39C3 GSUP fixes

[Takuto88]

This branch contains a couple of bugfixes that were authored during 39C3 for the GSUP implementation of the insert subscriber data / update location procedures. They originally stem from @lynxis's branch here but have been cleaned up for upstreaming: https://github.com/lynxis/pyhss/commits/39c3/

The only commit that I am not sure of is 5abb85d. There was an issue with the ISRs not being triggered correctly. I suspect that this was a first shot at fixing this and the "real" fixes were actually made in 6fcebf1. Maybe @lynxis can chime in and explain?

In any case the change in 5abb85d does not break anything and it did work correctly during the event, therefore I included it.

[osmith42]

Commits look good, feel free to merge!

The only commit that I am not sure of is 5abb85d. There was an issue with the ISRs not being triggered correctly.

I suspect this is needed because json_data may not have all these attributes used in the sendDiameterRequest() call when doing the API call, but the database object has them:

                    diameterClient.sendDiameterRequest(
                        requestType='CLR',
                        hostname=json_data['serving_mme'],
                        imsi=json_data['imsi'], 
                        DestinationHost=json_data['serving_mme'], 
                        DestinationRealm=json_data['serving_mme_realm'], 
                        CancellationType=1
                    )

In any case, as you wrote, it doesn't hurt to have this commit and given that this is what was running at 39C3 I think it makes a lot of sense to merge it.

[lynxis]

About 5abb85d:
I think we should not merge this specific commit yet, because it is fixing some parts, but breaks other a little bit.

@osmith42 is correct about his assumption.

So when looking over the code again. I think json_data isn't modified by UpdateObj() which is good.
But the code is still not correct.

We should check what was changed via the call:

a) if the enabled fields got changed,, lets send a CLR on 4G (as the code does) and also a CLR on 2G/3G
b) if the subscriber is still enabled, send a ISD on 2G/3G and a IDR on 4G.

[lynxis]

@Takuto88 great that you take care to upstream the changes. Thanks a lot!

[Takuto88]

I think we should not merge this specific commit yet, because it is fixing some parts, but breaks other a little bit.

@lynxis: Does it? What exactly? Because this commit ran during the event and I don't recall any issues.

But the code is still not correct.
We should check what was changed via the call:

a) if the enabled fields got changed,, lets send a CLR on 4G (as the code does) and also a CLR on 2G/3G
b) if the subscriber is still enabled, send a ISD on 2G/3G and a IDR on 4G.

That's pretty much what the code does apart from the CLR for 2G/3G. That would be a new feature though and I would consider that out of scope for this specific PR. But I agree that this should happen, so I have created #304 to track this.

Personally, I would merge this as is unless there is something with commit 5abb85d that I am missing.