/**
* Parse gtk-gnutella's version number in User-Agent/Server string `str'
* and extract timestamp into `ver'.
*/
static void
version_stamp(const char *str, version_t *ver)
{
static char stamp[256];
const char *p;
ver->timestamp = 0;
/*
* A typical vendor string with a timestamp would look like:
*
* gtk-gnutella/0.85 (04/04/2002; X11; FreeBSD 4.6-STABLE i386)
*
* The date stamp is formattted as DD/MM/YYYY here, but the date2time()
* routine is also able to parse the ISO format YYYY-MM-DD which is
* being used starting 2004-03-02.
*/
p = strchr(str, '(');
if (p) {
const char *end;
p++;
end = strchr(p, ';');
if (end == NULL)
end = strchr(p, ')'); /* Only date present: short version */
if (end) {
size_t size = end - p + 1;
/*
* Using date2time() will allow us to possibly change the date
* format in the future, without impacting the ability of older
* servents to parse it.
*/
g_strlcpy(stamp, p, MIN(size, sizeof(stamp)));
ver->timestamp = date2time(stamp, tm_time());
if (ver->timestamp == -1) {
ver->timestamp = 0;
g_warning("could not parse timestamp \"%s\" in \"%s\"", p, str);
}
} else
g_warning("no timestamp in \"%s\"", str);
}
}
/**
* Load spam database from the supplied FILE.
*
* The current file format is as follows:
*
* # Comment
* SHA1 <SHA-1>
* ADDED <date>
* END
*
* @returns the amount of entries loaded or -1 on failure.
*/
static G_GNUC_COLD gulong
spam_load(FILE *f)
{
static const struct spam_item zero_item;
struct spam_item item;
char line[1024];
guint line_no = 0;
bit_array_t tag_used[BIT_ARRAY_SIZE(NUM_SPAM_TAGS)];
gulong item_count = 0;
g_assert(f);
/* Reset state */
item = zero_item;
bit_array_init(tag_used, NUM_SPAM_TAGS);
while (fgets(line, sizeof line, f)) {
const char *tag_name, *value;
char *sp, *nl;
spam_tag_t tag;
line_no++;
nl = strchr(line, '\n');
if (!nl) {
/*
* If the line is too long or unterminated the file is either
* corrupt or was manually edited without respecting the
* exact format. If we continued, we would read from the
* middle of a line which could be the filename or ID.
*/
g_warning("spam_load(): "
"line too long or missing newline in line %u",
line_no);
break;
}
*nl = '\0';
/* Skip comments and empty lines */
if (*line == '#' || *line == '\0')
continue;
sp = strchr(line, ' ');
if (sp) {
*sp = '\0';
value = &sp[1];
} else {
value = strchr(line, '\0');
}
tag_name = line;
tag = spam_string_to_tag(tag_name);
g_assert(UNSIGNED(tag) < UNSIGNED(NUM_SPAM_TAGS));
if (SPAM_TAG_UNKNOWN != tag && !bit_array_flip(tag_used, tag)) {
g_warning("spam_load(): duplicate tag \"%s\" in entry in line %u",
tag_name, line_no);
continue;
}
switch (tag) {
case SPAM_TAG_ADDED:
{
time_t t;
t = date2time(value, tm_time());
if ((time_t) -1 == t) {
item.damaged = TRUE;
}
}
break;
case SPAM_TAG_SHA1:
{
if (strlen(value) != SHA1_BASE32_SIZE) {
item.damaged = TRUE;
g_warning("spam_load(): SHA-1 has wrong length.");
} else {
const struct sha1 *raw;
raw = base32_sha1(value);
if (raw)
item.sha1 = *raw;
else
item.damaged = TRUE;
}
}
break;
case SPAM_TAG_NAME:
{
if ('\0' == value[0]) {
item.damaged = TRUE;
g_warning("spam_load(): Missing filename pattern.");
} else if (!utf8_is_valid_string(value)) {
item.damaged = TRUE;
g_warning("spam_load(): Filename pattern is not UTF-8.");
} else {
item.name = h_strdup(value);
}
}
break;
case SPAM_TAG_SIZE:
{
const char *endptr;
guint64 u;
int error;
u = parse_uint64(value, &endptr, 10, &error);
if (error) {
item.damaged = TRUE;
g_warning("spam_load(): Cannot parse SIZE: %s", value);
} else {
item.min_size = u;
item.max_size = u;
if ('-' == endptr[0]) {
u = parse_uint64(&endptr[1], &endptr, 10, &error);
if (error) {
item.damaged = TRUE;
g_warning("spam_load(): Cannot parse SIZE: %s",
value);
}
if (u < item.min_size) {
item.damaged = TRUE;
g_warning("spam_load(): "
"Maximum size below minimum size");
} else {
item.max_size = u;
}
}
}
}
break;
case SPAM_TAG_END:
if (
!bit_array_get(tag_used, SPAM_TAG_SHA1) &&
!bit_array_get(tag_used, SPAM_TAG_NAME)
) {
g_warning("spam_load(): missing SHA1 or NAME tag");
item.damaged = TRUE;
}
if (!bit_array_get(tag_used, SPAM_TAG_ADDED)) {
g_warning("spam_load(): missing ADDED tag");
item.damaged = TRUE;
}
item.done = TRUE;
break;
case SPAM_TAG_UNKNOWN:
/* Ignore */
break;
case NUM_SPAM_TAGS:
g_assert_not_reached();
break;
}
if (item.done && !item.damaged) {
if (bit_array_get(tag_used, SPAM_TAG_SHA1)) {
spam_sha1_add(&item.sha1);
item_count++;
}
if (bit_array_get(tag_used, SPAM_TAG_NAME)) {
if (!bit_array_get(tag_used, SPAM_TAG_SIZE)) {
item.min_size = 0;
item.max_size = MAX_INT_VAL(filesize_t);
}
if (
spam_add_name_and_size(item.name,
item.min_size, item.max_size)
) {
item.damaged = TRUE;
} else {
item_count++;
}
}
}
if (item.damaged) {
g_warning("Damaged spam entry in line %u: "
"tag_name=\"%s\", value=\"%s\"",
line_no, tag_name, value);
}
if (item.done) {
/* Reset state */
HFREE_NULL(item.name);
item = zero_item;
bit_array_clear_range(tag_used, 0, NUM_SPAM_TAGS - 1U);
}
}
spam_sha1_sync();
return item_count;
}
