X-Git-Url: http://lists.indexdata.com/cgi-bin?a=blobdiff_plain;f=src%2Ffilter_http_rewrite.cpp;h=d417f46cead1f552277b757294e2315554dac812;hb=1751c7afd96e1984f5710284cdcb245a93d9ce6c;hp=e69aba649f8bc2610107b3b4917005b57fb14211;hpb=3ef4df94516a136b7ee18ec8a45e740ef9e9dc05;p=metaproxy-moved-to-github.git diff --git a/src/filter_http_rewrite.cpp b/src/filter_http_rewrite.cpp index e69aba6..d417f46 100644 --- a/src/filter_http_rewrite.cpp +++ b/src/filter_http_rewrite.cpp @@ -26,6 +26,7 @@ Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA #include #include +#include #include #include #include @@ -39,15 +40,13 @@ namespace metaproxy_1 { namespace filter { class HttpRewrite::Replace { public: - std::string regex; + boost::regex re; + boost::smatch what; std::string recipe; std::map group_index; - const std::string search_replace( - std::map & vars, - const std::string & txt) const; - std::string sub_vars ( + std::string sub_vars( const std::map & vars) const; - void parse_groups(); + void parse_groups(std::string pattern); }; class HttpRewrite::Rule { @@ -55,7 +54,7 @@ namespace metaproxy_1 { std::list replace_list; const std::string test_patterns( std::map & vars, - const std::string & txt) const; + const std::string & txt); }; class HttpRewrite::Within { public: @@ -68,7 +67,9 @@ namespace metaproxy_1 { class HttpRewrite::Phase { public: + Phase(); std::list within_list; + int m_verbose; void rewrite_reqline(mp::odr & o, Z_HTTP_Request *hreq, std::map & vars) const; void rewrite_headers(mp::odr & o, Z_HTTP_Header *headers, @@ -82,12 +83,13 @@ namespace metaproxy_1 { void anyTagEnd(const char *tag, int tag_len, int close_it); void attribute(const char *tag, int tag_len, const char *attr, int attr_len, - const char *value, int val_len); + const char *value, int val_len, + const char *sep); void closeTag(const char *tag, int tag_len); void text(const char *value, int len); const Phase *m_phase; WRBUF m_w; - std::list::const_iterator enabled_within; + std::stack::const_iterator> s_within; std::map &m_vars; public: Event(const Phase *p, std::map &vars); @@ -154,8 +156,12 @@ void yf::HttpRewrite::Phase::rewrite_reqline (mp::odr & o, else { //TODO what about proto + const char *host = z_HTTP_header_lookup(hreq->headers, "Host"); + if (!host) + return; + path += "http://"; - path += z_HTTP_header_lookup(hreq->headers, "Host"); + path += host; path += hreq->path; } @@ -225,6 +231,9 @@ void yf::HttpRewrite::Phase::rewrite_body(mp::odr & o, HTMLParser parser; Event ev(this, vars); + + parser.set_verbose(m_verbose); + std::string buf(*content_buf, *content_len); parser.parse(ev, buf.c_str()); @@ -239,7 +248,6 @@ yf::HttpRewrite::Event::Event(const Phase *p, ) : m_phase(p), m_vars(vars) { m_w = wrbuf_alloc(); - enabled_within = m_phase->within_list.end(); } yf::HttpRewrite::Event::~Event() @@ -254,23 +262,29 @@ const char *yf::HttpRewrite::Event::result() void yf::HttpRewrite::Event::openTagStart(const char *tag, int tag_len) { - // check if there is - if (enabled_within == m_phase->within_list.end()) + wrbuf_putc(m_w, '<'); + wrbuf_write(m_w, tag, tag_len); + + std::string t(tag, tag_len); + std::list::const_iterator it = m_phase->within_list.begin(); + for (; it != m_phase->within_list.end(); it++) { - std::string t(tag, tag_len); - std::list::const_iterator it = - m_phase->within_list.begin(); - for (; it != m_phase->within_list.end(); it++) + if (it->tag.length() > 0 && yaz_strcasecmp(it->tag.c_str(), + t.c_str()) == 0) { - if (it->tag.length() > 0 && yaz_strcasecmp(it->tag.c_str(), - t.c_str()) == 0) + std::vector attr; + boost::split(attr, it->attr, boost::is_any_of(",")); + size_t i; + for (i = 0; i < attr.size(); i++) { - enabled_within = it; + if (attr[i].compare("#text") == 0) + { + s_within.push(it); + return; + } } } } - wrbuf_putc(m_w, '<'); - wrbuf_write(m_w, tag, tag_len); } void yf::HttpRewrite::Event::anyTagEnd(const char *tag, int tag_len, @@ -278,14 +292,12 @@ void yf::HttpRewrite::Event::anyTagEnd(const char *tag, int tag_len, { if (close_it) { - std::list::const_iterator it = enabled_within; - if (it != m_phase->within_list.end()) + if (!s_within.empty()) { + std::list::const_iterator it = s_within.top(); std::string t(tag, tag_len); if (yaz_strcasecmp(it->tag.c_str(), t.c_str()) == 0) - { - enabled_within = m_phase->within_list.end(); - } + s_within.pop(); } } if (close_it) @@ -295,7 +307,8 @@ void yf::HttpRewrite::Event::anyTagEnd(const char *tag, int tag_len, void yf::HttpRewrite::Event::attribute(const char *tag, int tag_len, const char *attr, int attr_len, - const char *value, int val_len) + const char *value, int val_len, + const char *sep) { std::list::const_iterator it = m_phase->within_list.begin(); bool subst = false; @@ -323,31 +336,33 @@ void yf::HttpRewrite::Event::attribute(const char *tag, int tag_len, wrbuf_putc(m_w, ' '); wrbuf_write(m_w, attr, attr_len); - wrbuf_puts(m_w, "=\""); - - std::string output; - if (subst) + if (value) { - std::string input(value, val_len); - output = it->rule->test_patterns(m_vars, input); + wrbuf_puts(m_w, "="); + wrbuf_puts(m_w, sep); + + std::string output; + if (subst) + { + std::string input(value, val_len); + output = it->rule->test_patterns(m_vars, input); + } + if (output.empty()) + wrbuf_write(m_w, value, val_len); + else + wrbuf_puts(m_w, output.c_str()); + wrbuf_puts(m_w, sep); } - if (output.empty()) - wrbuf_write(m_w, value, val_len); - else - wrbuf_puts(m_w, output.c_str()); - wrbuf_puts(m_w, "\""); } void yf::HttpRewrite::Event::closeTag(const char *tag, int tag_len) { - std::list::const_iterator it = enabled_within; - if (it != m_phase->within_list.end()) + if (!s_within.empty()) { + std::list::const_iterator it = s_within.top(); std::string t(tag, tag_len); if (yaz_strcasecmp(it->tag.c_str(), t.c_str()) == 0) - { - enabled_within = m_phase->within_list.end(); - } + s_within.pop(); } wrbuf_puts(m_w, "::const_iterator it = enabled_within; - bool subst = false; - - if (it != m_phase->within_list.end()) - { - subst = true; - if (it->attr.length() > 0) - { - subst = false; - std::vector 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") == 0) - { - subst = true; - } - } - } - } + std::list::const_iterator it = m_phase->within_list.end(); + if (!s_within.empty()) + it = s_within.top(); std::string output; - if (subst) + if (it != m_phase->within_list.end()) { std::string input(value, len); output = it->rule->test_patterns(m_vars, input); @@ -388,70 +385,65 @@ void yf::HttpRewrite::Event::text(const char *value, int len) wrbuf_puts(m_w, output.c_str()); } - -/** - * Tests pattern from the vector in order and executes recipe on - the first match. - */ const std::string yf::HttpRewrite::Rule::test_patterns( std::map & vars, - const std::string & txt) const + const std::string & txt) { - std::list::const_iterator it = replace_list.begin(); - - for (; it != replace_list.end(); it++) - { - std::string out = it->search_replace(vars, txt); - if (!out.empty()) return out; - } - return ""; -} - -const std::string yf::HttpRewrite::Replace::search_replace( - std::map & vars, - const std::string & txt) const -{ - //exec regex against value - boost::regex re(regex); - boost::smatch what; + std::string out; std::string::const_iterator start, end; start = txt.begin(); end = txt.end(); - std::string out; - while (regex_search(start, end, what, re)) //find next full match + while (1) { + std::list::iterator bit = replace_list.end(); + { + std::string::const_iterator best_pos = txt.end(); + std::list::iterator it = replace_list.begin(); + for (; it != replace_list.end(); it++) + { + if (regex_search(start, end, it->what, it->re)) + { + if (it->what[0].first < best_pos) + { + best_pos = it->what[0].first; + bit = it; + } + } + } + if (bit == replace_list.end()) + break; + } + size_t i; - for (i = 1; i < what.size(); ++i) + for (i = 1; i < bit->what.size(); ++i) { //check if the group is named - std::map::const_iterator it - = group_index.find(i); - if (it != group_index.end()) + std::map::const_iterator git + = bit->group_index.find(i); + if (git != bit->group_index.end()) { //it is - if (!what[i].str().empty()) - vars[it->second] = what[i]; + vars[git->second] = bit->what[i]; } } //prepare replacement string - std::string rvalue = sub_vars(vars); + std::string rvalue = bit->sub_vars(vars); yaz_log(YLOG_LOG, "! Rewritten '%s' to '%s'", - what.str(0).c_str(), rvalue.c_str()); - out.append(start, what[0].first); + bit->what.str(0).c_str(), rvalue.c_str()); + out.append(start, bit->what[0].first); out.append(rvalue); - start = what[0].second; //move search forward + start = bit->what[0].second; //move search forward } - //if we had a match cat the last part if (start != txt.begin()) out.append(start, end); return out; } -void yf::HttpRewrite::Replace::parse_groups() +void yf::HttpRewrite::Replace::parse_groups(std::string pattern) { int gnum = 0; bool esc = false; - const std::string & str = regex; + const std::string &str = pattern; std::string res; yaz_log(YLOG_LOG, "Parsing groups from '%s'", str.c_str()); for (size_t i = 0; i < str.size(); ++i) @@ -507,11 +499,11 @@ void yf::HttpRewrite::Replace::parse_groups() } esc = false; } - regex = res; + re = res; } -std::string yf::HttpRewrite::Replace::sub_vars ( - const std::map & vars) const +std::string yf::HttpRewrite::Replace::sub_vars( + const std::map & vars) const { std::string out; bool esc = false; @@ -560,9 +552,19 @@ std::string yf::HttpRewrite::Replace::sub_vars ( return out; } +yf::HttpRewrite::Phase::Phase() : m_verbose(0) +{ +} void yf::HttpRewrite::configure_phase(const xmlNode *ptr, Phase &phase) { + static const char *names[2] = { "verbose", 0 }; + std::string values[1]; + values[0] = "0"; + mp::xml::parse_attr(ptr, names, values); + + phase.m_verbose = atoi(values[0].c_str()); + std::map rules; for (ptr = ptr->children; ptr; ptr = ptr->next) { @@ -583,11 +585,12 @@ void yf::HttpRewrite::configure_phase(const xmlNode *ptr, Phase &phase) if (!strcmp((const char *) p->name, "rewrite")) { Replace replace; + std::string from; const struct _xmlAttr *attr; for (attr = p->properties; attr; attr = attr->next) { if (!strcmp((const char *) attr->name, "from")) - replace.regex = mp::xml::get_text(attr->children); + from = mp::xml::get_text(attr->children); else if (!strcmp((const char *) attr->name, "to")) replace.recipe = mp::xml::get_text(attr->children); else @@ -597,10 +600,12 @@ void yf::HttpRewrite::configure_phase(const xmlNode *ptr, Phase &phase) + " in rewrite section of http_rewrite"); } yaz_log(YLOG_LOG, "Found rewrite rule from '%s' to '%s'", - replace.regex.c_str(), replace.recipe.c_str()); - replace.parse_groups(); - if (!replace.regex.empty()) + from.c_str(), replace.recipe.c_str()); + if (!from.empty()) + { + replace.parse_groups(from); rule->replace_list.push_back(replace); + } } else throw mp::filter::FilterException