Having inserts with text fragments tends to be painfull because oracle sees many metacharacters (eg & < >).

But it is easy turned off:

set scan off;


create sequence PAGE_PROPERTIES_ID;


create or replace trigger PAGE_PROPERTIES_ID
before insert on PAGE_PROPERTIES
for each row
begin
if :new.id is null then
select PAGE_PROPERTIES_ID.nextval into :new.id from dual;
end if;
end;
/


First the base Object

Javascript has a rather unusual aproach to inheritance. It relies on prototypes. Each object has a prototype that defines it’s blueprint – very much like a class. Lets start with the parent object

  function Parent() {
    this.someVal = ‘some value of the parent’;
    this.getVal = function () {
      return this.someVal;
    }

    this.doStuff = function () {
      return ‘runnig Parent.doStuff()’;
    }
  }

Now create a Child object that inherits

So far for the basics. Lets add a child to this Parent

  function Child() {
    this.someVal = ‘some value of the child’;
    this.getParentVal = function() {
      return Child.prototype.someVal;
    }
    this.doChildishStuff = function() {
      return this.doStuff() + ‘ plus doing my own stuff’;
    }
  }
  Child.prototype = new Parent();

Child.prototype = new Parent();  tells the child Class that it should go look into the Parent class if you call some function on it that it does not have itself. If Parent does not have it, maybe there is a .prototype of Gradparent that has it.

Now lets try it out

  var par = new Parent();
  alert(‘par.someVar: ‘ + par.someVal);
  alert(‘par.getVal(): ‘ + par.getVal() );
  alert(‘par.doStuff(): ‘ + par.doStuff() )

  var kid = new Child();
  alert(‘kid.someVal: ‘ + kid.someVal);
  alert(‘kid.doStuff(): ‘ + kid.doStuff())
  alert(‘kid.doChildishStuff(): ‘ + kid.doChildishStuff())
  alert(‘kid.getParentVal(): ‘ + kid.getParentVal())

Works neatly. You can go along and access the prototypes properties if you need it, you simply overwrite a method that does not behave like you would like it. This should help you if you need to do inheritance in Javascript.

There is another way to reuse code: parasitic inheritance

But you can do other tricks to avoid doublicate coding. There is a trick i like to call parasitic inheritance. As functions are variables and accessable you can borrow them from other objects. Without the hassle of inheritance you can reuse code. Let’s see

  function Other () {}
  Other.prototype.someFunction = function() {
    return ‘parasiticFunction original from Unrelated’;
  }

  Child.prototype.doTrick = Other.prototype.someFunction;
  alert(‘Kid does a neat trick :’ + kid.doTrick())

Other is a different Object with a method someFunction. By assigning Child.prototype.doTrick to that variable you get the function.

Changing just one Instance of a running Object

But maybe you do not want to change the behaviour of all Child objects, just that of one instance. Well, there is a way to do that

  function onTheFly() {
    return ‘adding functions to an instance at runtime ‘;
  }
  kid.otherTrick = onTheFly;

  alert(‘kid got a new trick: ‘ + kid.otherTrick() );

Instead of messing with Child.prototype that changes all running Childs you can just create a new variable in kid (or overwrite another) with a new function

The documentation about the more advanced features of javascript are rather poor. So instead of complaining I’ll put something together myself.

The basic function-based approch to javascript coding is basic and well documented. Let’s skip this part. I’ll assume you have some knowlange of obect orientated programing and understand the difference between object and class. The class beeing the blueprint, the object a concreate instance to work with.

Object orientation in JS is proberbly different to what most coders are familiar with. And there are several ways to do it.

About Javascipt variables

JS is a loosly typed language. A variable may either be primitive (eg an integer) an object (eg String) or a function. You can assign any of those types to variables. Objects have properties. So you can go along setting properties to functions. Further you can go along on a object assign a property of type function to it – just so, on the fly. As it is a loosley typed language you can even assign a function with a different signature to an objects property. Now show me how to do that in your language.

Lets start with a object literal.

I like to think of it as kind of inline singelton. You can go ahead and write the object directly. No class required, just rolling it as you go along. Perfect for a fire-and-forget object to pass to another function – much like jquery and extjs expect it as a function parameter. Or a singelton – that is an object that can only have one and only one instance. Like a Validator. Write one, call anywhere. I admit the syntax looks a bit strange :

var Validator = {
  isNumeric: function(val) {
    return !isNaN(val);
  },
  
  isInt: function(val) {
    var res = false;
    var tested = val.toString();
    if ( val.match(/^\d+$/) ) {
      return true;
    }
    return res;
  }
}

if (! Validator.isNumeric(“tintifax”)) {
  alert(“not an Number”)
}

So between the {} of  Validator you write the property followed by a “:” and then the value – in this case the function. After each variable definition you enter a “,” with the exception of the last entry. As soon as the code is parsed you have an object literal. Thats the quickest and easiest way to write objects.

Now for classes

function Car(name) {

  // a public property. you can access it from outside
  this.name = “noname”;
  
  // nasty hack for accessing private vars out of public context
  var self = this;
  
  // contructor code here – init the cars name
  if (name) {
    this.name = name;
  }

  // a private property, No way of accessing it from outside
  var isCleaned = false;
  
  // a public function
  this.honk = function() {
    return “basic car honk from ” + this.name + ” <br />”; // accessing public property
  };

  // public function getting a private property
  this.getCleaned = function() {
    return isCleaned;
  };

  // public function for setting a private property
  this.setCleaned = function(isCleaned) {
    this.cleaned = isCleaned;
    return this.cleaned;
  };

  // private function
  var emptyAshTray = function() {
    ashtrayFull = false;
    return “Ashtray of ” + self.name + ” emptied (using correct self.name )<br />”;
  };

  // public function calling private functions
  this.clean = function () {
    var retVal = “”;
    retVal += emptyAshTray();
    isCleaned = true;
    retVal += “basic car ” + this.name + ” is now clean<br />”;
    return retVal;
  };

}

// alternative version for writing public methods
Car.prototype.goFast = function() {
  return “car ” + this.name + ” goes fast<br />”;
};

var focus = new Car(“Focus”);
alert(“Car name: ” + focus.name);
alert(“car goes fast: ” + focus.goFast());
alert(“clean the car: ” + focus.clean())

The source code should speak for itself. Defining a Function creates the Class. Anything prefixed with “this” is public. Anything declared with “var” is private. There is a pitfall for accessing public variables from private scope (look for the hack with the self variable). There is second way of adding public variables (that can refer to functions or methods, remember) using the Cars prototype.

Explaining the Prototype

Javascript uses something called prototype-based inheritance. Every Class has a prototype shared by all its instances. You can modify the prototype to extend or alter all running instances. Read that again – modify the prototype to modify all running instances. As soon as you type String.prototype.shorten = fuction() {} all Strings have a shorten method. This is powerfull. You can not do this with object literals as they have no class, but you can alter the Validator object literal by saying Validator.isPink = function {}
 

Next: The Javascript way of doing Inheritance

Seit über einer Dekade arbeite ich im Web-Bereich. Immer wieder hatte ich mit Grafiker zu tun die gut für den Print-Bereich ausgebildet und erfahren waren. Aber leider sind nicht alle Grafiker auch erfahren im Web-Bereich und nicht alles Wissen aus dem Print ist direkt anwendbar. In einem größeren Projekt ist dieser Guideline entstanden das recht viele Web-Fragen beantwortet.

@todo: translate to english

Guidelines für Design im Web

If you rather drive your html guy screaming mad follow this article: 20 things to drive web developers crazy

MySQL features a query cache that keeps Select Statements in memory. My first tests look very promising.

To turn it on and allow 100 MB of RAM to be used edit your my.conf:

query_cache_type = 1
query_cache_size = 104857600

to check how its doing:
SHOW STATUS where variable_name like “%cache%”

See http://dev.mysql.com/doc/refman/5.1/en/query-cache.html

Now this is cool:

svnadmin dump /var/oforge/svn/speedTest/  | ssh tinifax@collab.kuplen.at “cat > /tinifax/tmp/speedtest.oforge.dmp”

Streams a big file (in fact any standard out) trough an ssh tunnel to another server

It’s good not everything is done as admin in the newer windows versions, but some things are just a pain to do if you do not have a admin shell. But there is a way

Hit the Windows button and type cmd (nothing new here). But now enter CTL + SHIFT + ENTER confirm to run it as admin and you have your root shell

It’s possible to send data from the reverse proxy to the server doing the work using HTTP Headers. This includes information if HTTPS is used (only the front server knows that) or geoIP information.

Example Apache conf

<Location /eParcel>
# tells if we are running http or https s
RequestHeader set X-Forwarded-HTTPS “%{HTTPS}s”
<Location>

ProxyPass /tintifax http://example.com/tintifax/
ProxyPassReverse /tintifax http://example.com/tintifax/

Yes, we can use CURL to talk to a Sharepoint server and retrieve a document:

curl –ntlm -u username:password http://server/my/personal/tintifax/Shared%20Documents/10MBrandom.doc > example.log/