{quote} IMHO recursive interpolation is what should be the "exception " and "simple" interpolation should be the norm. {quote} But is the problem really recursive interpolation? Is it not just the question on what to do with unresolved parameters which is really an orthogonal problem?
{quote} recursive interpolation seems very edge case to me and I was surprised it was the default. {quote} TBH, I think the current interpolation algorithm might indeed be a bit too complicated. However, we are bound to the spec.
[~gunnar.morling] , I agree that writing and providing a custom message interpolator is possible and in the intend of the spec. However, so are provider specific options. With the custom message interpolator you need to write code, whereas with the provider specific option you should set a property.
Regarding the custom interpolator. I don't think that simple delegation, followed by removal of '{' and '}' in the end will do. This does not take into account literal use of '{' and '{'. So now the interpolator needs to parse the message itself again to determine the actual parameters.
|