nautilus/eel/eel-string.c

/*
 *  eel-string.c: String routines to augment <string.h>.
 *
 *  Copyright (C) 2000 Eazel, Inc.
 *
 *  The Gnome Library is free software; you can redistribute it and/or
 *  modify it under the terms of the GNU Library General Public License as
 *  published by the Free Software Foundation; either version 2 of the
 *  License, or (at your option) any later version.
 *
 *  The Gnome 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
 *  Library General Public License for more details.
 *
 *  You should have received a copy of the GNU Library General Public
 *  License along with the Gnome Library; see the file COPYING.LIB.  If not,
 *  see <http://www.gnu.org/licenses/>.
 *
 *  Authors: Darin Adler <darin@eazel.com>
 */

#include <config.h>
#include "eel-string.h"

#include <errno.h>
#include <locale.h>
#include <stdlib.h>
#include <string.h>
#include <eel-glib-extensions.h>

#if !defined (EEL_OMIT_SELF_CHECK)
#include "eel-lib-self-check-functions.h"
#endif

char *
eel_str_double_underscores (const char *string)
{
    int underscores;
    const char *p;
    char *q;
    char *escaped;

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

    underscores = 0;
    for (p = string; *p != '\0'; p++)
    {
        underscores += (*p == '_');
    }

    if (underscores == 0)
    {
        return g_strdup (string);
    }

    escaped = g_new (char, strlen (string) + underscores + 1);
    for (p = string, q = escaped; *p != '\0'; p++, q++)
    {
        /* Add an extra underscore. */
        if (*p == '_')
        {
            *q++ = '_';
        }
        *q = *p;
    }
    *q = '\0';

    return escaped;
}

char *
eel_str_capitalize (const char *string)
{
    char *capitalized;

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

    capitalized = g_strdup (string);

    capitalized[0] = g_ascii_toupper (capitalized[0]);

    return capitalized;
}

gchar *
eel_str_middle_truncate (const gchar *string,
                         guint        truncate_length)
{
    const gchar ellipsis[] = "…";
    glong ellipsis_length;
    glong length;
    glong num_left_chars;
    glong num_right_chars;
    g_autofree gchar *left_substring = NULL;
    g_autofree gchar *right_substring = NULL;

    g_return_val_if_fail (string != NULL, NULL);
    g_return_val_if_fail (truncate_length > 0, NULL);

    ellipsis_length = g_utf8_strlen (ellipsis, -1);

    /* Our ellipsis string + one character on each side. */
    if (truncate_length < ellipsis_length + 2)
    {
        return g_strdup (string);
    }

    length = g_utf8_strlen (string, -1);

    if (length <= truncate_length)
    {
        return g_strdup (string);
    }

    num_left_chars = (truncate_length - ellipsis_length) / 2;
    num_right_chars = truncate_length - num_left_chars - ellipsis_length;

    g_assert (num_left_chars > 0);
    g_assert (num_right_chars > 0);

    left_substring = g_utf8_substring (string, 0, num_left_chars);
    right_substring = g_utf8_substring (string, length - num_right_chars, length);

    return g_strconcat (left_substring, ellipsis, right_substring, NULL);
}

char *
eel_str_strip_substring_and_after (const char *string,
                                   const char *substring)
{
    const char *substring_position;

    g_return_val_if_fail (substring != NULL, g_strdup (string));
    g_return_val_if_fail (substring[0] != '\0', g_strdup (string));

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

    substring_position = strstr (string, substring);
    if (substring_position == NULL)
    {
        return g_strdup (string);
    }

    return g_strndup (string,
                      substring_position - string);
}

char *
eel_str_replace_substring (const char *string,
                           const char *substring,
                           const char *replacement)
{
    int substring_length, replacement_length, result_length, remaining_length;
    const char *p, *substring_position;
    char *result, *result_position;

    g_return_val_if_fail (substring != NULL, g_strdup (string));
    g_return_val_if_fail (substring[0] != '\0', g_strdup (string));

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

    substring_length = substring ? strlen (substring) : 0;
    replacement_length = replacement ? strlen (replacement) : 0;

    result_length = strlen (string);
    for (p = string;; p = substring_position + substring_length)
    {
        substring_position = strstr (p, substring);
        if (substring_position == NULL)
        {
            break;
        }
        result_length += replacement_length - substring_length;
    }

    result = g_malloc (result_length + 1);

    result_position = result;
    for (p = string;; p = substring_position + substring_length)
    {
        substring_position = strstr (p, substring);
        if (substring_position == NULL)
        {
            remaining_length = strlen (p);
            memcpy (result_position, p, remaining_length);
            result_position += remaining_length;
            break;
        }
        memcpy (result_position, p, substring_position - p);
        result_position += substring_position - p;
        memcpy (result_position, replacement, replacement_length);
        result_position += replacement_length;
    }
    g_assert (result_position - result == result_length);
    result_position[0] = '\0';

    return result;
}

/**
 * get_common_prefix_length:
 * @str_a: first string
 * @str_b: second string
 * @min_required_len: the minimum number of characters required in the prefix
 *
 * Returns: the size of the common prefix of two strings, in characters.
 * If there's no common prefix, or the common prefix is smaller than
 * min_required_len, this will return -1
 */
static int
get_common_prefix_length (char *str_a,
                          char *str_b,
                          int   min_required_len)
{
    int a_len;
    int b_len;
    int intersection_len;
    int matching_chars;
    char *a;
    char *b;

    a_len = g_utf8_strlen (str_a, -1);
    b_len = g_utf8_strlen (str_b, -1);

    intersection_len = MIN (a_len, b_len);
    if (intersection_len < min_required_len)
    {
        return -1;
    }

    matching_chars = 0;
    a = str_a;
    b = str_b;
    while (matching_chars < intersection_len)
    {
        if (g_utf8_get_char (a) != g_utf8_get_char (b))
        {
            break;
        }

        ++matching_chars;

        a = g_utf8_next_char (a);
        b = g_utf8_next_char (b);
    }

    if (matching_chars < min_required_len)
    {
        return -1;
    }

    return matching_chars;
}

char *
eel_str_get_common_prefix (GList *strs,
                           int    min_required_len)
{
    GList *l;
    char *common_part;
    char *name;
    char *truncated;
    int matching_chars;

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

    common_part = NULL;
    for (l = strs; l != NULL; l = l->next)
    {
        name = l->data;
        if (name == NULL)
        {
            g_free (common_part);
            return NULL;
        }

        if (l->prev == NULL)
        {
            common_part = g_strdup (name);
            continue;
        }

        matching_chars = get_common_prefix_length (common_part, name, min_required_len);

        if (matching_chars == -1)
        {
            g_free (common_part);
            return NULL;
        }

        truncated = g_utf8_substring (common_part, 0, matching_chars);
        g_free (common_part);
        common_part = truncated;
    }

    matching_chars = g_utf8_strlen (common_part, -1);
    if (matching_chars < min_required_len)
    {
        g_free (common_part);
        return NULL;
    }

    return common_part;
}

/**************** Custom printf ***********/

typedef struct
{
    const char *start;
    const char *end;
    GString *format;
    int arg_pos;
    int width_pos;
    int width_format_index;
    int precision_pos;
    int precision_format_index;
} ConversionInfo;

enum
{
    ARG_TYPE_INVALID,
    ARG_TYPE_INT,
    ARG_TYPE_LONG,
    ARG_TYPE_LONG_LONG,
    ARG_TYPE_SIZE,
    ARG_TYPE_LONG_DOUBLE,
    ARG_TYPE_DOUBLE,
    ARG_TYPE_POINTER
};

/*********** refcounted strings ****************/

G_LOCK_DEFINE_STATIC (unique_ref_strs);
static GHashTable *unique_ref_strs = NULL;

static eel_ref_str
eel_ref_str_new_internal (const char *string,
                          int         start_count)
{
    char *res;
    volatile gint *count;
    gsize len;

    len = strlen (string);
    res = g_malloc (sizeof (gint) + len + 1);
    count = (volatile gint *) res;
    *count = start_count;
    res += sizeof (gint);
    memcpy (res, string, len + 1);
    return res;
}

eel_ref_str
eel_ref_str_new (const char *string)
{
    if (string == NULL)
    {
        return NULL;
    }

    return eel_ref_str_new_internal (string, 1);
}

eel_ref_str
eel_ref_str_get_unique (const char *string)
{
    eel_ref_str res;

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

    G_LOCK (unique_ref_strs);
    if (unique_ref_strs == NULL)
    {
        unique_ref_strs =
            g_hash_table_new (g_str_hash, g_str_equal);
    }

    res = g_hash_table_lookup (unique_ref_strs, string);
    if (res != NULL)
    {
        eel_ref_str_ref (res);
    }
    else
    {
        res = eel_ref_str_new_internal (string, 0x80000001);
        g_hash_table_insert (unique_ref_strs, res, res);
    }

    G_UNLOCK (unique_ref_strs);

    return res;
}

eel_ref_str
eel_ref_str_ref (eel_ref_str str)
{
    volatile gint *count;

    count = (volatile gint *) ((char *) str - sizeof (gint));
    g_atomic_int_add (count, 1);

    return str;
}

void
eel_ref_str_unref (eel_ref_str str)
{
    volatile gint *count;
    gint old_ref;

    if (str == NULL)
    {
        return;
    }

    count = (volatile gint *) ((char *) str - sizeof (gint));

retry_atomic_decrement:
    old_ref = g_atomic_int_get (count);
    if (old_ref == 1)
    {
        g_free ((char *) count);
    }
    else if (old_ref == 0x80000001)
    {
        G_LOCK (unique_ref_strs);
        /* Need to recheck after taking lock to avoid races with _get_unique() */
        if (g_atomic_int_add (count, -1) == 0x80000001)
        {
            g_hash_table_remove (unique_ref_strs, (char *) str);
            g_free ((char *) count);
        }
        G_UNLOCK (unique_ref_strs);
    }
    else if (!g_atomic_int_compare_and_exchange (count,
                                                 old_ref, old_ref - 1))
    {
        goto retry_atomic_decrement;
    }
}


#if !defined (EEL_OMIT_SELF_CHECK)

void
eel_self_check_string (void)
{
    EEL_CHECK_STRING_RESULT (eel_str_double_underscores (NULL), NULL);
    EEL_CHECK_STRING_RESULT (eel_str_double_underscores (""), "");
    EEL_CHECK_STRING_RESULT (eel_str_double_underscores ("_"), "__");
    EEL_CHECK_STRING_RESULT (eel_str_double_underscores ("foo"), "foo");
    EEL_CHECK_STRING_RESULT (eel_str_double_underscores ("foo_bar"), "foo__bar");
    EEL_CHECK_STRING_RESULT (eel_str_double_underscores ("foo_bar_2"), "foo__bar__2");
    EEL_CHECK_STRING_RESULT (eel_str_double_underscores ("_foo"), "__foo");
    EEL_CHECK_STRING_RESULT (eel_str_double_underscores ("foo_"), "foo__");

    EEL_CHECK_STRING_RESULT (eel_str_capitalize (NULL), NULL);
    EEL_CHECK_STRING_RESULT (eel_str_capitalize (""), "");
    EEL_CHECK_STRING_RESULT (eel_str_capitalize ("foo"), "Foo");
    EEL_CHECK_STRING_RESULT (eel_str_capitalize ("Foo"), "Foo");

    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("foo", 0), "foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("foo", 1), "foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("foo", 3), "foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("foo", 4), "foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("foo", 5), "foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("foo", 6), "foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("foo", 7), "foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("a_much_longer_foo", 0), "a_much_longer_foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("a_much_longer_foo", 1), "a_much_longer_foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("a_much_longer_foo", 2), "a_much_longer_foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("a_much_longer_foo", 3), "a_much_longer_foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("a_much_longer_foo", 4), "a_much_longer_foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("a_much_longer_foo", 5), "a…o");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("a_much_longer_foo", 6), "a…oo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("a_much_longer_foo", 7), "a_…oo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("a_much_longer_foo", 8), "a_…foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("a_much_longer_foo", 9), "a_m…foo");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_even", 8), "so…ven");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_odd", 8), "so…odd");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_even", 9), "som…ven");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_odd", 9), "som…odd");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_even", 10), "som…even");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_odd", 10), "som…_odd");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_even", 11), "some…even");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_odd", 11), "some…_odd");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_even", 12), "some…_even");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_odd", 12), "some…g_odd");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_even", 13), "somet…_even");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_odd", 13), "something_odd");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_even", 14), "something_even");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("something_odd", 13), "something_odd");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("ääääääääää", 5), "ä…ä");
    EEL_CHECK_STRING_RESULT (eel_str_middle_truncate ("あぃいぅうぇえぉ", 7), "あぃ…えぉ");

    EEL_CHECK_STRING_RESULT (eel_str_strip_substring_and_after (NULL, "bar"), NULL);
    EEL_CHECK_STRING_RESULT (eel_str_strip_substring_and_after ("", "bar"), "");
    EEL_CHECK_STRING_RESULT (eel_str_strip_substring_and_after ("foo", "bar"), "foo");
    EEL_CHECK_STRING_RESULT (eel_str_strip_substring_and_after ("foo bar", "bar"), "foo ");
    EEL_CHECK_STRING_RESULT (eel_str_strip_substring_and_after ("foo bar xxx", "bar"), "foo ");
    EEL_CHECK_STRING_RESULT (eel_str_strip_substring_and_after ("bar", "bar"), "");

    EEL_CHECK_STRING_RESULT (eel_str_replace_substring (NULL, "foo", NULL), NULL);
    EEL_CHECK_STRING_RESULT (eel_str_replace_substring (NULL, "foo", "bar"), NULL);
    EEL_CHECK_STRING_RESULT (eel_str_replace_substring ("bar", "foo", NULL), "bar");
    EEL_CHECK_STRING_RESULT (eel_str_replace_substring ("", "foo", ""), "");
    EEL_CHECK_STRING_RESULT (eel_str_replace_substring ("", "foo", "bar"), "");
    EEL_CHECK_STRING_RESULT (eel_str_replace_substring ("bar", "foo", ""), "bar");
    EEL_CHECK_STRING_RESULT (eel_str_replace_substring ("xxx", "x", "foo"), "foofoofoo");
    EEL_CHECK_STRING_RESULT (eel_str_replace_substring ("fff", "f", "foo"), "foofoofoo");
    EEL_CHECK_STRING_RESULT (eel_str_replace_substring ("foofoofoo", "foo", "f"), "fff");
    EEL_CHECK_STRING_RESULT (eel_str_replace_substring ("foofoofoo", "f", ""), "oooooo");
}

#endif /* !EEL_OMIT_SELF_CHECK */