I agree. It is quite likely that CVE-2018-17144 attacks have happened which is why updating to a fixed wallet is important (2.0.2 unofficial or ohashi3d sources does not matter in this aspect - only difference is 2.0.2 unofficial makes it possible to check how many nodes have updated by advertizing a unique version string). I don't think the chain split is related though. If you look at chainparams.cpp lines 93-96 you see deployment of of BIP68, BIP112, and BIP113 has started 2018-10-01 (i think you noticed this yourself too) and i have seen a 2.0.1 node reject a 1.3.1 block for consensus reason (block versions between 1.3.1 and 2.0.1 are very different and 2.0.1 has started to enforce it's own version). Sadly i did not save the log but in my opinion it that is the reason for the split.
Yeah, BIP68, BIP112 and BIP113 have started from 2018-10-01...though I've not studied them, OP_CSV enough.
I'm not sure how OP_CSV related to the attacks or not... I need more and more studies about blockchain.
There were two block/hash as the candidate at height 248958:
(a) 00000022193d53b1aa17ac7b86fce2ceecbaba36f0e6d2a0a483c227d7aa4d67
(b) 0000002390e6be74008db3fd88a144657de442cf03f72e9f8a2e446fd6f304f1
Here is my quick check thead on my twitter (sorry in Japanese) again:
https://twitter.com/ohashi3d/status/1054404573526294528The block/hash (a) has block version = 2,
and if block version=2 and BIP66 activated, such a block/hash will be rejected by v2.0.1-based wallet:
https://github.com/ohashi3d/yenten/blob/master/src/validation.cpp#L2973and BIP66 was activated at block height 198765 in v2.0.1-based wallet:
https://github.com/conan-equal-newone/yenten/blob/master/src/chainparams.cpp#L81The block/hash (b) has ridiculous/absurd transactions as mentioned in my previous post.
Both block/hash (a)(b) look suspicious for me...
So...
Conan's original v2.0.1-fix (this '-fix' doesn't mean CVE-2018-17144) wallet rejected block/hash (a),
and picked block/hash (b), though it might be polluted by duplicate input attack.
In opposite, v1.3.1 wallet picked block/hash (a) and rejected (b) as duplicate input attack.
...this is my guess for blockchain split mechanism, and CVE-2018-17144-fixed wallet rejected both block/hash in this situation, and got stuck...