Arrange so we can now open the parent of a document (e.g. chm file instead of temp copy of html page inside chm), even when the parent is itself embedded in an archive

src/doc/user/usermanual.sgml

@@ -215,7 +215,7 @@
           <application>Python</application>
           programming interface</link>, a <link linkend="rcl.search.kio">
           <application>KDE</application> KIO slave module</link>, and
-        a <application>Ubuntu Unity Lens</application> module.
+        a <ulink url="http://bitbucket.org/medoc/recoll/wiki/UnityLens">Ubuntu Unity Lens</ulink> module.
         </para>
 
     </sect1>
@@ -2825,8 +2825,29 @@ dir:recoll dir:src -dir:utils -dir:common
       <title>Desktop integration</title>
 
       <para>Being independant of the desktop type has its drawbacks: &RCL;
-      desktop integration is minimal. Here follow a few things that may
+      desktop integration is minimal. However there are a few tools
-      help.</para> 
+      available:
+      <itemizedlist>
+          <listitem>
+            <para>The <application>KDE</application> KIO Slave was
+            described in a <link linkend="rcl.search.kio">previous 
+            section</link>.</para>
+          </listitem>
+        <listitem>
+            <para>If you use a recent version of Ubuntu Linux, you may
+             find the <ulink
+            url="http://bitbucket.org/medoc/recoll/wiki/UnityLens">Ubuntu Unity
+           Lens</ulink> module useful.</para>
+          </listitem>
+          <listitem>
+            <para>There is also an independantly developed
+                 <ulink
+      url="http://kde-apps.org/content/show.php/recollrunner?content=128203">
+            Krunner plugin</ulink>.</para>
+          </listitem>
+        </itemizedlist>
+
+       <para>Here follow a few other things that may help.</para>
 
       <sect2 id="rcl.search.shortcut">
         <title>Hotkeying recoll</title>
@@ -2844,6 +2865,7 @@ dir:recoll dir:src -dir:utils -dir:common
       <sect2 id="rcl.kicker-applet">
         <title>The KDE Kicker Recoll applet</title>
 
+        <para>This is probably obsolete now. Anyway:</para>
         <para>The &RCL; source tree contains the source code to the
         <application>recoll_applet</application>, a small application derived
         from the <application>find_applet</application>. This can be used to

src/filters/rclchm

@@ -190,7 +190,6 @@ class rclCHM:
     def __init__(self, em):
         self.contents = []
         self.chm = chm.CHMFile()
-        self.currentindex = 0
         self.em = em
         if rclchm_catenate:
             self.em.setmimetype("text/plain")
@@ -240,7 +239,7 @@ class rclCHM:
         """Open the chm file and build the contents list by extracting and
         parsing the Topics object"""
 
-        self.currentindex = 0
+        self.currentindex = -1
         self.contents = []
         
         filename = params["filename:"]
@@ -248,8 +247,6 @@ class rclCHM:
             self.em.rclog("LoadCHM failed")
             return False
 
-        self.sfn = os.path.basename(filename)
-
         #self.em.rclog("home [%s] topics [%s] title [%s]" %
         #              (self.chm.home, self.chm.topics, self.chm.title))
 
@@ -293,6 +290,16 @@ class rclCHM:
             else:
                 return (False, "", "", rclexecm.RclExecM.eofnow)
 
+        if self.currentindex == -1:
+            # Return "self" doc
+            self.currentindex = 0
+            self.em.setmimetype('text/plain')
+            if len(self.contents) == 0:
+                eof = rclexecm.RclExecM.eofnext
+            else:
+                eof = rclexecm.RclExecM.noteof
+            return (True, "", "", eof)
+
         if self.currentindex >= len(self.contents):
             return (False, "", "", rclexecm.RclExecM.eofnow)
         else:

src/filters/rclepub

@@ -48,7 +48,7 @@ class rclEPUB:
 
     def openfile(self, params):
         """Open the EPUB file, create a contents array"""
-        self.currentindex = 0
+        self.currentindex = -1
         self.contents = []
         try:
             self.book = epub.open(params["filename:"])
@@ -65,6 +65,17 @@ class rclEPUB:
         return self.extractone(params["ipath:"])
         
     def getnext(self, params):
+
+        if self.currentindex == -1:
+            # Return "self" doc
+            self.currentindex = 0
+            self.em.setmimetype('text/plain')
+            if len(self.contents) == 0:
+                eof = rclexecm.RclExecM.eofnext
+            else:
+                eof = rclexecm.RclExecM.noteof
+            return (True, "", "", eof)
+
         if self.currentindex >= len(self.contents):
             return (False, "", "", rclexecm.RclExecM.eofnow)
         else:

src/filters/rclics

@@ -60,7 +60,7 @@ class IcalExtractor:
             self.em.rclog("Openfile: open: %s" % str(e))
             return False
 
-        self.currentindex = 0
+        self.currentindex = -1
 
         if usemodule == 'internal':
             self.contents = ICalSimpleSplitter().splitcalendar(calstr)
@@ -96,6 +96,17 @@ class IcalExtractor:
         return self.extractone(index)
         
     def getnext(self, params):
+
+        if self.currentindex == -1:
+            # Return "self" doc
+            self.currentindex = 0
+            self.em.setmimetype('text/plain')
+            if len(self.contents) == 0:
+                eof = rclexecm.RclExecM.eofnext
+            else:
+                eof = rclexecm.RclExecM.noteof
+            return (True, "", "", eof)
+
         if self.currentindex >= len(self.contents):
             self.em.rclog("getnext: EOF hit")
             return (False, "", "", rclexecm.RclExecM.eofnow)

src/filters/rclinfo

@@ -73,7 +73,7 @@ class InfoExtractor:
             sys.exit(1);
 
 
-        self.currentindex = 0
+        self.currentindex = -1
 
         self.contents = InfoSimpleSplitter().splitinfo(self.file, infostream)
 
@@ -90,6 +90,17 @@ class InfoExtractor:
 
     # Extract next in list
     def getnext(self, params):
+
+        if self.currentindex == -1:
+            # Return "self" doc
+            self.currentindex = 0
+            self.em.setmimetype('text/plain')
+            if len(self.contents) == 0:
+                eof = rclexecm.RclExecM.eofnext
+            else:
+                eof = rclexecm.RclExecM.noteof
+            return (True, "", "", eof)
+
         if self.currentindex >= len(self.contents):
             self.em.rclog("getnext: EOF hit")
             return (False, "", "", rclexecm.RclExecM.eofnow)

src/filters/rclrar

@@ -73,7 +73,7 @@ class RarExtractor:
 
     ###### File type handler api, used by rclexecm ---------->
     def openfile(self, params):
-        self.currentindex = 0
+        self.currentindex = -1
         try:
             self.rar = RarFile(params["filename:"])
             return True
@@ -93,6 +93,17 @@ class RarExtractor:
             return (ok, data, ipath, eof)
         
     def getnext(self, params):
+
+        if self.currentindex == -1:
+            # Return "self" doc
+            self.currentindex = 0
+            self.em.setmimetype('text/plain')
+            if len(self.rar.namelist()) == 0:
+                eof = rclexecm.RclExecM.eofnext
+            else:
+                eof = rclexecm.RclExecM.noteof
+            return (True, "", "", eof)
+
         if self.currentindex >= len(self.rar.namelist()):
             #self.em.rclog("getnext: EOF hit")
             return (False, "", "", rclexecm.RclExecM.eofnow)

src/filters/rcltar

@@ -43,7 +43,7 @@ class TarExtractor:
         return (ok, docdata, ipath, iseof)
 
     def openfile(self, params):
-        self.currentindex = 0
+        self.currentindex = -1
         try:
             self.tar = open(name=params["filename:"],mode='r')
             self.namen = [ y.name for y in filter(lambda z:z.isfile(),self.tar.getmembers())]
@@ -63,6 +63,17 @@ class TarExtractor:
             return (ok, data, ipath, eof)
 
     def getnext(self, params):
+
+        if self.currentindex == -1:
+            # Return "self" doc
+            self.currentindex = 0
+            self.em.setmimetype('text/plain')
+            if len(self.namen) == 0:
+                eof = rclexecm.RclExecM.eofnext
+            else:
+                eof = rclexecm.RclExecM.noteof
+            return (True, "", "", eof)
+
         if self.currentindex >= len(self.namen):
             self.namen=[]
             return (False, "", "", rclexecm.RclExecM.eofnow)

src/filters/rclwar

@@ -22,7 +22,7 @@ class WarExtractor:
 
     ###### File type handler api, used by rclexecm ---------->
     def openfile(self, params):
-        self.currentindex = 0
+        self.currentindex = -1
         try:
             self.tar = tarfile.open(params["filename:"])
             return True
@@ -40,6 +40,11 @@ class WarExtractor:
         return self.extractone(tarinfo)
 
     def getnext(self, params):
+        if self.currentindex == -1:
+            # Return "self" doc
+            self.currentindex = 0
+            return (True, "", "", rclexecm.RclExecM.noteof)
+
         tarinfo = self.tar.next()
         if tarinfo is None:
             #self.em.rclog("getnext: EOF hit")

src/filters/rclzip

@@ -67,7 +67,7 @@ class ZipExtractor:
 
     ###### File type handler api, used by rclexecm ---------->
     def openfile(self, params):
-        self.currentindex = 0
+        self.currentindex = -1
         try:
             self.zip = ZipFile(params["filename:"])
             return True
@@ -87,6 +87,16 @@ class ZipExtractor:
             return (ok, data, ipath, eof)
         
     def getnext(self, params):
+        if self.currentindex == -1:
+            # Return "self" doc
+            self.currentindex = 0
+            self.em.setmimetype('text/plain')
+            if len(self.zip.namelist()) == 0:
+                eof = rclexecm.RclExecM.eofnext
+            else:
+                eof = rclexecm.RclExecM.noteof
+            return (True, "", "", eof)
+
         if self.currentindex >= len(self.zip.namelist()):
             #self.em.rclog("getnext: EOF hit")
             return (False, "", "", rclexecm.RclExecM.eofnow)

src/internfile/internfile.cpp

@@ -126,6 +126,16 @@ bool FileInterner::getEnclosing(const string &url, const string &ipath,
     return true;
 }
 
+string FileInterner::getLastIpathElt(const string& ipath)
+{
+    string::size_type sep;
+    if ((sep =  ipath.find_last_of(cstr_isep)) != string::npos) {
+	return ipath.substr(sep + 1);
+    } else {
+	return ipath;
+    }
+}
+
 // Uncompress input file into a temporary one, by executing the appropriate
 // script.
 static bool uncompressfile(RclConfig *conf, const string& ifn, 

src/internfile/internfile.h

@@ -83,6 +83,9 @@ class FileInterner {
     static bool getEnclosing(const string &url, const string &ipath,
 			     string &eurl, string &eipath, string& udi);
 
+    /** Return last element in ipath, like basename */
+    static std::string getLastIpathElt(const std::string& ipath);
+
     /** Constructors take the initial step to preprocess the data object and
      *  create the top filter */
 

src/qtgui/rclmain_w.cpp

@@ -1501,13 +1501,17 @@ static bool lookForHtmlBrowser(string &exefile)
 
 void RclMain::startNativeViewer(Rcl::Doc doc, int pagenum, QString term)
 {
-    LOGDEB(("RclMain::startNativeViewer: page %d\n", pagenum));
-    // Look for appropriate viewer
     string apptag;
     doc.getmeta(Rcl::Doc::keyapptg, &apptag);
+    LOGDEB(("RclMain::startNativeViewer: mtype [%s] apptag [%s] page %d "
+	    "term [%s] url [%s] ipath [%s]\n", 
+	    doc.mimetype.c_str(), apptag.c_str(), pagenum, 
+	    (const char *)(term.toUtf8()), doc.url.c_str(), doc.ipath.c_str()
+	       ));
+
+    // Look for appropriate viewer
     string cmdplusattr = theconfig->getMimeViewerDef(doc.mimetype, apptag, 
 						     prefs.useDesktopOpen);
-
     if (cmdplusattr.empty()) {
 	QMessageBox::warning(0, "Recoll", 
 			     tr("No external viewer configured for mime type [")
@@ -1515,25 +1519,12 @@ void RclMain::startNativeViewer(Rcl::Doc doc, int pagenum, QString term)
 	return;
     }
 
-    if (pagenum == -1) {
+    // Separate command string and viewer attributes (if any)
-	pagenum = 1;
+    ConfSimple viewerattrs;
-	string lterm;
-	if (m_source.isNotNull())
-	    pagenum = m_source->getFirstMatchPage(doc, lterm);
-	if (pagenum == -1)
-	    pagenum = 1;
-	else
-	    term = QString::fromUtf8(lterm.c_str());
-    }
-    char cpagenum[20];
-    sprintf(cpagenum, "%d", pagenum);
-
-    // Extract possible viewer attributes
-    ConfSimple attrs;
     string cmd;
-    theconfig->valueSplitAttributes(cmdplusattr, cmd, attrs);
+    theconfig->valueSplitAttributes(cmdplusattr, cmd, viewerattrs);
     bool ignoreipath = false;
-    if (attrs.get("ignoreipath", cmdplusattr))
+    if (viewerattrs.get("ignoreipath", cmdplusattr))
         ignoreipath = stringToBool(cmdplusattr);
 
     // Split the command line
@@ -1541,7 +1532,7 @@ void RclMain::startNativeViewer(Rcl::Doc doc, int pagenum, QString term)
     if (!stringToStrings(cmd, lcmd)) {
 	QMessageBox::warning(0, "Recoll", 
 			     tr("Bad viewer command line for %1: [%2]\n"
-				"Please check the mimeconf file")
+				"Please check the mimeview file")
 			     .arg(QString::fromAscii(doc.mimetype.c_str()))
 			     .arg(QString::fromLocal8Bit(cmd.c_str())));
 	return;
@@ -1549,23 +1540,23 @@ void RclMain::startNativeViewer(Rcl::Doc doc, int pagenum, QString term)
 
     // Look for the command to execute in the exec path and the filters 
     // directory
-    string cmdpath;
+    string execpath;
-    if (!ExecCmd::which(lcmd.front(), cmdpath)) {
+    if (!ExecCmd::which(lcmd.front(), execpath)) {
-	cmdpath = theconfig->findFilter(lcmd.front());
+	execpath = theconfig->findFilter(lcmd.front());
 	// findFilter returns its input param if the filter is not in
 	// the normal places. As we already looked in the path, we
 	// have no use for a simple command name here (as opposed to
 	// mimehandler which will just let execvp do its thing). Erase
-	// cmdpath so that the user dialog will be started further
+	// execpath so that the user dialog will be started further
 	// down.
-	if (!cmdpath.compare(lcmd.front())) 
+	if (!execpath.compare(lcmd.front())) 
-	    cmdpath.erase();
+	    execpath.erase();
 
 	// Specialcase text/html because of the help browser need
-	if (cmdpath.empty() && !doc.mimetype.compare("text/html")) {
+	if (execpath.empty() && !doc.mimetype.compare("text/html")) {
-	    if (lookForHtmlBrowser(cmdpath)) {
+	    if (lookForHtmlBrowser(execpath)) {
 		lcmd.clear();
-		lcmd.push_back(cmdpath);
+		lcmd.push_back(execpath);
 		lcmd.push_back("%u");
 	    }
 	}
@@ -1573,7 +1564,7 @@ void RclMain::startNativeViewer(Rcl::Doc doc, int pagenum, QString term)
 
 
     // Command not found: start the user dialog to help find another one:
-    if (cmdpath.empty()) {
+    if (execpath.empty()) {
 	QString mt = QString::fromAscii(doc.mimetype.c_str());
 	QString message = tr("The viewer specified in mimeview for %1: %2"
 			     " is not found.\nDo you want to start the "
@@ -1594,22 +1585,68 @@ void RclMain::startNativeViewer(Rcl::Doc doc, int pagenum, QString term)
         // new command.
 	return;
     }
+    // Get rid of the command name. lcmd is now argv[1...n]
+    lcmd.erase(lcmd.begin());
 
-    // We may need a temp file, or not, depending on the command
+
-    // arguments and the fact that this is a subdoc or not.
+    // Process the command arguments to determine if we need to create
-    bool wantsipath = (cmd.find("%i") != string::npos) || ignoreipath;
+    // a temporary file.
+
+    // If the command has a %i parameter it will manage the
+    // un-embedding. Else if ipath is not empty, we need a temp file.
+    // This can be overridden with the "ignoreipath" attribute
+    bool groksipath = (cmd.find("%i") != string::npos) || ignoreipath;
+
+    // wantsfile: do we actually need a local file ? The only other
+    // case here is an url %u (ie: for web history).
     bool wantsfile = cmd.find("%f") != string::npos;
-    bool istempfile = false;
+    bool wantsparentfile = cmd.find("%F") != string::npos;
-    string fn = fileurltolocalpath(doc.url);
+
-    string orgfn = fn;
+    if (wantsfile && wantsparentfile) {
+	QMessageBox::warning(0, "Recoll", 
+			     tr("Viewer command line for %1 specifies both "
+				"file and parent file value: unsupported")
+			     .arg(QString::fromAscii(doc.mimetype.c_str())));
+	return;
+    }
+	
     string url = doc.url;
+    string fn = fileurltolocalpath(doc.url);
+    Rcl::Doc pdoc;
+    if (wantsparentfile) {
+	// We want the path for the parent document. For example to
+	// open the chm file, not the internal page. Note that we just
+	// override the other file name in this case.
+	if (m_source.isNull() || !m_source->getEnclosing(doc, pdoc)) {
+	    QMessageBox::warning(0, "Recoll",
+				 tr("Cannot find parent document"));
+	    return;
+	}
+	// Override fn with the parent's : 
+	fn = fileurltolocalpath(pdoc.url);
+
+	// If the parent document has an ipath too, we need to create
+	// a temp file even if the command takes an ipath
+	// parameter. We have no viewer which could handle a double
+	// embedding. Will have to change if such a one appears.
+	if (!pdoc.ipath.empty()) {
+	    groksipath = false;
+	}
+    }
+
+    bool istempfile = false;
+
+    LOGDEB(("RclMain::startNV: groksipath %d wantsf %d wantsparentf %d\n", 
+	    groksipath, wantsfile, wantsparentfile));
 
     // If the command wants a file but this is not a file url, or
     // there is an ipath that it won't understand, we need a temp file:
     theconfig->setKeyDir(path_getfather(fn));
-    if ((wantsfile && fn.empty()) || (!wantsipath && !doc.ipath.empty())) {
+    if (((wantsfile || wantsparentfile) && fn.empty()) ||
+	(!groksipath && !doc.ipath.empty())) {
 	TempFile temp;
-	if (!FileInterner::idocToFile(temp, string(), theconfig, doc)) {
+	Rcl::Doc& thedoc = wantsparentfile ? pdoc : doc;
+	if (!FileInterner::idocToFile(temp, string(), theconfig, thedoc)) {
 	    QMessageBox::warning(0, "Recoll",
 				 tr("Cannot extract document or create "
 				    "temporary file"));
@@ -1647,8 +1684,21 @@ void RclMain::startNativeViewer(Rcl::Doc doc, int pagenum, QString term)
         }
     }
 
-    // Get rid of the command name. lcmd is now argv[1...n]
+    // If we are not called with a page number (which would happen for a call
-    lcmd.erase(lcmd.begin());
+    // from the snippets window), see if we can compute a page number anyway.
+    if (pagenum == -1) {
+	pagenum = 1;
+	string lterm;
+	if (m_source.isNotNull())
+	    pagenum = m_source->getFirstMatchPage(doc, lterm);
+	if (pagenum == -1)
+	    pagenum = 1;
+	else // We get the match term used to compute the page
+	    term = QString::fromUtf8(lterm.c_str());
+    }
+    char cpagenum[20];
+    sprintf(cpagenum, "%d", pagenum);
+
 
     // Substitute %xx inside arguments
     string efftime;
@@ -1662,8 +1712,8 @@ void RclMain::startNativeViewer(Rcl::Doc doc, int pagenum, QString term)
     map<string, string> subs;
     subs["D"] = efftime;
     subs["f"] = fn;
-    subs["F"] = orgfn;
+    subs["F"] = fn;
-    subs["i"] = doc.ipath;
+    subs["i"] = FileInterner::getLastIpathElt(doc.ipath);
     subs["M"] = doc.mimetype;
     subs["p"] = cpagenum;
     subs["s"] = (const char*)term.toLocal8Bit();
@@ -1705,7 +1755,7 @@ void RclMain::startNativeViewer(Rcl::Doc doc, int pagenum, QString term)
     // We keep pushing back and never deleting. This can't be good...
     ExecCmd *ecmd = new ExecCmd;
     m_viewers.push_back(ecmd);
-    ecmd->startExec(cmdpath, lcmd, false, false);
+    ecmd->startExec(execpath, lcmd, false, false);
 }
 
 void RclMain::startManual()

tests/chm/chm.txt

@@ -1,7 +1,8 @@
 1 results
 text/html	[file:///home/dockes/projets/fulltext/testrecoll/chm/Nokia_Nseries_Help_jpn.chm]	[デジタル著作権管理(DRM)]	2948	bytes	
-1 results
+2 results
 text/html	[file:///home/dockes/projets/fulltext/testrecoll/chm/soundrec.chm]	[Superposer (mixer) des fichiers son]	2279	bytes	
+text/html	[file:///home/dockes/projets/fulltext/testrecoll/chm/zippedsoundrec.zip]	[Superposer (mixer) des fichiers son]	2279	bytes	
 1 results
 text/html	[file:///home/dockes/projets/fulltext/testrecoll/chm/Django-1.1a1-r9905.chm]	[User authentication in Django]	115179	bytes	
 1 results