For example, compared to migration from 3G to 4G networks. As I understand, from the launch of 4G to complete shutdown of 3G it took around 12—14 years.
A reasonably fair comparison. The ISPs had a much stronger incentive to finish the migration, though, because the 3g spectrum could just get turned around and used for 4g after rollout. IPv6 doesn't really have that strong of an incentive structure now that CGNAT is a well-developed technology.
One major difference in the 3G->4G and now 4G->5G conversion was that was largely a single-party change in the end to actually implement. The client and the server hosting an application doesn't care about whether that traffic is over 3G or 4G or IP over Avian Carriers as long as the packets get there in a reasonable time. Going from IPv4 to IPv6 requires lots of very different players to all work together to make the transition, meanwhile for a carrier to go from 3G to 4G its largely on them and their direct contractors.
World IPv6 Day was in 2011, so 15 years since then. This is also requiring a consumer hardware and software upgrade on both the client and server (resource they're accessing). GitHub doesn't have to implement 4G support.