/**
* Callback to determine if a node is a linking element whose target has been
* visited.
*
* \param pw HTML document
* \param node DOM node
* \param match Pointer to location to receive result
* \return CSS_OK.
*
* \post \a match will contain true if the node matches and false otherwise.
*/
css_error node_is_visited(void *pw, void *node, bool *match)
{
nscss_select_ctx *ctx = pw;
nsurl *url;
nserror error;
const struct url_data *data;
dom_exception exc;
dom_node *n = node;
dom_string *s = NULL;
*match = false;
exc = dom_node_get_node_name(n, &s);
if ((exc != DOM_NO_ERR) || (s == NULL)) {
return CSS_NOMEM;
}
if (!dom_string_caseless_lwc_isequal(s, corestring_lwc_a)) {
/* Can't be visited; not ancher element */
dom_string_unref(s);
return CSS_OK;
}
/* Finished with node name string */
dom_string_unref(s);
s = NULL;
exc = dom_element_get_attribute(n, corestring_dom_href, &s);
if ((exc != DOM_NO_ERR) || (s == NULL)) {
/* Can't be visited; not got a URL */
return CSS_OK;
}
/* Make href absolute */
/* TODO: this duplicates what we do for box->href
* should we put the absolute URL on the dom node? */
error = nsurl_join(ctx->base_url, dom_string_data(s), &url);
/* Finished with href string */
dom_string_unref(s);
if (error != NSERROR_OK) {
/* Couldn't make nsurl object */
return CSS_NOMEM;
}
data = urldb_get_url_data(url);
/* Visited if in the db and has
* non-zero visit count */
if (data != NULL && data->visits > 0)
*match = true;
nsurl_unref(url);
return CSS_OK;
}
END_TEST
/**
* check join asserts on NULL parameter
*/
START_TEST(nsurl_api_assert_join1_test)
{
const char *rel = "moo";
nsurl *res;
nserror err;
err = nsurl_join(NULL, rel, &res);
ck_assert(err != NSERROR_OK);
}
static bool save_complete_rewrite_url_value(save_complete_ctx *ctx,
const char *value, size_t value_len)
{
nsurl *url;
hlcache_handle *content;
char *escaped;
nserror error;
utf8_convert_ret ret;
error = nsurl_join(ctx->base, value, &url);
if (error == NSERROR_NOMEM)
return false;
if (url != NULL) {
content = save_complete_ctx_find_content(ctx, url);
if (content != NULL) {
/* found a match */
nsurl_unref(url);
fprintf(ctx->fp, "\"%p\"", content);
} else {
/* no match found */
ret = utf8_to_html(nsurl_access(url), "UTF-8",
nsurl_length(url), &escaped);
nsurl_unref(url);
if (ret != UTF8_CONVERT_OK)
return false;
fprintf(ctx->fp, "\"%s\"", escaped);
free(escaped);
}
} else {
ret = utf8_to_html(value, "UTF-8", value_len, &escaped);
if (ret != UTF8_CONVERT_OK)
return false;
fprintf(ctx->fp, "\"%s\"", escaped);
free(escaped);
}
return true;
}
END_TEST
/**
* check join asserts on NULL parameter
*/
START_TEST(nsurl_api_assert_join2_test)
{
nsurl *url;
nsurl *res;
nserror err;
err = nsurl_create(base_str, &url);
ck_assert(err == NSERROR_OK);
err = nsurl_join(url, NULL, &res);
ck_assert(err != NSERROR_OK);
nsurl_unref(url);
}
/**
* URL resolution callback for libcss
*
* \param pw Resolution context
* \param base Base URI
* \param rel Relative URL
* \param abs Pointer to location to receive resolved URL
* \return CSS_OK on success,
* CSS_NOMEM on memory exhaustion,
* CSS_INVALID if resolution failed.
*/
css_error nscss_resolve_url(void *pw, const char *base,
lwc_string *rel, lwc_string **abs)
{
lwc_error lerror;
nserror error;
nsurl *nsbase;
nsurl *nsabs;
/* Create nsurl from base */
/* TODO: avoid this */
error = nsurl_create(base, &nsbase);
if (error != NSERROR_OK) {
return error == NSERROR_NOMEM ? CSS_NOMEM : CSS_INVALID;
}
/* Resolve URI */
error = nsurl_join(nsbase, lwc_string_data(rel), &nsabs);
if (error != NSERROR_OK) {
nsurl_unref(nsbase);
return error == NSERROR_NOMEM ? CSS_NOMEM : CSS_INVALID;
}
nsurl_unref(nsbase);
/* Intern it */
lerror = lwc_intern_string(nsurl_access(nsabs),
nsurl_length(nsabs), abs);
if (lerror != lwc_error_ok) {
*abs = NULL;
nsurl_unref(nsabs);
return lerror == lwc_error_oom ? CSS_NOMEM : CSS_INVALID;
}
nsurl_unref(nsabs);
return CSS_OK;
}
/**
* Test nsurl
*/
int main(void)
{
nsurl *base;
nsurl *joined;
char *string;
size_t len;
const char *url;
const struct test_pairs *test;
int passed = 0;
int count = 0;
/* Create base URL */
if (nsurl_create("http://a/b/c/d;p?q", &base) != NSERROR_OK) {
assert(0 && "Failed to create base URL.");
}
if (nsurl_get(base, NSURL_WITH_FRAGMENT, &string, &len) != NSERROR_OK) {
LOG(("Failed to get string"));
} else {
LOG(("Testing nsurl_join with base %s", string));
free(string);
}
for (test = join_tests; test->test != NULL; test++) {
if (nsurl_join(base, test->test, &joined) != NSERROR_OK) {
LOG(("Failed to join test URL."));
} else {
if (nsurl_get(joined, NSURL_WITH_FRAGMENT,
&string, &len) !=
NSERROR_OK) {
LOG(("Failed to get string"));
} else {
if (strcmp(test->res, string) == 0) {
LOG(("\tPASS: \"%s\"\t--> %s",
test->test,
string));
passed++;
} else {
LOG(("\tFAIL: \"%s\"\t--> %s",
test->test,
string));
LOG(("\t\tExpecting: %s",
test->res));
}
free(string);
}
nsurl_unref(joined);
}
count++;
}
nsurl_unref(base);
/* Create tests */
LOG(("Testing nsurl_create"));
for (test = create_tests; test->test != NULL; test++) {
if (nsurl_create(test->test, &base) != NSERROR_OK) {
LOG(("Failed to create URL:\n\t\t%s.", test->test));
} else {
if (strcmp(nsurl_access(base), test->res) == 0) {
LOG(("\tPASS: \"%s\"\t--> %s",
test->test, nsurl_access(base)));
passed++;
} else {
LOG(("\tFAIL: \"%s\"\t--> %s",
test->test, nsurl_access(base)));
LOG(("\t\tExpecting %s", test->res));
}
nsurl_unref(base);
}
count++;
}
if (passed == count) {
LOG(("Testing complete: SUCCESS"));
} else {
LOG(("Testing complete: FAILURE"));
LOG(("Failed %d out of %d", count - passed, count));
}
return 0;
}
bool html_css_process_link(html_content *htmlc, dom_node *node)
{
dom_string *rel, *type_attr, *media, *href;
struct html_stylesheet *stylesheets;
nsurl *joined;
dom_exception exc;
nserror ns_error;
hlcache_child_context child;
/* rel=<space separated list, including 'stylesheet'> */
exc = dom_element_get_attribute(node, corestring_dom_rel, &rel);
if (exc != DOM_NO_ERR || rel == NULL)
return true;
if (strcasestr(dom_string_data(rel), "stylesheet") == 0) {
dom_string_unref(rel);
return true;
} else if (strcasestr(dom_string_data(rel), "alternate") != 0) {
/* Ignore alternate stylesheets */
dom_string_unref(rel);
return true;
}
dom_string_unref(rel);
/* type='text/css' or not present */
exc = dom_element_get_attribute(node, corestring_dom_type, &type_attr);
if (exc == DOM_NO_ERR && type_attr != NULL) {
if (!dom_string_caseless_lwc_isequal(type_attr,
corestring_lwc_text_css)) {
dom_string_unref(type_attr);
return true;
}
dom_string_unref(type_attr);
}
/* media contains 'screen' or 'all' or not present */
exc = dom_element_get_attribute(node, corestring_dom_media, &media);
if (exc == DOM_NO_ERR && media != NULL) {
if (strcasestr(dom_string_data(media), "screen") == NULL &&
strcasestr(dom_string_data(media), "all") == NULL) {
dom_string_unref(media);
return true;
}
dom_string_unref(media);
}
/* href='...' */
exc = dom_element_get_attribute(node, corestring_dom_href, &href);
if (exc != DOM_NO_ERR || href == NULL)
return true;
/* TODO: only the first preferred stylesheets (ie.
* those with a title attribute) should be loaded
* (see HTML4 14.3) */
ns_error = nsurl_join(htmlc->base_url, dom_string_data(href), &joined);
if (ns_error != NSERROR_OK) {
dom_string_unref(href);
goto no_memory;
}
dom_string_unref(href);
LOG("linked stylesheet %i '%s'", htmlc->stylesheet_count, nsurl_access(joined));
/* extend stylesheets array to allow for new sheet */
stylesheets = realloc(htmlc->stylesheets,
sizeof(struct html_stylesheet) *
(htmlc->stylesheet_count + 1));
if (stylesheets == NULL) {
nsurl_unref(joined);
ns_error = NSERROR_NOMEM;
goto no_memory;
}
htmlc->stylesheets = stylesheets;
htmlc->stylesheets[htmlc->stylesheet_count].node = NULL;
htmlc->stylesheets[htmlc->stylesheet_count].modified = false;
/* start fetch */
child.charset = htmlc->encoding;
child.quirks = htmlc->base.quirks;
ns_error = hlcache_handle_retrieve(joined, 0,
content_get_url(&htmlc->base),
NULL, html_convert_css_callback,
htmlc, &child, CONTENT_CSS,
&htmlc->stylesheets[htmlc->stylesheet_count].sheet);
nsurl_unref(joined);
if (ns_error != NSERROR_OK)
goto no_memory;
htmlc->stylesheet_count++;
htmlc->base.active++;
LOG("%d fetches active", htmlc->base.active);
return true;
no_memory:
content_broadcast_errorcode(&htmlc->base, ns_error);
return false;
}
/**
* process a script with a src tag
*/
static dom_hubbub_error
exec_src_script(html_content *c,
dom_node *node,
dom_string *mimetype,
dom_string *src)
{
nserror ns_error;
nsurl *joined;
hlcache_child_context child;
struct html_script *nscript;
union content_msg_data msg_data;
bool async;
bool defer;
enum html_script_type script_type;
hlcache_handle_callback script_cb;
dom_hubbub_error ret = DOM_HUBBUB_OK;
dom_exception exc; /* returned by libdom functions */
/* src url */
ns_error = nsurl_join(c->base_url, dom_string_data(src), &joined);
if (ns_error != NSERROR_OK) {
msg_data.error = messages_get("NoMemory");
content_broadcast(&c->base, CONTENT_MSG_ERROR, msg_data);
return DOM_HUBBUB_NOMEM;
}
LOG(("script %i '%s'", c->scripts_count, nsurl_access(joined)));
/* there are three ways to process the script tag at this point:
*
* Syncronously pause the parent parse and continue after
* the script has downloaded and executed. (default)
* Async Start the script downloading and execute it when it
* becomes available.
* Defered Start the script downloading and execute it when
* the page has completed parsing, may be set along
* with async where it is ignored.
*/
/* we interpret the presence of the async and defer attribute
* as true and ignore its value, technically only the empty
* value or the attribute name itself are valid. However
* various browsers interpret this in various ways the most
* compatible approach is to be liberal and accept any
* value. Note setting the values to "false" still makes them true!
*/
exc = dom_element_has_attribute(node, corestring_dom_async, &async);
if (exc != DOM_NO_ERR) {
return DOM_HUBBUB_OK; /* dom error */
}
if (async) {
/* asyncronous script */
script_type = HTML_SCRIPT_ASYNC;
script_cb = convert_script_async_cb;
} else {
exc = dom_element_has_attribute(node,
corestring_dom_defer, &defer);
if (exc != DOM_NO_ERR) {
return DOM_HUBBUB_OK; /* dom error */
}
if (defer) {
/* defered script */
script_type = HTML_SCRIPT_DEFER;
script_cb = convert_script_defer_cb;
} else {
/* syncronous script */
script_type = HTML_SCRIPT_SYNC;
script_cb = convert_script_sync_cb;
}
}
nscript = html_process_new_script(c, mimetype, script_type);
if (nscript == NULL) {
nsurl_unref(joined);
msg_data.error = messages_get("NoMemory");
content_broadcast(&c->base, CONTENT_MSG_ERROR, msg_data);
return DOM_HUBBUB_NOMEM;
}
/* set up child fetch encoding and quirks */
child.charset = c->encoding;
child.quirks = c->base.quirks;
ns_error = hlcache_handle_retrieve(joined,
0,
content_get_url(&c->base),
NULL,
script_cb,
c,
&child,
CONTENT_SCRIPT,
&nscript->data.handle);
nsurl_unref(joined);
if (ns_error != NSERROR_OK) {
/* @todo Deal with fetch error better. currently assume
* fetch never became active
*/
/* mark duff script fetch as already started */
nscript->already_started = true;
LOG(("Fetch failed with error %d",ns_error));
} else {
/* update base content active fetch count */
c->base.active++;
LOG(("%d fetches active", c->base.active));
switch (script_type) {
case HTML_SCRIPT_SYNC:
ret = DOM_HUBBUB_HUBBUB_ERR | HUBBUB_PAUSED;
case HTML_SCRIPT_ASYNC:
break;
case HTML_SCRIPT_DEFER:
break;
default:
assert(0);
}
}
return ret;
}
static char *save_complete_rewrite_stylesheet_urls(save_complete_ctx *ctx,
const char *source, unsigned long size, const nsurl *base,
unsigned long *osize)
{
char *rewritten;
unsigned long offset = 0;
unsigned int imports = 0;
nserror error;
/* count number occurrences of @import to (over)estimate result size */
/* can't use strstr because source is not 0-terminated string */
for (offset = 0; SLEN("@import") < size &&
offset <= size - SLEN("@import"); offset++) {
if (source[offset] == '@' &&
tolower(source[offset + 1]) == 'i' &&
tolower(source[offset + 2]) == 'm' &&
tolower(source[offset + 3]) == 'p' &&
tolower(source[offset + 4]) == 'o' &&
tolower(source[offset + 5]) == 'r' &&
tolower(source[offset + 6]) == 't')
imports++;
}
rewritten = malloc(size + imports * 20);
if (rewritten == NULL)
return NULL;
*osize = 0;
offset = 0;
while (offset < size) {
const char *import_url = NULL;
char *import_url_copy;
int import_url_len = 0;
nsurl *url = NULL;
regmatch_t match[11];
int m = regexec(&save_complete_import_re, source + offset,
11, match, 0);
if (m)
break;
if (match[2].rm_so != -1) {
import_url = source + offset + match[2].rm_so;
import_url_len = match[2].rm_eo - match[2].rm_so;
} else if (match[4].rm_so != -1) {
import_url = source + offset + match[4].rm_so;
import_url_len = match[4].rm_eo - match[4].rm_so;
} else if (match[6].rm_so != -1) {
import_url = source + offset + match[6].rm_so;
import_url_len = match[6].rm_eo - match[6].rm_so;
} else if (match[8].rm_so != -1) {
import_url = source + offset + match[8].rm_so;
import_url_len = match[8].rm_eo - match[8].rm_so;
} else if (match[10].rm_so != -1) {
import_url = source + offset + match[10].rm_so;
import_url_len = match[10].rm_eo - match[10].rm_so;
}
assert(import_url != NULL);
import_url_copy = strndup(import_url, import_url_len);
if (import_url_copy == NULL) {
free(rewritten);
return NULL;
}
error = nsurl_join(base, import_url_copy, &url);
free(import_url_copy);
if (error == NSERROR_NOMEM) {
free(rewritten);
return NULL;
}
/* copy data before match */
memcpy(rewritten + *osize, source + offset, match[0].rm_so);
*osize += match[0].rm_so;
if (url != NULL) {
hlcache_handle *content;
content = save_complete_ctx_find_content(ctx, url);
if (content != NULL) {
/* replace import */
char buf[64];
snprintf(buf, sizeof buf, "@import '%p'",
content);
memcpy(rewritten + *osize, buf, strlen(buf));
*osize += strlen(buf);
} else {
/* copy import */
memcpy(rewritten + *osize,
source + offset + match[0].rm_so,
match[0].rm_eo - match[0].rm_so);
*osize += match[0].rm_eo - match[0].rm_so;
}
nsurl_unref(url);
} else {
/* copy import */
memcpy(rewritten + *osize,
source + offset + match[0].rm_so,
match[0].rm_eo - match[0].rm_so);
*osize += match[0].rm_eo - match[0].rm_so;
}
assert(0 < match[0].rm_eo);
offset += match[0].rm_eo;
}
/* copy rest of source */
if (offset < size) {
memcpy(rewritten + *osize, source + offset, size - offset);
*osize += size - offset;
}
return rewritten;
}
