};
class HttpRewrite::Within {
public:
- std::string header;
- std::string attr;
- std::string tag;
+ boost::regex header;
+ boost::regex attr;
+ boost::regex tag;
std::string type;
bool reqline;
RulePtr rule;
std::list<Within>::const_iterator it = cit->within_list.begin();
for (; it != cit->within_list.end(); it++)
{
- if (it->header.length() > 0 &&
- yaz_strcasecmp(it->header.c_str(), header->name) == 0)
+ if (!it->header.empty() &&
+ regex_match(header->name, it->header))
{
+#ifdef OLDHEADERMATCH
+ // Matches and replaces the whole header line.
+ // This is good if you want to play with the header name too,
+ // but useless for patterns that want to anchor to the beginning
+ // or end of the header value, as we want to do with host-relative
+ // links. This code should probably be removed.
std::string sheader(header->name);
sheader += ": ";
sheader += header->value;
header->value = odr_strdup(
o, sheader.substr(pos + 2, std::string::npos).c_str());
}
+#else
+ // Match and replace only the header value
+ std::string hval(header->value);
+ if (it->exec(vars, hval, true))
+ {
+ header->value = odr_strdup(o, hval.c_str());
+ }
+
+#endif
}
}
}
{
if (*content_len == 0)
return;
+ if (!content_type) {
+ yaz_log(YLOG_LOG, "rewrite_body: null content_type, can not rewrite");
+ return;
+ }
std::list<Content>::const_iterator cit = content_list.begin();
for (; cit != content_list.end(); cit++)
{
&& regex_match(content_type, cit->content_re))
break;
}
- if (cit == content_list.end())
+ if (cit == content_list.end()) {
+ yaz_log(YLOG_LOG,"rewrite_body: No content rule matched %s, not rewriting",
+ content_type );
return;
+ }
int i;
for (i = 0; i < *content_len; i++)
- if ((*content_buf)[i] == 0)
+ if ((*content_buf)[i] == 0) {
+ yaz_log(YLOG_LOG,"rewrite_body: Looks like binary stuff, not rewriting");
return; // binary content. skip
+ }
std::string content(*content_buf, *content_len);
cit->parse(m_verbose, content, vars);
std::list<Within>::const_iterator it = m_content->within_list.begin();
for (; it != m_content->within_list.end(); it++)
{
- if (it->tag.length() > 0 && yaz_strcasecmp(it->tag.c_str(),
- t.c_str()) == 0)
+ if (!it->tag.empty() && regex_match(t, it->tag))
{
- std::vector<std::string> attr;
- boost::split(attr, it->attr, boost::is_any_of(","));
- size_t i;
- for (i = 0; i < attr.size(); i++)
+ if (!it->attr.empty() && regex_match("#text", it->attr))
{
- if (attr[i].compare("#text") == 0)
- {
- s_within.push(it);
- return;
- }
+ s_within.push(it);
+ return;
}
}
}
{
std::list<Within>::const_iterator it = s_within.top();
std::string t(tag, tag_len);
- if (yaz_strcasecmp(it->tag.c_str(), t.c_str()) == 0)
+ if (regex_match(t, it->tag))
s_within.pop();
}
}
for (; it != m_content->within_list.end(); it++)
{
std::string t(tag, tag_len);
- if (it->tag.length() == 0 ||
- yaz_strcasecmp(it->tag.c_str(), t.c_str()) == 0)
+ if (it->tag.empty() || regex_match(t, it->tag))
{
std::string a(attr, attr_len);
- std::vector<std::string> attr;
- boost::split(attr, it->attr, boost::is_any_of(","));
- size_t i;
- for (i = 0; i < attr.size(); i++)
- {
- if (attr[i].compare("#text") &&
- yaz_strcasecmp(attr[i].c_str(), a.c_str()) == 0)
- subst = true;
- }
+ if (!it->attr.empty() && regex_match(a, it->attr))
+ subst = true;
}
if (subst)
break;
{
std::list<Within>::const_iterator it = s_within.top();
std::string t(tag, tag_len);
- if (yaz_strcasecmp(it->tag.c_str(), t.c_str()) == 0)
+ if (regex_match(t, it->tag))
s_within.pop();
}
wrbuf_puts(m_w, "</");
std::string values[6];
mp::xml::parse_attr(ptr, names, values);
Within w;
- w.header = values[0];
- w.attr = values[1];
- w.tag = values[2];
- std::map<std::string,RulePtr>::const_iterator it =
- rules.find(values[3]);
- if (it == rules.end())
+ if (values[0].length() > 0)
+ w.header.assign(values[0], boost::regex_constants::icase);
+ if (values[1].length() > 0)
+ w.attr.assign(values[1], boost::regex_constants::icase);
+ if (values[2].length() > 0)
+ w.tag.assign(values[2], boost::regex_constants::icase);
+
+ std::vector<std::string> rulenames;
+ boost::split(rulenames, values[3], boost::is_any_of(","));
+ if (rulenames.size() == 0)
+ {
throw mp::filter::FilterException
- ("Reference to non-existing rule '" + values[3] +
+ ("Empty rule in '" + values[3] +
"' in http_rewrite filter");
- w.rule = it->second;
+ }
+ else if (rulenames.size() == 1)
+ {
+ std::map<std::string,RulePtr>::const_iterator it =
+ rules.find(rulenames[0]);
+ if (it == rules.end())
+ throw mp::filter::FilterException
+ ("Reference to non-existing rule '" + rulenames[0] +
+ "' in http_rewrite filter");
+ w.rule = it->second;
+
+ }
+ else
+ {
+ RulePtr rule(new Rule);
+ size_t i;
+ for (i = 0; i < rulenames.size(); i++)
+ {
+ std::map<std::string,RulePtr>::const_iterator it =
+ rules.find(rulenames[i]);
+ if (it == rules.end())
+ throw mp::filter::FilterException
+ ("Reference to non-existing rule '" + rulenames[i] +
+ "' in http_rewrite filter");
+ RulePtr subRule = it->second;
+ std::list<Replace>::iterator rit =
+ subRule->replace_list.begin();
+ for (; rit != subRule->replace_list.end(); rit++)
+ rule->replace_list.push_back(*rit);
+ }
+ w.rule = rule;
+ }
w.reqline = values[4] == "1";
w.type = values[5];
if (w.type.empty() || w.type == "quoted-literal")
Content c;
c.type = values[0];
- // if (!values[1].empty())
- c.content_re = values[1];
+ if (!values[1].empty())
+ c.content_re.assign(values[1], boost::regex::icase);
c.configure(ptr->children, rules);
phase.content_list.push_back(c);
}