importmulti falsely warns private keys present when importing descriptor #15742

Command:

importmulti '[{"desc": "sh(wsh(multi(2,[aeee1e6a/49h/1h/0h]tpubDC5bVdaveQYxGFztMShFAjU637kKjrs9FD8ne1oHR7z8tLeQJExsaxWRVeSpcCcYMdp5ecf2LQhuVaeYFZ7YEpvWTJr4nDGUJrmuvLuTSsp/1/*,[f04973fd/49h/1h/0h]tpubDDSx3fQ8QyVN21PqRwobfAeDnjesTRMG3U6MPs2aUW5tYg9Arhz4tBRiUWTc9VrJBxRVXJqSrULfGs5HJZTqMJC7PBtDGT7cVtsySf76wmS/1/*,[149b5aca/49h/1h/0h]tpubDCCrKqf5e4eiXxRsma3QkcoKbd6MS8zxK5NTfsvNerTPVZcZqFMbn12D3qtWNLUBdxFkRhw2ZToRCMCNQzxG1T8dCVNMRuLwCmUngyf3CbB/1/*)))", "keypool": true, "timestamp": "now", "watchonly": true, "internal": true, "range": {"end": 100}}]' '{"rescan": false}'

Result:

[
  {
    "success": true,
    "warnings": [
      "All private keys are provided, outputs will be considered spendable. If this is intentional, do not specify the watchonly flag."
    ]
  }
]

And according to getdescriptoinfo:

getdescriptorinfo "sh(wsh(multi(2,[aeee1e6a/49h/1h/0h]tpubDC5bVdaveQYxGFztMShFAjU637kKjrs9FD8ne1oHR7z8tLeQJExsaxWRVeSpcCcYMdp5ecf2LQhuVaeYFZ7YEpvWTJr4nDGUJrmuvLuTSsp/1/*,[f04973fd/49h/1h/0h]tpubDDSx3fQ8QyVN21PqRwobfAeDnjesTRMG3U6MPs2aUW5tYg9Arhz4tBRiUWTc9VrJBxRVXJqSrULfGs5HJZTqMJC7PBtDGT7cVtsySf76wmS/1/*,[149b5aca/49h/1h/0h]tpubDCCrKqf5e4eiXxRsma3QkcoKbd6MS8zxK5NTfsvNerTPVZcZqFMbn12D3qtWNLUBdxFkRhw2ZToRCMCNQzxG1T8dCVNMRuLwCmUngyf3CbB/1/*)))"

{
  "descriptor": "sh(wsh(multi(2,[aeee1e6a/49'/1'/0']tpubDC5bVdaveQYxGFztMShFAjU637kKjrs9FD8ne1oHR7z8tLeQJExsaxWRVeSpcCcYMdp5ecf2LQhuVaeYFZ7YEpvWTJr4nDGUJrmuvLuTSsp/1/*,[f04973fd/49'/1'/0']tpubDDSx3fQ8QyVN21PqRwobfAeDnjesTRMG3U6MPs2aUW5tYg9Arhz4tBRiUWTc9VrJBxRVXJqSrULfGs5HJZTqMJC7PBtDGT7cVtsySf76wmS/1/*,[149b5aca/49'/1'/0']tpubDCCrKqf5e4eiXxRsma3QkcoKbd6MS8zxK5NTfsvNerTPVZcZqFMbn12D3qtWNLUBdxFkRhw2ZToRCMCNQzxG1T8dCVNMRuLwCmUngyf3CbB/1/*)))#sddrjdjm",
  "isrange": true,
  "issolvable": true,
  "hasprivatekeys": false
}

It seems that ProcessImportDescriptor finds no pubkeys in this descriptor (pubkey_map is empty), so it concludes that, trivially, all privkeys are provided.

This probably has the same cause as #15743.

Yeah I'm going to close this as a duplicate of #15743, which is more substantive, because the cause is the same.

Reopened because the fix for #15743 won't end up fixing this.

THAT fix is going to cause key origins to be successfully record for multisigs in ProcessImportDescriptor, but according to @sipa it is intentional that pubkeys are not imported, and this will not be changed.

Possibly all that needs to be changed is the comment:

    // Check if all the public keys have corresponding private keys in the import for spendability.
    // This does not take into account threshold multisigs which could be spendable without all keys.	    
    // Thus, threshold multisigs without all keys will be considered not spendable here, even if they are,
    // perhaps triggering a false warning message. This is consistent with the current wallet IsMine check.

The comment claims that we will falsely warn that threshold multisigs are not spendable, even when they are, but ACTUALLY we falsely warn that they are spendable, even when they are not.

Right, there is an additional issue here; namely that only to-be-imported pubkeys are being analyzed for spendability, rather than all involved keys.

I'll PR a fix soon.

Fixed by https://github.com/bitcoin/bitcoin/pull/15749