allow opening parent/enclosing doc with native editor in reslist

src/utils/fileudi.h

@@ -24,9 +24,8 @@ using std::string;
 // indexer).  Document Ids are built from a concatenation of the file
 // path and the internal path (ie: email number inside
 // folder/attachment number/etc.)  As the size of Xapian terms is
-// limited, the path is truncated to a maximum length, and completed
-// by a hash of the remainder. So the unique id looks like:
-// /some/truncated/paHASHVALUE|ipath
+// limited, the Id path is truncated to a maximum length, and completed
+// by a hash of the remainder (including the ipath)
 
 extern void make_udi(const string& fn, const string& ipath, string &udi);
 

src/utils/pathut.cpp

@@ -176,13 +176,13 @@ TempFileInternal::~TempFileInternal()
 }
 
 
-void path_catslash(std::string &s) {
+void path_catslash(string &s) {
     if (s.empty() || s[s.length() - 1] != '/')
 	s += '/';
 }
 
-std::string path_cat(const std::string &s1, const std::string &s2) {
-    std::string res = s1;
+string path_cat(const string &s1, const string &s2) {
+    string res = s1;
     path_catslash(res);
     res +=  s2;
     return res;
@@ -274,7 +274,7 @@ extern string path_tildexpand(const string &s)
     return o;
 }
 
-extern std::string path_absolute(const std::string &is)
+extern string path_absolute(const string &is)
 {
     if (is.length() == 0)
 	return is;
@@ -290,7 +290,7 @@ extern std::string path_absolute(const std::string &is)
 }
 
 #include <smallut.h>
-extern std::string path_canon(const std::string &is)
+extern string path_canon(const string &is)
 {
     if (is.length() == 0)
 	return is;
@@ -330,8 +330,8 @@ extern std::string path_canon(const std::string &is)
 
 #include <glob.h>
 #include <sys/stat.h>
-list<std::string> path_dirglob(const std::string &dir, 
-				    const std::string pattern)
+list<string> path_dirglob(const string &dir, 
+				    const string pattern)
 {
     list<string> res;
     glob_t mglob;
@@ -356,7 +356,7 @@ bool path_isdir(const string& path)
     return false;
 }
 
-std::string url_encode(const std::string url, string::size_type offs)
+string url_encode(const string& url, string::size_type offs)
 {
     string out = url.substr(0, offs);
     const char *cp = url.c_str();
@@ -392,6 +392,26 @@ std::string url_encode(const std::string url, string::size_type offs)
     return out;
 }
 
+string url_gpath(const string& url)
+{
+    // Remove the access schema part (or whatever it's called)
+    string::size_type colon = url.find_first_of(":");
+    if (colon == string::npos || colon == url.size() - 1)
+        return url;
+    // If there are non-alphanum chars before the ':', then there
+    // probably is no scheme. Whatever...
+    for (string::size_type i = 0; i < colon; i++) {
+        if (!isalnum(url.at(i)))
+            return url;
+    }
+
+    // In addition we canonize the path to remove empty host parts
+    // (for compatibility with older versions of recoll where file://
+    // was hardcoded, but the local path was used for doc
+    // identification.
+    return path_canon(url.substr(colon+1));
+}
+
 // Printable url: this is used to transcode from the system charset
 // into either utf-8 if transcoding succeeds, or url-encoded
 bool printableUrl(const string &fcharset, const string &in, string &out)

src/utils/pathut.h

@@ -50,11 +50,16 @@ extern string path_canon(const string &s);
 extern list<string> path_dirglob(const string &dir, 
 					   const string pattern);
 /// Encode according to rfc 1738
-extern string url_encode(const string url, 
+extern string url_encode(const string& url, 
 			      string::size_type offs = 0);
+/// Transcode to utf-8 if possible or url encoding, for display.
 extern bool printableUrl(const string &fcharset, 
 			 const string &in, string &out);
 
+/// Return the host+path part of an url. This is not a general
+/// routine, it does the right thing only in the recoll context
+extern string url_gpath(const string& url);
+
 /// Stat parameter and check if it's a directory
 extern bool path_isdir(const string& path);