Make use of butl url encode() and decode() functions

8 files changed, 71 insertions, 127 deletions

diff --git a/mod/build-config.cxx b/mod/build-config.cxx
index aa25511..cbde252 100644
--- a/mod/build-config.cxx
+++ b/mod/build-config.cxx
@@ -118,9 +118,9 @@ namespace brep
     // the package version into the URL path part and so don't encode it.
     //
     string url (host + root.representation () +
-                mime_url_encode (b.package_name) + '/' +
+                mime_url_encode (b.package_name, false) + '/' +
                 b.package_version.string () + "/log/" +
-                mime_url_encode (b.configuration) + '/' +
+                mime_url_encode (b.configuration, false) + '/' +
                 b.toolchain_version.string ());
 
     if (op != nullptr)
diff --git a/mod/mod-package-details.cxx b/mod/mod-package-details.cxx
index dce7f3b..ca2b94a 100644
--- a/mod/mod-package-details.cxx
+++ b/mod/mod-package-details.cxx
@@ -75,7 +75,7 @@ handle (request& rq, response& rs)
   const dir_path& root (options_->root ());
 
   const string& name (*rq.path ().rbegin ());
-  const string ename (mime_url_encode (name));
+  const string ename (mime_url_encode (name, false));
 
   params::package_details params;
   bool full;
diff --git a/mod/mod-package-version-details.cxx b/mod/mod-package-version-details.cxx
index 134fb9a..26838f7 100644
--- a/mod/mod-package-version-details.cxx
+++ b/mod/mod-package-version-details.cxx
@@ -130,7 +130,9 @@ handle (request& rq, response& rs)
 
   s <<       DIV(ID="heading")
     <<         H1
-    <<           A(HREF=root / path (mime_url_encode (name))) << name << ~A
+    <<           A(HREF=root / path (mime_url_encode (name, false)))
+    <<             name
+    <<           ~A
     <<           "/"
     <<           A(HREF=url ()) << sver << ~A
     <<         ~H1
@@ -255,7 +257,7 @@ handle (request& rq, response& rs)
 
         const auto& dcon (d.constraint);
         const string& dname (p->id.name);
-        string ename (mime_url_encode (dname));
+        string ename (mime_url_encode (dname, false));
 
         if (r->url)
         {
diff --git a/mod/mod-repository-details.cxx b/mod/mod-repository-details.cxx
index 177f49d..53cd3aa 100644
--- a/mod/mod-repository-details.cxx
+++ b/mod/mod-repository-details.cxx
@@ -99,7 +99,9 @@ handle (request& rq, response& rs)
     //
     string id (html_id (r.name));
     s << H1(ID=id)
-      <<   A(HREF="#" + web::mime_url_encode (id)) << r.display_name << ~A
+      <<   A(HREF="#" + web::mime_url_encode (id, false))
+      <<     r.display_name
+      <<   ~A
       << ~H1;
 
     if (r.summary)
diff --git a/mod/page.cxx b/mod/page.cxx
index a7ec087..566406e 100644
--- a/mod/page.cxx
+++ b/mod/page.cxx
@@ -188,7 +188,7 @@ namespace brep
 
       // Propagate search criteria to the package details page.
       //
-      <<         root_ / path (mime_url_encode (name_)) << query_param_
+      <<         root_ / path (mime_url_encode (name_, false)) << query_param_
 
       <<       ~HREF
       <<          name_
@@ -218,7 +218,7 @@ namespace brep
     else
     {
       assert (root_ != nullptr);
-      s << A(HREF=*root_ / dir_path (mime_url_encode (*package_)) /
+      s << A(HREF=*root_ / dir_path (mime_url_encode (*package_, false)) /
                   path (version_))
         <<   version_;
 
@@ -393,7 +393,7 @@ namespace brep
             ? p->internal_repository.load ()
             : p->other_repositories[0].load ());
 
-          auto en (mime_url_encode (n));
+          auto en (mime_url_encode (n, false));
 
           if (r->url)
             s << A(HREF=*r->url + en) << n << ~A;
@@ -542,7 +542,7 @@ namespace brep
       <<     SPAN(CLASS="value")
       <<       A
       <<       HREF
-      <<         root_ << "?about#" << mime_url_encode (html_id (name_))
+      <<         root_ << "?about#" << mime_url_encode (html_id (name_), false)
       <<       ~HREF
       <<         name_
       <<       ~A
diff --git a/web/apache/request.cxx b/web/apache/request.cxx
index 3393555..51590e3 100644
--- a/web/apache/request.cxx
+++ b/web/apache/request.cxx
@@ -18,7 +18,6 @@
 #include <memory>    // unique_ptr
 #include <string>
 #include <cassert>
-#include <sstream>
 #include <ostream>
 #include <istream>
 #include <cstring>   // str*(), memcpy(), size_t
@@ -353,7 +352,7 @@ namespace web
     {
       if (path_.empty ())
       {
-        path_ = path_type (rec_->uri);
+        path_ = path_type (rec_->uri); // Is already URL-decoded.
 
         // Module request handler can not be called if URI is empty.
         //
@@ -514,10 +513,9 @@ namespace web
     {
       assert (!buffer); // Cookie buffering is not implemented yet.
 
-      ostringstream s;
-      mime_url_encode (name, s);
-      s << "=";
-      mime_url_encode (value, s);
+      string s (mime_url_encode (name));
+      s += "=";
+      s += mime_url_encode (value);
 
       if (max_age)
       {
@@ -528,20 +526,27 @@ namespace web
         //
         char b[100];
         strftime (b, sizeof (b), "%a, %d-%b-%Y %H:%M:%S GMT", gmtime (&t));
-        s << "; Expires=" << b;
+        s += "; Expires=";
+        s += b;
       }
 
       if (path)
-        s << ";Path=" << path;
+      {
+        s += ";Path=";
+        s += path;
+      }
 
       if (domain)
-        s << ";Domain=" << domain;
+      {
+        s += ";Domain=";
+        s += domain;
+      }
 
       if (secure)
-        s << ";Secure";
+        s += ";Secure";
 
       state (request_state::headers);
-      apr_table_add (rec_->err_headers_out, "Set-Cookie", s.str ().c_str ());
+      apr_table_add (rec_->err_headers_out, "Set-Cookie", s.c_str ());
     }
 
     void request::
diff --git a/web/mime-url-encoding.cxx b/web/mime-url-encoding.cxx
index e0022bc..00f8852 100644
--- a/web/mime-url-encoding.cxx
+++ b/web/mime-url-encoding.cxx
@@ -4,132 +4,64 @@
 
 #include <web/mime-url-encoding.hxx>
 
-#include <ios>       // hex, uppercase, right
 #include <string>
-#include <iomanip>   // setw(), setfill()
-#include <ostream>
-#include <sstream>
-#include <cstring>   // size_t, strspn()
-#include <stdexcept> // invalid_argument
+#include <iterator> // back_inserter
+
+#include <libbutl/url.mxx>
 
 using namespace std;
+using namespace butl;
 
 namespace web
 {
-  // Encode characters different from unreserved ones specified in
-  // "2.3. Unreserved Characters" of http://tools.ietf.org/html/rfc3986.
-  //
-  void
-  mime_url_encode (const char* v, ostream& o)
+  inline static bool
+  encode_query (char& c)
   {
-    char f (o.fill ());
-    ostream::fmtflags g (o.flags ());
-    o << hex << uppercase << right << setfill ('0');
-
-    char c;
-    while ((c = *v++) != '\0')
+    if (c == ' ')
     {
-      if ((c >= 'A' && c <= 'Z') || (c >= 'a' && c <= 'z') ||
-          (c >= '0' && c <= '9'))
-      {
-        o << c;
-      }
-      else
-      {
-        switch (c)
-        {
-        case ' ': o << '+'; break;
-        case '.':
-        case '_':
-        case '-':
-        case '~': o << c; break;
-        default:
-          {
-            o << "%" << setw (2) << static_cast<unsigned short> (c);
-            break;
-          }
-        }
-      }
+      c = '+';
+      return false;
     }
 
-    o.flags (g);
-    o.fill (f);
+    return !url::unreserved (c);
   }
 
   string
-  mime_url_encode (const char* v)
+  mime_url_encode (const char* v, bool query)
   {
-    stringstream o;
-    mime_url_encode (v, o);
-    return o.str ();
+    return query ? url::encode (v, encode_query) : url::encode (v);
   }
 
   string
-  mime_url_encode (const string& v)
+  mime_url_encode (const string& v, bool query)
   {
-    return mime_url_encode (v.c_str ());
+    return query ? url::encode (v, encode_query) : url::encode (v);
   }
 
   string
-  mime_url_decode (const char* b, const char* e, bool trim)
+  mime_url_decode (const char* b, const char* e, bool trim, bool query)
   {
     if (trim)
     {
-      b += strspn (b, " ");
+      for (; b != e && *b == ' '; ++b) ;
 
-      if (b >= e)
+      if (b == e)
         return string ();
 
       while (*--e == ' ');
       ++e;
     }
 
-    string value;
-    value.reserve (e - b);
-
-    char bf[3];
-    bf[2] = '\0';
-
-    while (b != e)
-    {
-      char c (*b++);
-      switch (c)
-      {
-      case '+':
-        {
-          value.append (" ");
-          break;
-        }
-      case '%':
-        {
-          if (*b == '\0' || b[1] == '\0')
-          {
-            throw invalid_argument ("::web::mime_url_decode short");
-          }
-
-          *bf = *b;
-          bf[1] = b[1];
-
-          char* ebf (nullptr);
-          size_t vl (strtoul (bf, &ebf, 16));
-
-          if (*ebf != '\0')
-          {
-            throw invalid_argument ("::web::mime_url_decode wrong");
-          }
-
-          value.append (1, static_cast<char> (vl));
-          b += 2;
-          break;
-        }
-      default:
-        {
-          value.append (1, c);
-          break;
-        }
-      }
-    }
-
-    return value;
+    string r;
+    if (!query)
+      url::decode (b, e, back_inserter (r));
+    else
+      url::decode (b, e, back_inserter (r),
+                   [] (char& c)
+                   {
+                     if (c == '+')
+                       c = ' ';
+                   });
+    return r;
   }
 }
diff --git a/web/mime-url-encoding.hxx b/web/mime-url-encoding.hxx
index d45f4ca..f748738 100644
--- a/web/mime-url-encoding.hxx
+++ b/web/mime-url-encoding.hxx
@@ -6,25 +6,28 @@
 #define WEB_MIME_URL_ENCODING_HXX
 
 #include <string>
-#include <iosfwd>
 
 namespace web
 {
-  // @@ Add the query flag (true by default). If true, then the encoding is
-  //    applied to the URL query part, and so the plus character is used to
-  //    encode the space character. Audit use cases afterwards.
+  // URL-encode characters other than unreserved (see RFC3986). If the query
+  // flag is true, then the encoding is applied to the URL query part, and so
+  // convert space characters to plus characters rather than percent-encode
+  // them.
   //
-  void
-  mime_url_encode (const char* v, std::ostream& o);
-
   std::string
-  mime_url_encode (const char* v);
+  mime_url_encode (const char*, bool query = true);
 
   std::string
-  mime_url_encode (const std::string& v);
+  mime_url_encode (const std::string&, bool query = true);
 
+  // If the query flag is true, then convert plus characters to space
+  // characters (see above). Throw std::invalid_argument if an invalid encoding
+  // sequence is encountered.
+  //
   std::string
-  mime_url_decode (const char* b, const char* e, bool trim = false);
+  mime_url_decode (const char* b, const char* e,
+                   bool trim = false,
+                   bool query = true);
 }
 
 #endif // WEB_MIME_URL_ENCODING_HXX