Sending HTML via SOAP API

Question

We use Perl Salesforce library WWW::Salesforce to connect our web shop via SOAP API with Salesforce and Financial Force.

For invoices we send a string with HTML code to a richtext field in c2g__InvoiceDescription__c. The string and HTML special chars like < > are not encoded in any way. This worked until some days ago. We didn't change our Perl code, but on updating we get error message:

unable to update c2g__InvoiceDescription__c for a0z2s000002AW0PAAW: Unexpected element {}table during simple type deserialization

I can set the API version in WWW::Salesforce but error seems to be independet from version.

Richtext field definition for WWW::Salesforce is

WWW::Salesforce::Constants->register_type( 'c2g__InvoiceDescription__c', 'Items__c', 'xsd:string' );

For updating invoices upsert method of API is used which gets a hash with fields and values:

{
    'id' => $salesforce_id,
    'Items__c' => '<table>...</table>',
}

What is the correct way to send HTML code via SOAP API to a richtext field? A simple example for sending HTML via SOAP API would be very appreciated (doesn't have to be Perl)

Answer 1

The SOAP API uses XML, which is a variant of SGML (the Standardized General Markup Language), which is also used by HTML. As such, when you place unsafe characters, which are &, <, >, and ", inside a string literally and it's later formatted into XML, you end up with something like:

<Items__c>
  <table>
    ...
  </table>
</Items__c>

This causes the XML to fail to parse with errors like Unexpected element {}table during simple type deserialization. In order to fix it, you need to escape these special characters. The easiest way would probably be with HTML::Entities:

use HTML::Entities;
...
'Items__c' => encode_entities('<table>...</table>');

Technically, the WWW::Salesforce client should do this for you automatically. I'm going to check the source and find out if you're just missing something, or if it's a bug, but either way, you now have a solution to your problem.