While developing solutions I have encountered two issues recently and I figured it might be useful to share out here:
1. GetNamedNodeProperty(navigator, property) behaviour when the property has not been set:
a. On the rich client it returns an empty string
b. In the browser it throws an ArgumentException
There are several problems with that:
- Testing on the rich client does not reveal the problem until the form is deployed
- In order to get this to work in the browser one needs to catch ArgumentException - exceptions are bad for flow
- The rich client version cannot differentiate between "not there" and "value is there and it's an empty string"
2. XML-based secondary data sources
a. In the rich client the PIs are stripped
b. In the browser the PIs are present
The rich client implementation breaks a scenario where a user wants to submit a document that is based off a different template to a SharePoint document library.
(more on this here: http://www.infopathdev.com/forums/p/6894/25218.aspx)
David Airapetyan / Software Services ArchitectQdabra® Software / Streamline data gathering to turn process into knowledge