I am writing to see if you could kindly have a look at the following
patch, which, instead of treating None always as NULL, passes NULL to
the constructor of a class when %implicitconv is defined. The
motivation of this patch is that I would like to make None a valid
input of a customized type but the way SWIG handles None makes it
impossible to pass None to the constructor of the class. I have
attached a simple example that demonstrates the current behavior. The
patch seems to work but I do not know any side effect of it.