Hi Viktor,
indeed, you found a bug I introduced in osmo-ggsn 2d6a69e69a4b4cb2b8cc63c4810dae44e5a4d8f6, sorry for that, I didn't notice it since I was testing it with osmo-sgsn instead of using sgsnemu.
You patch is fine, but I'd argue it would be more correct to use "< 1" for the condition since it now returns the amount of addresses parsed (or -1 in case of error). And better add whitespace in the comparison.
Can you please submit the patch to gerrit? Add a reference to its description: Fixes: 2d6a69e69a4b4cb2b8cc63c4810dae44e5a4d8f6 ("Add support for IPv4v6 End User Addresses")
You can add me as reviewer.
Also please note that sgsnemu doesn't yet support IPv4v6 EUAs as opposed to osmo-ggsn. I only added support for it (in the commit introducing the bug) because modems in osmo-gsm-tester were using that one and I needed to support it in osmo-ggsn to have the tests working. As I never used sgsnemu before, I only patched it to account for the changes in libgtp but without adding the support for the new feature there.
Regards,