Webmaster Forum / HTML, CSS, Scripts / JavaScript / April 2006

As it happens viewing the page in Firefox did exhibit the problem, and
so allow its cause to be identified.

The newsgroup didn't add any wrapping, that was done by your positing
software and so is your personal responsibility. The base way of
avoiding it is as described in the FAQ (or the liked material).

It is a good idea, and if you use the TARGET attribute in the A element
you can also pass that onto pop-up window scripts as - this.target -,
and it will turn out that adding - "tartget="_blank" - to your like may
solve your problem in the simplest way.

Now to the issue itself. My first problem with your code was that it
should 'work' properly. That is, if it encounters a pop-up blocker or a
browser that does not implement - window.open - it won't actually work
at all (or will work unexpectedly) but on a common desktop browser with
pop-ups enabled and not blocked it should do what it was written to do.
As a result trying it in a couple of browsers and finding that it did
indeed work made this look like a non-issue.

In case you haven't guessed by now, VK's explanation is another of his
fantasies; a fiction made up off the top of his head and without any
meaning in relation to javascript and browser scripting. He seems to
believe this nonsense himself, no matter how often it is demonstrated
wrong, leaving everyone else with little choice but attribute his
postings to mental illness. Unfortunately people who don't know
javascript can be taken in by his ravings, and in fact he has stated
that impressing those who don't know any better is one of his aims.

Any analyses of an issue must account for the symptoms. Here the
symptoms are that your onclick handler successfully opens the pop-up
window and the page itself is navigated to the URL in its HREF
attribute. When a page navigates to the URL specified in its HREF, where
the link has an onclick attribute that returns false, the usual
explanation is that the onclick handler actually errored and so did not
return at all. This is quite common when the onclick code is attempting
to open pop-up windows because attempts to open, or interact with,
windows in the presence of pop-up blockers (of various sorts) can cause
exceptions to be thrown.

However, the action of pop-up blockers can be dismissed entirely when
one of the symptoms is that the window actually opened. If an error is
generated then it can only happen after the call to - window.open -, as
that call must have executed successfully. And there are only two
statements executed following the call to - window.open -, two return
statements. The return statement cannot throw an exception unless an
error is generated when evaluating the expression that is returned,
and - false - cannot fail to be successfully resolved in javascript as
it is a direct reference to one of the Boolean values. The fact that no
error is reported (in, for example, firefox's javascript console,
confirms this impression.

Thus, the code used, and posted here, cannot account for the symptoms
described; it should (broadly) work as expected. So it is necessary to
look of external explanations for the phenomenon. Viewing the source of
the page in a browser that exhibits the symptoms reveals that the
geocities server is inserting many elements into the page, including a
number of SCRIPT elements. Examining the scripts inserted with these
elements identified the cause of the unexpected page navigation. This
file:-

<quote cite="http://us.geocities.com/js_source/geovck08.js">
var d=document,w=window,
   h=location.hostname,fC=false,fO=false,day=new Date();
var id=''+day.getFullYear()+day.getDate()+'';
var tF='http://' + h + '/js_source/tab04.html';
var aF='http://' + h + '/js_source/adframe07.html';
var mP=location.href,cP=mP;
var tW="16",oW="184",cW="0",aW=oW;
var ts=(new Date()).getTime();
var mFr="m"+ts,tFr="t"+ts,sFr="s"+ts;
var oFR = '<frameset id="FR" cols="*,'+tW+','+aW+
   '" frameborder="NO" border="0" framespacing="0">';
var oMF = '<frame id="'+mFr+'" name="'+mFr+
   '" noresize="noresize" frameborder="0" src="';
var cMF = '">';
var oTF = '<frame id="'+tFr+'" name="'+tFr+
   '" noresize="noresize" frameborder="0" src="' + tF + ' ">';
var oAF = '<frame id="'+sFr+'" name="'+sFr+
   '" noresize="noresize" frameborder="0" src="' + aF +'">';
var cFR = '</frameset>';

function wt(){
 top.d.title = eval('top.'+mFr+'.document.title');
}

function ssF(){
 if(d.getElementById==null){
   d.getElementById=d.all
 };
 if(fO==false){
   eval(
     'top.d.getElementById("FR").setAttribute("cols","*,'+
      tW+','+aW+'");'
   );
   fO=true;
 }else{
   aW=cW;
   eval(
     'top.d.getElementById("FR").setAttribute("cols","*,'+
     tW+','+aW+'");'
   );
   fO = false; aW=oW;
 }
 wt();
}

if ((self==top)){
 if(cP.indexOf(id)<0){
   w.onload=function(){
     ssF();
   }
   if (cP.indexOf("#")>-1){
     sg=cP.split("#");
     mP=sg[0]+'?'+id+'#'+sg[1];
   }else{
     mP+='?'+id;
   }
   d.write(oFR+oMF+mP+cMF+oTF+oAF+cFR);
 }else{
   fC=true;
 }
}

function isBase(){
 var numB=0;
 numB=document.getElementsByTagName("BASE").length;
 isB=(numB>0) ? true : false; return isB;
}

function cH(e){
 e=e||window.event;
 var tgt=e.target||e.srcElement;
 if((e.shiftKey)||(e.ctrlKey)||(e.button>=2)){
   return e;
 }
 if ((tgt)&&(tgt.nodeName=="A")){
   var uPath=tgt.pathname;
   var uHash=tgt.hash;
   var uHref=tgt.href;
   var uHostname=tgt.hostname;
   var uTarget=tgt.target;
   var uNodename=tgt.nodeName;
   var uProt=tgt.protocol;
 }else if (
   (tgt.parentNode)&&(tgt.parentNode.nodeName=="A")
 ) {
   var uPath=tgt.parentNode.pathname;
   var uHash=tgt.parentNode.hash;
   var uHref=tgt.parentNode.href;
   var uHostname=tgt.parentNode.hostname;
   var uTarget=tgt.parentNode.target;
   var uNodename=tgt.parentNode.nodeName;
   var uProt=tgt.parentNode.protocol;
 }else if (
   (tgt.parentNode.parentNode)&&
   (tgt.parentNode.parentNode.nodeName=="A")
 ) {
   var uPath=tgt.parentNode.parentNode.pathname;
   var uHash=tgt.parentNode.parentNode.hash;
   var uHref=tgt.parentNode.parentNode.href;
   var uHostname=tgt.parentNode.parentNode.hostname;
   var uTarget=tgt.parentNode.parentNode.target;
   var uNodename=tgt.parentNode.parentNode.nodeName;
   var uProt=tgt.parentNode.parentNode.protocol;
 }else if (
   (tgt.parentNode.parentNode.parentNode)&&
   (tgt.parentNode.parentNode.parentNode.nodeName=="A")
 ) {
   var uPath=tgt.parentNode.parentNode.parentNode.pathname;
   var uHash=tgt.parentNode.parentNode.parentNode.hash;
   var uHref=tgt.parentNode.parentNode.parentNode.href;
   var uHostname=tgt.parentNode.parentNode.parentNode.hostname;
   var uTarget=tgt.parentNode.parentNode.parentNode.target;
   var uNodename=tgt.parentNode.parentNode.parentNode.nodeName;
   var uProt=tgt.parentNode.parentNode.parentNode.protocol;
 }else{
   return e;
 }
 if (uProt!="http:") return;
 if (d.getElementsByTagName("IFRAME").length>0){
   var foundOne=false;
   var nI=d.getElementsByTagName("IFRAME").length;
   for (i=0; i<nI; i++){
     if(
       (d.getElementsByTagName("IFRAME")[i].name!="")&&
       ((d.getElementsByTagName("IFRAME")[i].name)==(uTarget))
     ){
       foundOne=true;
       break;
     }
   }
   if(foundOne==true){
     return;
   }
   foundOne=false;
 }
 if((top.w.frames[0].length<=0)){
   if (uHref.indexOf(id)>-1){
     tHref = uHref.split("?");
     uHref=tHref[0]+uHash;
     eval("location.href=uHref;");
   }else if(uHref.indexOf('#')>-1){
     eval("location.href=uHref;");
   }
 }
 var ytg = uTarget;
 ytg=ytg.toLowerCase();
 if ((ytg.indexOf("new")>-1)||(ytg.indexOf("blank")>-1)){
   return;
 }
 if ((ytg.indexOf("self")>-1)&&(top.w.frames[0].length>0)){
   return;
 }
 if((uHref)&&(top.w.frames[0].length!=0)&&(uHref.indexOf("#")>-1)){
   return;
 }
 if ((tgt)&&(!fC)) {
   if (uNodename=="A") {
     if ((uHostname.indexOf("geocities.com"))>-1){
       if((uHref.indexOf("#")<=-1)&&(frames.name==mFr)){
         fO=false;ssF();
         top.location.href=uHref;
       }else if(
         (
           (top.w.frames[0].length)-
           (d.getElementsByTagName("IFRAME").length)!=0)&&
         (uHref.indexOf("#")<=-1)
       ){
         if ((uTarget)&&(uTarget!=sFr)&&(uTarget!=tFr)) {

         }else{
           if(isBase()) {
             uTarget=document.getElementsByTagName("BASE")[0].target;
             return;
           }
           location.href=uHref;
         }
       }else{
         top.location.href=uHref;
         return false;
       }
     }else{
       top.d.location.href=uHref;
     }
   }else{
     return e;
   }
 }
}

d.onclick=cH;
</quote>

- included in the page (presented here re-formatted for clarity).

This file adds an - onclick handler to the document. This handler
attempts to compensate for geocities efforts to add advertising. That
advertising would be presented in a frameset, and so when navigation
within a page is expected geocities's addition of a frameset would make
it navigation inside the frameset, and then geocities would (attempt to)
make that new page into another frameset. To avoid this they use the
document's onclick handler to intercept clicks on A elements and try to
do what the link would have done otherwise. Unfortunately geocities
employ very unskilled script authors and their efforts are misguided and
error prone. Your issue is a direct result, as they cannot tell that the
click event was cancelled by your onclick handler so they try to
navigate the page to the URL in the HREF by assigning the URL to the -
location.href - property of the page.

You can easily see that the geocities script authors are inept because
the code includes:-

eval("location.href=uHref;");

Which means that the script's author doesn't even know enough to see why
the above can be substituted with the simpler and more direct:-

location.href=uHref;

- to precisely the same effect.

It looks like the author of the script did actually realise that
attempts to open links in new windows did not require then action of
this script, so if you add the attribute - target="_blank" - the script
will assume that it does not need to act upon the clicking on the link.

An alternative solution would be to find a hosting company that does not
insert scripts into your pages, that will eliminate the possibility of
future problems resulting form the less than intelligent scripts
inserted by geocities.

Richard.

I am wondering if there are any other ideas to address my original
query:

-------ORIGINAL QUERY ------------------------------------------
I am trying to pop-up a window when the user clicks on an image.  The
problem is that when the user clicks on the image and the window pops
up OK, but the window underneath also proceeds to the image.  The
desired behavior is that when the pop-up is invoked, I want the
underlying window to stay put.

I don't have this problem when I run the code on my local computer but
I do have it when I run the code on geocities.

The webpage code is simple:

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0 Transitional//EN">
<html>
<head>
<title>Page title</title>
<script language="JavaScript" type="text/javascript">
var tallParm =
'scrollbars=1,toolbar=0,location=0,directories=0,status=0,menubar=0,resizable=1,height='+(screen.height
- 200)+',width=' + (screen.width - 80);

function pop_tall_image_win (ref) {
       window.open(ref,"x",tallParm);
       return false;
}

</script>
</head>
<body>
<a href="large.gif" onclick="return
pop_tall_image_win('large.gif')"><img src="small.gif" border="0"></a>
</body>
</html>

The place where I have the page on geocities is:
http://www.geocities.com/sample_issue/index.htm

As a side-note, I still want the link on the image to go to the target,
in case the user decides to use right-mouse open-new-window.  That is
why I have the HREF reference.

END OF ORIGINAL QUERY ------------------------------------------

1. This was suggested and made no difference:
<a href="large.gif" onclick="pop_tall_image_win('large.gif'); return
false;"><img src="small.gif" border="0"></a>

2. Also, someone queried the line-wrapping and that is just the
newgroup posting adding carriage returns.  The full code at:
http://www.geocities.com/sample_issue/index.htm has no extra carriage
returns.

3. I am seeing this in Firefox 1.0, IE 6.0-SP2 and Netscape 8.1 when
accessing the link above (and others have reported it too).  Someone
suggested that it might not be happening in Opera but I don't have
Opera so I cannot comment on that.