Disabling the PHPSESSID in the Query String

15 08 2006

If you’ve worked with PHP sessions, probably your URL looked something like this:

http://www.yourdomain.com/page.php?PHPSESSID=59aa95ad46cd67d82ba0f812407326dd

If you don’t like the PHPSSID being displayed in the query string, you might have wondered how to disable it.

Well, a way to disable it is by setting session.use_trans_sid to off in your php.ini. However, if you don’t have access to the php.ini file because you’re hosting your site on a shared server, then, you can accomplish this by using a .htaccess file*.

The following example worked for me:

<IfModule mod_php4.c>
     php_flag session.use_trans_sid off
</IfModule>

That’s it! I hope this helps.
––––

*Make sure your hosting provider allows .htaccess files.



/usr/bin/wget: Permission denied

10 08 2006

If you’ve ever got the error /bin/sh: /usr/bin/wget: Permission denied when trying to execute a .php script using a cron job, spent minutes/hours trying to find a solution to this, well, I hope this post will help you to save some time.

It’s possible that your hosting provider, for security reasons, is not allowing you to execute the wget command. If you want to use the wget (or GET) command, you can do two things:

a) send an email to your hosting provider’s support department. Or

b) you can try using the following command:

php -q /path/to/script.php

The above example worked for me nicely. However, if it still doesn’t work for you, then contact your hosting provider; maybe there are other things that they need to take care of (or fix) first before you can work with cron jobs.

Cheers!



The textContent and innerText Properties

18 04 2006

In my original post, Using the innerText Property in Firefox, I wrote my example using document.all to determine when to use innerText or textContent.

Because the main purpose of that post was to explain to the reader that Firefox does not support the innerText property but the textContent property, I failed to consider other browsers… yes, shame on me! (thanks Paul for bringing this to my attention).

You see, there are browsers that although don’t support document.all, they DO support the innerText property: Safari and Konqueror.

So, it is more efficient to check if the innerText property is supported by the user agent (thanks, Matthias). The way I do it is like this:

var hasInnerText =
(document.getElementsByTagName("body")[0].innerText != undefined) ? true : false;
var elem = document.getElementById('id');
var elem2 = document.getElementById ('id2');

if(!hasInnerText){
    elem.textContent = value;
    elem2.textContent = value;
} else{
    elem.innerText = value;
    elem2.innerText = value;
}

Why am I using document.getElementsByTagName("body")[0]? Well, because, as you know, document.getElementsByTagName returns an array of elements with that tag name, and since there’s only ONE body tag, I’m calling the first and only index of that array at once. This is so, as you can see in the example above, if you have more than one element that you need to assign a value to.

I have an example. If you wish, check the source out to see how it works.

Compatibility:

  • Linux:
    • Konqueror 3.5.2
    • Firefox 1.5.0.1
  • Mac X:
    • Safari 2.0.3
    • Camino 1.0
  • Windows:
    • Firefox 1.5.0.2
    • Opera 8.54
    • Internet Explorer 6
    • Internet Explorer 7.0.5346.5 Beta 2

Remember, this is one way to do it. Cheers! ;)



Displaying PNG 24-bit Alpha Transparency on IE with CSS

10 01 2006

After reading Justin Koivisto’s great article: Normal Display of PNG Alpha Transparency with MSIE and reading his PHP code, I said to myself: “myself, there should be a much simpler (lazy, if you will :D ) way to accomplish this”. True, I could just copy, paste and use his code, modify the .htaccess file (or add the MIME type to the Apache server) but still… :P

Another approach is Bob Osola’s JavaScript script (this is pretty cool too) found in his article: The PNG problem in Windows Internet Explorer.

Personally, I wouldn’t use more than 2 or 3 images on a page, so I believe the use of CSS will do.

I have an example where you can see the results (please see the page source to see the code). The drawback of this approach (as it is written right now – using the background-image property) is that no image(s) will be displayed if the user has CSS disabled.

Again, if you will have lots of .png images and don’t want to have a bunch of CSS classes, you’re better off using Koivisto’s or Osola’s approach.

CSS Compatibility

  • Firefox 1.5 (Win XP and Linux)
  • Internet Explorer 6 (duh!)
  • Opera 8.51
  • Konqueror 3.5.0

I don’t have Safari but my guess is that it will work :D

Enjoy!



Simple CSS Layout

3 11 2005

Note: By mistake I deleted the original post (dated 08.08.05), so I’m posting it again :P

While I was working on a big project at work, I happened to use CSS a lot and gained a bit of experience and a better understanding on how to manipulate elements and all that stuff.

Well, by doing lots of ‘experiments’ I came up with a technique (which, by the way, I haven’t seen anywhere else) that would make a page have its header and footer to the top and bottom of the browser window as if it was made using tables. I have an example where you can see what I mean.

The layout, consists only of a header, contents and footer. This is just the ‘begining’ because I am still experimenting to expand this technique to use it with columns too.

Now going back to the topic, in my research to find a way to accomplish what I wanted, I always found examples where the container would hold the header, contents and footer… but this wouldn’t produce the results I wanted.

By the way, although I’m not a writer, I will do my best to explain this technique; please bear with me.

This technique consists of placing the header and the contents inside the container only. The footer will be placed outside the container; this will allow us to ‘push’ the footer all the way down and manipulate its position.

In the definition of the container id, we define a height of 100% for Internet Explorer and create a filter for Firefox and other CSS2 compliant browsers like so:

html>body #container {
   height: auto;
   min-height: 100%;
}

Internet Explorer will automatically expand if the contents are larger than the 100% of the screen. In the filter, we define a height with its value set to auto so if the contents are larger than the 100% the screen they won’t overflow outside the container; if we don’t set the min-height to 100%, the contents div will not expand all the way to the bottom of the screen if the contents div is either empty or with few text.

Then, we define the contents div like so:

#contents{
   padding: 10px 10px 100px 10px;
}

As you can see, we set the bottom padding to 100px. The reason for this is so the footer won’t overlap over the header if there is no text in the contents div and to leave some space in between both of them if the browser window is resized to a small size.

Finally, we define the footer div as follows:

#footer{
   position: relative;
   bottom: 0;
   background-color: #f90;
   margin: -100px auto 0 auto;
   height: 100px;
   width: 760px;
}

Since the position of the footer is set to relative, it will be pushed down all the way to the bottom, even beyond the 100% of the user’s screen. For this reason, as you can see in the definition, we give it a top margin of -100px (which is the same value of the footer’s height) and then we give it the same positioning and width values as the container so that they align as they should.

That’s it! You can take it from here and tweak it as needed. See the source code for more details.I hope this technique helps and saves you some time ;) .

Buggy in Opera 7.5. It won’t automatically resize until you reload the page.

Works fine in:

  • Windows XP
    • Firefox 1.4
    • Internet Explorer 6.0
  • Macintosh
    • Firefox 1.0.1
    • Safari 2.0
  • Linux (Fedora Core 3)
    • Firefox 1.4
    • Konqueror 3.4

Doesn’t work:

  • Internet Explorer 5.2 (Macintosh)


Multiple Submit Buttons on a Single Form

4 10 2005

At some point, people ask themselves “How can I have two submit buttons with different actions in one form?” Well, the approach I’ve used to accomplish this task is to ‘emulate’ the submit buttons with regular button controls and changing the form action based on what button was clicked.

In the following example I have a text field and 3 different buttons that will send the string to be searched to Google, MSN or Yahoo depending on what button the user clicks.

The Code

So, the JavaScript looks like this:

<script type="text/javascript">
<!--
    function IsBlankField(field){
	var j = 0;
	var chr;

	for(j; j < field.length; j++){
	    chr = field.charAt(j);

	    if(chr != "" && chr != " " && chr != "\n" && chr != "\t"){
		return false;
	    }
	}

	return true;
    }

    function SendTo(id){
        //shortening form elements
        var myForm      = document.getElementById('myForm');
        var txtSearch   = document.getElementById('wrdSearch');
	var srchURL;    // this variable will hold the URL of the search engine to be used
        var srchVar;    // this variable will hold the value of the text field 'name' which will hold the string to be searched 

        if(IsBlankField(txtSearch.value)){
            //if the form doesn't validate, send this alert
            alert("there's nothing to search for!");
        } else{
            if(id == "google"){
		srchURL = "http://www.google.com/search";
		srchVar = "q";
            } else if(id == "msn"){
		srchURL = "http://search.msn.com/results.aspx";
		srchVar = "q";
            } else{
		srchURL = "http://search.yahoo.com/search";
		srchVar = "p";
            }

            //adding the attribute name and values to the form and text field elements
	    myForm.setAttribute("action", srchURL);
	    txtSearch.setAttribute("name", srchVar);
	    myForm.setAttribute("target", "_blank"); //this, of course, will open a new browser window
	    myForm.setAttribute("method", "get");

            myForm.submit();           

        }
    }
//-->
</script>

And the HTML part looks like this:

<form name="myForm" id="myForm">
    Search: <input type="text" id="wrdSearch"/>
    <input type="button" name="google" id="google" value="Google" onClick="SendTo(this.id)"/>
    <input type="button" name="msn" id="msn" value="MSN" onClick="SendTo(this.id)"/>
    <input type="button" name="yahoo" id="yahoo" value="Yahoo" onClick="SendTo(this.id)"/>
</form>

Explanation

In the JavaScript, the IsBlankField(field) function takes as its argument the text field value and will check if it is empty or has space, new line or tab characters. If it does, the function will return true and the form will not validate.

if(IsBlankField(txtSearch.value)){
    //if the form doesn't validate, send this alert
     alert("there's nothing to search for!");
}

The SendTo(id) function takes one argument (the button id) so that we know what button triggered the event. Then we will add the attributes, with the correct values, to the form and text field elements using the DOM setAttribute() method. As you can see in the code sample, this method takes two parameters: the attribute name and the attribute value.

function SendTo(id){
    ....
    ....
    ....

    myForm.setAttribute("action", srchURL);
    txtSearch.setAttribute("name", srchVar); //this will add the name attribute to the text field element
    myForm.setAttribute("target", "_blank");
    myForm.setAttribute("method", "get");

Then, we submit the form.

    myForm.submit();
    ....
    ....
}

In the HTML, notice that the form does not have its name, action and method attributes (nor the text field has its name attribute). As I explained before, they are being added with the DOM setAttribute() method.

<form name="myForm" id="myForm">
    Search: <input type="text" id="wrdSearch"/>
    ....
    ....
</form>

See the demo.

Compatibility

This will work on Firefox 1.0.7, IE 6, Opera 8.01 and Konqueror 3.4.2 (I haven’t tested this on Safari, but most likely it’ll work :D )

That’s it! Hope this can help to get you started if you ever need to have 2 or more ‘submit’ buttons on one form. And remember, this is one way to do it ;)

J



Using the innerText Property in Firefox

22 09 2005

I’ve read on different forums questions of people asking how they can make the innerText property work in Firefox. Many have suggested to use the innerHTML property instead, but that would not be useful because, obviously, the HTML tags would be either rendered or displayed (an example of the latter would be if we want such text to be displayed in a textarea or text field) giving undesired effects.

Well, in short, Firefox does not support the innerText property. Instead, it supports the textContent property.

So, you could ‘sniff’ what browser the user is using and use the correct property accordingly.
So, you could check for the browser’s feature support to use the correct property accordingly (this is better than sniffing ;) )

Example (updated):

if(document.all){
     document.getElementById('element').innerText = "my text";
} else{
    document.getElementById('element').textContent = "my text";
}

That’s it. Hope it helps ;)

J

Addendum:
See also: The textContent and innerText Properties. This post will help you to make the innerText property work in Safari too ;)






LiveSTRONG