Using UTF-8 encoding

Place the following code in your WOApplication's dispatchRequest

worequest.setHeader("text/html; charset=UTF-8; encoding=UTF-8", "content-type");

This will make every page use UTF8 encoding when sending content back to the browser.


To use UTF-8 encoding you need four steps:

databaseEncoding = "UTF-8";

flag, so you can create a new database specifying the encoding with:

createdb -U postgres -E UNICODE
{"WebObjects Release" = "WebObjects 5.0"; encoding = NSUTF8StringEncoding; }

If they are not yet encoded in UTF-8 you can copy-paste the above line into your .woo file. Then you can change the encoding of the .wod and .html files (in pbx) by going to the "Format" menu and choosing "file Encodings".

The reason for doing this step is to minimize conversion at run time. If all your component are UTF-8 encoded then WO does not need to do any conversion during the page generation process. I have never actually tested the performance so it may be unwarranted.

<META HTTP-EQUIV="Content-Type" CONTENT="text/html; charset=UTF-8">
public void appendToResponse(WOResponse aResponse, WOContext aContext) {
    // set the encoding before anything is garbled
    aResponse.setContentEncoding( _NSUtilities.UTF8StringEncoding );
    super.appendToResponse(aResponse, aContext);

public void takeValuesFromRequest(WORequest aRequest, WOContext  aContext) {
    aRequest.setDefaultFormValueEncoding( _NSUtilities.UTF8StringEncoding );
    super.takeValuesFromRequest(aRequest, aContext);

Beware of the two spelling when setting the encoding in a WOMessage the name is "UTF8"; or use the private function _NSUtilities.UTF8StringEncoding. Everywhere else it is "UTF-8"

Is this really necessary? Is not enough to add WOMessage.setDefaultEncoding("UTF8") in the application constructor?

And by the way, I don't understand well the difference between WOMessage.defaultEncoding and WORequest.defaultFormValueEncoding

One is about the URL an the other about the data within the form? I'm a bit confused.

David Paules

The first step in modifying the EOModel's connection dictionary was unnecesary for me. I used OpenBase as my backend database. OpenBase allows you to create a database with one of n different encodings (UTF8 is one of them). The EO Framework was capable of storing and retrieving UTF8 character data from OpenBase without any customization of the connection dictionary (ie, the connection dictionary only had the jdbc driver and database URL and user/password).

Since you must modify the session's appendToResponse method, you can simplify the process of adding the META tags to all your pages by doing it in code once. Simply add

aResponse.setHeader("text/html; charset=UTF-8; encoding=UTF-8", "content-type");

before calling

super.appendToResponse(aResponse, aContext);

Another person adds: Note, this is NOT the same as adding a meta tag to all your pages at once. It's adding a header to all your responses. But that should be just as good, or better. The META tag is really just meant to simulate the header.

For WebObjects 4.x using ObjectiveC you will need to following code in the WOComponent class:

-(void) appendToResponse: (WOResponse *) r inContext: (WOContext *) c
  [r setHeader.at"text-html. charsetUTF-8. encoding=UTF8". forKey.at"Content-type"];
  [r setContentEncoding.NSUTF8StringEncoding];
  [super appendToResponse.r inContext.c];

-(void) takeValuesFromRequest: (WORequest *) r inContext: (WOContext *) c
  [r setFormValueEncodingDetectionEnabled.NO];
  [r setDefaultFormValueEncoding.NSUTF8StringEncoding];
  [super takeValuesFromRequest.r inContext.c];

You also need to disable the encoding detection... at least I had to do it in my case.