static void _add_piece(_metalink_context_t *ctx, const char *value)
{
wget_metalink_t *metalink = ctx->metalink;
sscanf(value, "%127s", ctx->hash);
if (ctx->length && *ctx->hash_type && *ctx->hash) {
// hash for a piece of the file
wget_metalink_piece_t piece, *piecep;
if (!metalink->pieces)
metalink->pieces = wget_vector_create(32, 32, NULL);
piece.length = ctx->length;
strlcpy(piece.hash.type, ctx->hash_type, sizeof(piece.hash.type));
strlcpy(piece.hash.hash_hex, ctx->hash, sizeof(piece.hash.hash_hex));
piecep = wget_vector_get(metalink->pieces, wget_vector_size(metalink->pieces) - 1);
if (piecep)
piece.position = piecep->position + piecep->length;
else
piece.position = 0;
wget_vector_add(metalink->pieces, &piece, sizeof(wget_metalink_piece_t));
}
*ctx->hash = 0;
}
void job_create_parts(JOB *job)
{
PART part;
wget_metalink_t *metalink;
ssize_t fsize;
if (!(metalink = job->metalink))
return;
memset(&part, 0, sizeof(PART));
// create space to hold enough parts
if (!job->parts)
job->parts = wget_vector_create(wget_vector_size(metalink->pieces), 4, NULL);
else
wget_vector_clear(job->parts);
fsize = metalink->size;
for (int it = 0; it < wget_vector_size(metalink->pieces); it++) {
wget_metalink_piece_t *piece = wget_vector_get(metalink->pieces, it);
if (fsize >= piece->length) {
part.length = piece->length;
} else {
part.length = fsize;
}
part.id = it + 1;
wget_vector_add(job->parts, &part, sizeof(PART));
part.position += part.length;
fsize -= piece->length;
}
}
// Callback function, called from CSS parser for each URI found.
static void _css_get_url(void *context, const char *url, size_t len, size_t pos)
{
_CSS_CONTEXT *ctx = context;
WGET_PARSED_URL parsed_url = { .len = len, .pos = pos, .url = wget_strmemdup(url, len), .abs_url = NULL };
if (!ctx->uris) {
ctx->uris = wget_vector_create(16, -2, NULL);
wget_vector_set_destructor(ctx->uris, (wget_vector_destructor_t)_free_url);
}
wget_vector_add(ctx->uris, &parsed_url, sizeof(parsed_url));
}
static void _urls_to_absolute(wget_vector_t *urls, wget_iri_t *base)
{
if (base && urls) {
wget_buffer_t buf;
wget_buffer_init(&buf, NULL, 1024);
for (int it = 0; it < wget_vector_size(urls); it++) {
WGET_PARSED_URL *url = wget_vector_get(urls, it);
if (wget_iri_relative_to_abs(base, url->url, url->len, &buf))
url->abs_url = wget_strmemdup(buf.data, buf.length);
else
error_printf("Cannot resolve relative URI '%s'\n", url->url);
}
wget_buffer_deinit(&buf);
}
}
wget_vector_t *wget_css_get_urls(const char *css, size_t len, wget_iri_t *base, const char **encoding)
{
_CSS_CONTEXT context = { .encoding = encoding };
wget_css_parse_buffer(css, len, _css_get_url, encoding ? _css_get_encoding : NULL, &context);
_urls_to_absolute(context.uris, base);
return context.uris;
}
int job_validate_file(JOB *job)
{
PART part;
wget_metalink_t *metalink;
off_t fsize;
int fd, rc = -1;
struct stat st;
if (!(metalink = job->metalink))
return 0;
memset(&part, 0, sizeof(PART));
// Metalink may be used without pieces...
if (!metalink->pieces) {
wget_metalink_piece_t piece;
wget_metalink_hash_t *hash = wget_vector_get(metalink->hashes, 0);
if (!hash)
return 1;
piece.length = metalink->size;
piece.position = 0;
strlcpy(piece.hash.type, hash->type, sizeof(piece.hash.type));
strlcpy(piece.hash.hash_hex, hash->hash_hex, sizeof(piece.hash.hash_hex));
metalink->pieces = wget_vector_create(1, 1, NULL);
wget_vector_add(metalink->pieces, &piece, sizeof(wget_metalink_piece_t));
}
// create space to hold enough parts
if (!job->parts)
job->parts = wget_vector_create(wget_vector_size(metalink->pieces), 4, NULL);
else
wget_vector_clear(job->parts);
fsize = metalink->size;
if (wget_vector_size(metalink->hashes) == 0) {
// multipart non-metalink download: do not clobber if file has expected size
if (stat(metalink->name, &st) == 0 && st.st_size == fsize) {
return 1; // we are done
}
}
// truncate file if needed
if (stat(metalink->name, &st) == 0 && st.st_size > fsize) {
if (truncate(metalink->name, fsize) == -1)
error_printf(_("Failed to truncate %s\n from %llu to %llu bytes\n"),
metalink->name, (unsigned long long)st.st_size, (unsigned long long)fsize);
}
if ((fd = open(metalink->name, O_RDONLY)) != -1) {
// file exists, check which piece is invalid and requeue it
for (int it = 0; errno != EINTR && it < wget_vector_size(metalink->hashes); it++) {
wget_metalink_hash_t *hash = wget_vector_get(metalink->hashes, it);
if ((rc = _check_file_fd(hash, fd)) == -1)
continue; // hash type not available, try next
break;
}
if (rc == 1) {
info_printf(_("Checksum OK for '%s'\n"), metalink->name);
close(fd);
return 1; // we are done
}
else if (rc == -1) {
// failed to check file, continue as if file is ok
info_printf(_("Failed to build checksum, assuming file to be OK\n"));
close(fd);
return 1; // we are done
} else
info_printf(_("Bad checksum for '%s'\n"), metalink->name);
// if (vec_size(metalink->pieces) < 1)
// return;
for (int it = 0; errno != EINTR && it < wget_vector_size(metalink->pieces); it++) {
wget_metalink_piece_t *piece = wget_vector_get(metalink->pieces, it);
wget_metalink_hash_t *hash = &piece->hash;
if (fsize >= piece->length) {
part.length = piece->length;
} else {
part.length = (size_t)fsize;
}
part.id = it + 1;
if ((rc = check_piece_hash(hash, fd, part.position, part.length)) != 1) {
info_printf(_("Piece %d/%d not OK - requeuing\n"), it + 1, wget_vector_size(metalink->pieces));
wget_vector_add(job->parts, &part, sizeof(PART));
debug_printf(" need to download %llu bytes from pos=%llu\n",
(unsigned long long)part.length, (unsigned long long)part.position);
}
part.position += part.length;
fsize -= piece->length;
}
close(fd);
} else {
for (int it = 0; it < wget_vector_size(metalink->pieces); it++) {
wget_metalink_piece_t *piece = wget_vector_get(metalink->pieces, it);
if (fsize >= piece->length) {
part.length = piece->length;
} else {
part.length = fsize;
}
part.id = it + 1;
wget_vector_add(job->parts, &part, sizeof(PART));
part.position += part.length;
fsize -= piece->length;
}
}
return 0;
}
static void html_parse_localfile(const char *fname)
{
char *data;
const char *encoding = NULL;
size_t len;
if ((data = wget_read_file(fname, &len))) {
if ((unsigned char)data[0] == 0xFE && (unsigned char)data[1] == 0xFF) {
// Big-endian UTF-16
encoding = "UTF-16BE";
// adjust behind BOM, ignore trailing single byte
data += 2;
len -= 2;
} else if ((unsigned char)data[0] == 0xFF && (unsigned char)data[1] == 0xFE) {
// Little-endian UTF-16
encoding = "UTF-16LE";
// adjust behind BOM
data += 2;
len -= 2;
} else if ((unsigned char)data[0] == 0xEF && (unsigned char)data[1] == 0xBB && (unsigned char)data[2] == 0xBF) {
// UTF-8
encoding = "UTF-8";
// adjust behind BOM
data += 3;
len -= 3;
}
if (encoding)
printf("URI encoding '%s' set by BOM\n", encoding);
if (!wget_strncasecmp_ascii(encoding, "UTF-16", 6)) {
size_t n;
char *utf8;
len -= len & 1; // ignore single trailing byte, else charset conversion fails
if (wget_memiconv(encoding, data, len, "UTF-8", &utf8, &n) == 0) {
printf("Convert non-ASCII encoding '%s' to UTF-8\n", encoding);
data = utf8;
} else {
printf("Failed to convert non-ASCII encoding '%s' to UTF-8, skip parsing\n", encoding);
return;
}
}
WGET_HTML_PARSED_RESULT *res = wget_html_get_urls_inline(data, NULL, NULL);
if (encoding) {
if (res->encoding && wget_strcasecmp_ascii(encoding, res->encoding))
printf("Encoding '%s' as stated in document has been ignored\n", encoding);
}
for (int it = 0; it < wget_vector_size(res->uris); it++) {
WGET_HTML_PARSED_URL *html_url = wget_vector_get(res->uris, it);
wget_string_t *url = &html_url->url;
printf(" %s.%s '%.*s'\n", html_url->dir, html_url->attr, (int) url->len, url->p);
}
wget_xfree(data);
wget_html_free_urls_inline(&res);
}
}
static void test_parse_challenge(void)
{
static const struct test_data {
const char *
input;
const char *
scheme[3];
} test_data[] = {
{ // simplebasic
"Basic realm=\"foo\"",
{ "Basic", NULL }
},
{ // simplebasicucase
"BASIC REALM=\"foo\"",
{ "Basic", NULL }
},
{ // simplebasicucase
"Basic , realm=\"foo\"",
{ "Basic", NULL }
},
{ //
"Basic realm=\"test realm\"",
{ "Basic", NULL }
},
{ //
"Basic realm=\"test-äöÜ\"",
{ "Basic", NULL }
},
{ //
"Basic realm=\"basic\", Newauth realm=\"newauth\"",
{ "Basic", "Newauth", NULL }
},
};
wget_vector_t *challenges;
wget_http_challenge_t *challenge;
// Testcases found here http://greenbytes.de/tech/tc/httpauth/
challenges = wget_vector_create(2, 2, NULL);
wget_vector_set_destructor(challenges, (void(*)(void *))wget_http_free_challenge);
for (unsigned it = 0; it < countof(test_data); it++) {
const struct test_data *t = &test_data[it];
wget_http_parse_challenges(t->input, challenges);
for (unsigned nchal = 0; nchal < countof(test_data[0].scheme) && t->scheme[nchal]; nchal++) {
challenge = wget_vector_get(challenges, nchal);
if (!t->scheme[nchal]) {
if (challenge) {
failed++;
info_printf("Failed [%u]: wget_http_parse_challenges(%s) found %d challenges (expected %u)\n", it, t->input, wget_vector_size(challenges), nchal);
}
break;
}
if (!challenge) {
failed++;
info_printf("Failed [%u]: wget_http_parse_challenges(%s) did not find enough challenges\n", it, t->input);
break;
}
if (!wget_strcasecmp_ascii(challenge->auth_scheme, t->scheme[nchal])) {
ok++;
} else {
failed++;
info_printf("Failed [%u]: wget_http_parse_challenges(%s) -> '%s' (expected '%s')\n", it, t->input, challenge->auth_scheme, t->scheme[nchal]);
}
}
wget_vector_clear(challenges);
}
wget_http_free_challenges(&challenges);
}