Summary
A 15 March 2017 message to bitcoin-dev (public-inbox copy) listed blocks the author's node had rejected with bad-version(0x00000002): version 2 blocks mined after BIP66 made version 3 mandatory. One of them, 363731, is an admitted record with its header. Five are in data/reported-blocks.jsonl by hash only:
| Height |
Hash |
| 363997 |
000000000000000003ae1223f4926ec86100885cfe1484dc52fd67e042a19b12 |
| 364261 |
00000000000000000b6adf92bc192b3c21210f456ab21b5e46951665c74cfab2 |
| 367195 |
0000000000000000116322b5f25826787b01f7a70fb322837b68dff8216cefc4 |
| 386682 |
0000000000000000083cbdbb25c1607527c8f3fdb16f0d048c4439a73b501cb6 |
| 387396 |
00000000000000000afc9fbe7cfe8a6b50502d509ba626beb2e2d6c15d1d3ee3 |
bad-version is a header rule, so an 80-byte header is enough to admit any of them. The heights come from the message. The last two also appear, by hash only, as predecessors in merged-mining records from December 2015.
What has been searched
The message has no attachment; the .bin file names in its body are text, and nothing in the thread links to the files. Current nodes return "Block not found" for all five. Wayback Machine index queries over about 140 explorer URL patterns per hash returned nothing, while the same queries do find pages for other stale blocks. archive.today has no captures. A December 2015 node data directory on the Internet Archive was indexed in full: that node started syncing on 8 December 2015 and holds no version-2 header after 363725. bitcoinsearch.xyz, GitHub code search and Common Crawl indexes return only copies of the message.
Everything except the Merkle root and the nonce can be inferred from the canonical chain (the previous block, the difficulty bits, the approximate time), which is not enough to rebuild a header. The message's author, bfd@cock.lu, has been emailed for the headers and has not replied. If you ran a node in July or December 2015 that kept rejected headers, they would close this.
Summary
A 15 March 2017 message to bitcoin-dev (public-inbox copy) listed blocks the author's node had rejected with
bad-version(0x00000002): version 2 blocks mined after BIP66 made version 3 mandatory. One of them, 363731, is an admitted record with its header. Five are indata/reported-blocks.jsonlby hash only:000000000000000003ae1223f4926ec86100885cfe1484dc52fd67e042a19b1200000000000000000b6adf92bc192b3c21210f456ab21b5e46951665c74cfab20000000000000000116322b5f25826787b01f7a70fb322837b68dff8216cefc40000000000000000083cbdbb25c1607527c8f3fdb16f0d048c4439a73b501cb600000000000000000afc9fbe7cfe8a6b50502d509ba626beb2e2d6c15d1d3ee3bad-versionis a header rule, so an 80-byte header is enough to admit any of them. The heights come from the message. The last two also appear, by hash only, as predecessors in merged-mining records from December 2015.What has been searched
The message has no attachment; the
.binfile names in its body are text, and nothing in the thread links to the files. Current nodes return "Block not found" for all five. Wayback Machine index queries over about 140 explorer URL patterns per hash returned nothing, while the same queries do find pages for other stale blocks. archive.today has no captures. A December 2015 node data directory on the Internet Archive was indexed in full: that node started syncing on 8 December 2015 and holds no version-2 header after 363725. bitcoinsearch.xyz, GitHub code search and Common Crawl indexes return only copies of the message.Everything except the Merkle root and the nonce can be inferred from the canonical chain (the previous block, the difficulty bits, the approximate time), which is not enough to rebuild a header. The message's author, bfd@cock.lu, has been emailed for the headers and has not replied. If you ran a node in July or December 2015 that kept rejected headers, they would close this.