Unable to correctly parse wsdl url

It appears that the wsdl download function cannot correctly save or create the wsdl url properly.

If the wsdl url is configured as ssl, https://crowd.domain.com and the uri is /crowd/services/SecurityServer?wsdl then the wsdl fails to download at all, and a zero byte crowdsoapservice.wsdl file is generated in \sites\default\files.

If the wsdl url is configured as non-ssl http://crowd.domain.com and the uri is /crowd/services/SecurityServer?wsdl then the wsdl is created, but only contains 4 random unicode characters.

The only workaround I have found is to manually generate the wsdl file, ie: wget https://crowd.domain.com/crowd/services/SecurityServer?wsdl --no-check-certificate and replace the empty file with the actual one.

The SSL root cert for the target crowd server is installed on the machine, however it is a self signed root cert and not issued from a real cert provider.

Some ideas:
Do we need a root certificate, or some method ignoring untrusted certificates when pulling from ssl wsdl urls? I know that when connecting a java tomcat instance in this fashion (with a self created root cert) to another ssl session you need to store a root cert in the certroot store, to avoid pkix path building errors.