;
|
||||
|
||||
if (flags & GM_SEARCHABLE_SEARCH_FORWARDS) {
|
||||
search = *end;
|
||||
do {
|
||||
/* This loop has an inefficient worst-case, where
|
||||
* gtk_text_iter_get_text () is called repeatedly on
|
||||
* a single line.
|
||||
*/
|
||||
|
||||
if (lines_match(&search, (const gchar**)lines, start, end)) {
|
||||
retval = TRUE;
|
||||
break;
|
||||
}
|
||||
} while (gtk_text_iter_forward_line(&search));
|
||||
} else {
|
||||
search = *start;
|
||||
while (TRUE)
|
||||
{
|
||||
/* This loop has an inefficient worst-case, where
|
||||
* gtk_text_iter_get_text () is called repeatedly on
|
||||
* a single line.
|
||||
*/
|
||||
|
||||
if (backward_lines_match (&search, (const gchar**)lines,
|
||||
start, end)) {
|
||||
retval = TRUE;
|
||||
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;
|
||||
}
|
||||
|
||||
gboolean
|
||||
gm_searchable_text_view_find_next(GtkTextView *text_view, const gchar *str,
|
||||
GmSearchableSearchFlags flags) {
|
||||
|
|
@ -59,7 +549,12 @@ gm_searchable_text_view_find_next(GtkTextView *text_view, const gchar *str,
|
|||
start = end;
|
||||
}
|
||||
|
||||
if (flags & GM_SEARCHABLE_SEARCH_FORWARDS) {
|
||||
if (flags & GM_SEARCHABLE_SEARCH_CASE_INSENSITIVE) {
|
||||
match_start = start;
|
||||
match_end = end;
|
||||
found = gm_searchable_text_view_search_case_insensitive(text_view,
|
||||
str, flags, &match_start, &match_end);
|
||||
} else if (flags & GM_SEARCHABLE_SEARCH_FORWARDS) {
|
||||
found = gtk_text_iter_forward_search(&end, str,
|
||||
GTK_TEXT_SEARCH_VISIBLE_ONLY | GTK_TEXT_SEARCH_TEXT_ONLY,
|
||||
&match_start, &match_end, NULL);
|
||||
|
|
|
|||
|
|
@ -20,7 +20,8 @@ typedef struct _GmSearchableInterface GmSearchableInterface;
|
|||
typedef enum _GmSearchableSearchFlags {
|
||||
GM_SEARCHABLE_SEARCH_NONE = 0,
|
||||
GM_SEARCHABLE_SEARCH_FORWARDS = 1 << 0,
|
||||
GM_SEARCHABLE_SEARCH_BACKWARDS = 1 << 1
|
||||
GM_SEARCHABLE_SEARCH_BACKWARDS = 1 << 1,
|
||||
GM_SEARCHABLE_SEARCH_CASE_INSENSITIVE = 1 << 2
|
||||
} GmSearchableSearchFlags;
|
||||
|
||||
struct _GmSearchableInterface {
|
||||
|
|
|
|||
Reference in New Issue