Add SegWit support to importmulti #14454

meshcollider · 2018-10-10T01:58:38Z

Add support for segwit to importmulti, supports P2WSH, P2WPKH, P2SH-P2WPKH, P2SH-P2WSH. Adds a new witnessscript parameter which must be used for the witness scripts in the relevant situations.

Also includes some tests for the various import types.

~~Also makes the change in #14019 redundant, but cherry-picks the test from that PR to test the behavior (@achow101).~~

Fixes #12253, also addresses the second point in #12703, and fixes #14407

DrahtBot · 2018-10-10T04:59:14Z

Reviewers, this pull request conflicts with the following ones:

rpc: Early call once CWallet::MarkDirty in import calls #14303 (rpc: Early call once CWallet::MarkDirty in import calls by promag)
Import watch only pubkeys to the keypool if private keys are disabled #14075 (Import watch only pubkeys to the keypool if private keys are disabled by achow101)

If you consider this pull request important, please also help to review the conflicting pull requests. Ideally, start with the one that should be merged first.

practicalswift · 2018-10-13T13:48:28Z

src/wallet/rpcdump.cpp


-                CTxDestination pubkey_dest = pubKey.GetID();
+            CTxDestination pubkey_dest = pubKey.GetID();


No longer needed?

practicalswift · 2018-10-13T13:50:03Z

src/wallet/rpcdump.cpp

-            if (pubKeys.size() && keys.size() == 0) {
-                const std::string& strPubKey = pubKeys[0].get_str();
+                // Generate the scripts
+                std::vector<unsigned char> vData(ParseHex(strWitnessScript));


Use another variable name to avoid shadowing existing local variable vData :-)

meshcollider · 2018-10-14T00:04:30Z

Thanks for reviewing @practicalswift, both comments addressed :)

achow101

Concept ACK

59f8e547130fc1cb756da7d990f65445ca196f5f could be squashed into 24092b3402693d8a2c39b640db18748b168863e9

acdba74ee011bfcafc9f28b528918815b1de4920 seems to be an unrelated change

achow101 · 2018-10-14T02:36:32Z

src/wallet/rpcdump.cpp

-        if (isP2SH && !IsHex(strRedeemScript)) {
-            throw JSONRPCError(RPC_INVALID_ADDRESS_OR_KEY, "Invalid redeem script");
+        // Force users to provide the witness script in its field rather than redeemscript
+        if (!strRedeemScript.empty() && script.IsPayToWitnessScriptHash()) {


Instead of checking that strRedeemScript is not empty, this should check that strWitnessScript is empty. Or it could check both: !strRedeemScript.empty() && strWitnessScript.empty().

It should be possible to provide a P2WSH scriptPubKey without providing the witness script for it, e.g. watch only, so I don't think that's right

I think checking both covers that case.

Checking both would mean that if witnessScript and redeemScript were both present, it would allow it, which makes no sense if its not a P2WSH address

P2WSH scriptPubKey without providing the witness script for it, e.g. watch only

Do we already support that? And test it? If not, maybe it's easier to just treat it as P2SH until we know the witnessScript.

@Sjors Yes, of course, just importaddress it. It probably works fine without this PR even.

@sipa so we currently support importing a 3... address as watch-only, but the wallet doesn't know or care if it's SegWit or not. The new code here is for a scenario where you do know it's SegWit script.IsPayToWitnessScriptHash(), which means you know the strRedeemScript, but you don't know the strWitnessScript. That seems a strange situation. Why would anyone give you the legacy redeem script but not the Segwit witness script?

@Sjors I have no idea what you're talking about. This is about the case where you're importing a bech32 P2WSH address (no P2SH), without revealing the script inside - which is a totally reasonable thing if you don't care about solvability.

@sipa I see, I may have gotten confused because Github is showing if (isP2SH && !IsHex(strRedeemScript as the deleted line, but this thread is about an earlier code branch.

achow101 · 2018-10-14T02:38:35Z

src/wallet/rpcdump.cpp

            // Import redeem script.
            std::vector<unsigned char> vData(ParseHex(strRedeemScript));
            CScript redeemScript = CScript(vData.begin(), vData.end());
+            CScriptID redeem_id(redeemScript);
+            CScript redeemDestination = GetScriptForDestination(redeem_id);


nit: snake_case

achow101 · 2018-10-14T02:38:42Z

src/wallet/rpcdump.cpp


        // Optional fields.
        const std::string& strRedeemScript = data.exists("redeemscript") ? data["redeemscript"].get_str() : "";
+        const std::string& strWitnessScript = data.exists("witnessscript") ? data["witnessscript"].get_str() : "";


nit: snake_case

achow101 · 2018-10-14T02:40:00Z

src/wallet/rpcdump.cpp

-                const std::string& strPubKey = pubKeys[0].get_str();
+                // Generate the scripts
+                std::vector<unsigned char> vData(ParseHex(strWitnessScript));
+                CScript witnessScript = CScript(vData.begin(), vData.end());


nit: snake_case

achow101 · 2018-10-14T02:40:06Z

src/wallet/rpcdump.cpp

+                std::vector<unsigned char> vData(ParseHex(strWitnessScript));
+                CScript witnessScript = CScript(vData.begin(), vData.end());
+                CScriptID witness_id(witnessScript);
+                CScript witnessDestination = GetScriptForDestination(WitnessV0ScriptHash(witnessScript));


nit: snake_case

achow101 · 2018-10-14T02:42:21Z

src/wallet/rpcdump.cpp

+        // P2PK/P2PKH/P2WPKH
+        } else if (dest.type() == typeid(CKeyID) || dest.type() == typeid(WitnessV0KeyHash)) {
+            if (keys.size() > 1 || pubKeys.size() > 1) {
+                throw JSONRPCError(RPC_INVALID_PARAMETER, "More than private key given for one address");


This error wouldn't make sense if only pubkeys were given.

achow101 · 2018-10-14T02:44:10Z

src/wallet/rpcdump.cpp

                }
+                pubkey = pubkey_temp;
+            }
+            if (pubkey.size()) {


Instead of checking pubkey size, check validity? (e.g. use IsValid() or IsFullyValid())

IsValid() just performs the same size check anyway, but yep I'll do that for the sake of clarity, I don't think we need an IsFullyValid() check here

meshcollider · 2018-10-14T09:22:24Z

Thanks @achow101, addressed all your points

sipa · 2018-10-15T21:16:16Z

@meshcollider I don't think it's necessary to import the pubkeys involved separately anymore since #14424, and in fact that sounds very dangerous (you could be tricked into importing a 2-of-3 multisig where you have 2 of the keys, but then receiving a payment to a P2PKH of the third key, and seeing it treated towards your watch-only balance). NACK until that is fixed (or at least restricted to not have that issue, but I believe that rebasing on 14424 will be sufficient).

meshcollider · 2018-10-16T00:15:02Z

@sipa good point, fixed, and rebased on master to include the public key fix

sipa · 2018-10-17T00:32:50Z

I think there are a few things that need fixing:

There doesn't seem to be code for P2SH-P2WPKH (which matters at least when passing in a pubkey in that case, which won't be imported).
The added tests only cover whether importmulti returns true, not whether it actually imported anything.
The added tests are restricted to watchonly and spendable cases, but no solvable-but-not-spendable ones.

meshcollider · 2018-10-17T02:50:27Z

I believe I've addressed the points, thanks for the feedback so far

sipa

Concept ACK

sipa · 2018-10-17T06:37:26Z

src/wallet/rpcdump.cpp

            }

-            // Import private keys.
+        // P2PK/P2PKH/P2WPKH


I think it may make sense to repeat the trick from the P2SH block of 'descending' into scriptPubKey/address of the redeemscript, to the P2WSH block, and then do P2PK/P2PKH/P2WPKH as a totally new case rather than an else branch. This would make the code work correctly for P2PK/P2PKH nested inside P2WSH (maybe track that you don't permit P2WPKH inside P2WSH though).

sipa · 2018-10-17T06:47:26Z

test/functional/wallet_importmulti.py

@@ -458,6 +468,74 @@ def run_test (self):
                "timestamp": "",
            }])

+        # Import P2WPKH address


There are a number more scenarios that make sense to test, I think:

Address based import (without key/script) of P2WSH-multisig

For all of P2WPKH, P2WSH-multisig, P2SH-P2WPKH, P2SH-P2WSH-multisig, versions with script/pubkey but no private key, and test that the result is solvable.

P2WPKH with private key, result must be spendable (not just watchonly)

P2WSH-multisig and P2SH-multisig with more than 1 key, where some of the keys are specified as private keys, and some as public keys.

Re "Address based import", should this result in the following?

"ismine": false, "iswatchonly": true, "issolvable": false, "isscript": true, "iswitness": true, "witness_version": 0

Re "must be spendable", this means "ismine": true and "iswatchonly": false?

@Sjors Correct.

Sjors

Concept ACK.

I tested 65ecf2d importing a watch-only bech32 address, which worked.

I also tested importing a watch-only p2sh-p2wpkh address like so: importmulti '[{"scriptPubKey": {"address": "2NFiDQiCdRYaRUdWMerff1qSTR4EC7zSBmB"}, "timestamp": "now", "pubkeys": ["03f33112792b07e9994933287ea71069d94176225f745822e47151d39e754d957f"]}]'. This is considered "issolvable": false. Once I add the private key it does know how to solve it.

Code looks good to me, but I don't pretend to understand all the edge cases involved. The tests really help illustrate it, so the more the better, see Sipa's suggestions:
480cba3#r225796932)

Sjors · 2018-10-17T06:07:44Z

src/wallet/rpcdump.cpp

-        // Should have script or JSON with "address".
-        if (!(scriptPubKey.getType() == UniValue::VOBJ && scriptPubKey.exists("address")) && !(scriptPubKey.getType() == UniValue::VSTR)) {
+        bool isScript = scriptPubKey.getType() == UniValue::VSTR;
+        if (!isScript && !(scriptPubKey.getType() == UniValue::VOBJ && scriptPubKey.exists("address"))) {
            throw JSONRPCError(RPC_INVALID_PARAMETER, "Invalid scriptPubKey");


Suggested change

throw JSONRPCError(RPC_INVALID_PARAMETER, "Invalid scriptPubKey");

throw JSONRPCError(RPC_INVALID_PARAMETER, "scriptPubKey must be string with script or JSON with address string");

I'm not sure how the code suggestion feature works. I suspect it creates a new commit, in which case it's better to just copy-paste the text and amend an existing commit.

Sjors · 2018-10-17T06:20:00Z

src/wallet/rpcdump.cpp

@@ -858,32 +853,31 @@ static UniValue ProcessImport(CWallet * const pwallet, const UniValue& data, con
            throw JSONRPCError(RPC_INVALID_PARAMETER, "Incompatibility found between watchonly and keys");
        }

-        // Internal + Label
+        // Internal addresses should not have a label
        if (internal && data.exists("label")) {
            throw JSONRPCError(RPC_INVALID_PARAMETER, "Incompatibility found between internal and label");


Suggest using the same error message as the above comment: Internal addresses should not have a label

Could we also get a test for this?

Sjors · 2018-10-17T06:33:45Z

src/wallet/rpcdump.cpp

-        if (isP2SH && !IsHex(strRedeemScript)) {
-            throw JSONRPCError(RPC_INVALID_ADDRESS_OR_KEY, "Invalid redeem script");
+        // Force users to provide the witness script in its field rather than redeemscript
+        if (!strRedeemScript.empty() && script.IsPayToWitnessScriptHash()) {


P2WSH scriptPubKey without providing the witness script for it, e.g. watch only

Do we already support that? And test it? If not, maybe it's easier to just treat it as P2SH until we know the witnessScript.

Sjors · 2018-10-17T06:40:42Z

src/wallet/rpcdump.cpp

@@ -809,29 +809,24 @@ UniValue dumpwallet(const JSONRPCRequest& request)
 static UniValue ProcessImport(CWallet * const pwallet, const UniValue& data, const int64_t timestamp) EXCLUSIVE_LOCKS_REQUIRED(pwallet->cs_wallet)


Maybe move this function to a different file? Unless dump means dumping ground :-)

I understand if you prefer to keep this PR simple though.

It's specific to importing things, though, which is the scope of rpcdump.cpp?

Sjors · 2018-10-17T07:07:09Z

test/functional/wallet_importmulti.py

@@ -458,6 +468,74 @@ def run_test (self):
                "timestamp": "",
            }])

+        # Import P2WPKH address


Re "Address based import", should this result in the following?

"ismine": false, "iswatchonly": true, "issolvable": false, "isscript": true, "iswitness": true, "witness_version": 0

Re "must be spendable", this means "ismine": true and "iswatchonly": false?

instagibbs · 2018-10-17T14:39:38Z

src/wallet/rpcdump.cpp

-        // Should have script or JSON with "address".
-        if (!(scriptPubKey.getType() == UniValue::VOBJ && scriptPubKey.exists("address")) && !(scriptPubKey.getType() == UniValue::VSTR)) {
+        bool isScript = scriptPubKey.getType() == UniValue::VSTR;
+        if (!isScript && !(scriptPubKey.getType() == UniValue::VOBJ && scriptPubKey.exists("address"))) {
            throw JSONRPCError(RPC_INVALID_PARAMETER, "Invalid scriptPubKey");


Good time to give a more descriptive error?

instagibbs · 2018-10-17T14:42:33Z

src/wallet/rpcdump.cpp

@@ -858,32 +853,31 @@ static UniValue ProcessImport(CWallet * const pwallet, const UniValue& data, con
            throw JSONRPCError(RPC_INVALID_PARAMETER, "Incompatibility found between watchonly and keys");
        }

-        // Internal + Label
+        // Internal addresses should not have a label
        if (internal && data.exists("label")) {
            throw JSONRPCError(RPC_INVALID_PARAMETER, "Incompatibility found between internal and label");


Could we also get a test for this?

instagibbs · 2018-10-17T14:47:11Z

src/wallet/rpcdump.cpp

-            throw JSONRPCError(RPC_INVALID_PARAMETER, "More than private key given for one address");
+        // Force users to provide the witness script in its field rather than redeemscript
+        if (!strRedeemScript.empty() && script.IsPayToWitnessScriptHash()) {
+            throw JSONRPCError(RPC_INVALID_PARAMETER, "Provide witnessscript not redeemscript for P2WSH address");


wording suggestion: "P2WSH addresses have an empty redeemscript. Please provide the witnessscript instead."

instagibbs · 2018-10-17T14:48:27Z

src/wallet/rpcdump.cpp

+        if (!strRedeemScript.empty() && script.IsPayToScriptHash()) {
+            // Check the redeemScript is valid
+            if (!IsHex(strRedeemScript)) {
+                throw JSONRPCError(RPC_INVALID_ADDRESS_OR_KEY, "Invalid redeem script");


suggestion: "Invalid redeem script: must be hex string"

instagibbs · 2018-10-17T14:54:11Z

src/wallet/rpcdump.cpp

            throw JSONRPCError(RPC_INVALID_PARAMETER, "Invalid scriptPubKey");
        }
+        const std::string& output = isScript ? scriptPubKey.get_str() : scriptPubKey["address"].get_str();


while we're here could we rename this to something more self-documenting?

@instagibbs what do you suggest? script_or_address?

instagibbs · 2018-10-17T14:58:55Z

src/wallet/rpcdump.cpp

            if (!pwallet->HaveCScript(redeem_id) && !pwallet->AddCScript(redeemScript)) {
                throw JSONRPCError(RPC_WALLET_ERROR, "Error adding p2sh redeemScript to wallet");
            }

-            CScript redeemDestination = GetScriptForDestination(redeem_id);
+            // Check for P2SH-P2WSH


I understand this is a transient commit, but could there be a TODO P2SH-P2WPKH comment somewhere around here?

instagibbs · 2018-10-17T19:43:14Z

src/wallet/rpcdump.cpp

-                    CKeyID vchAddress = pubkey.GetID();
-                    pwallet->MarkDirty();
-                    pwallet->SetAddressBook(vchAddress, label, "receive");
+        // P2WSH


// (P2SH-)P2WSH

Makes it more clear to me that the above block changed the script

And same thing with: // P2PK/P2PKH/P2WPKH

instagibbs · 2018-10-17T19:47:20Z

src/wallet/rpcdump.cpp

-        }
-
-        // Process. //
+        CScript original_script = script;


scriptpubkey_script more self-explanatory? "original" is context dependent.

instagibbs · 2018-10-17T19:47:33Z

src/wallet/rpcdump.cpp

-
-        // Process. //
+        CScript original_script = script;
+        CTxDestination original_dest = dest;


scriptpubkey_dest more self-explanatory? "original" is context dependent.

meshcollider · 2018-10-18T04:19:48Z

All comments so far are addressed, lots more tests added. Thanks @sipa, @instagibbs, @Sjors

Sjors · 2018-10-18T07:57:06Z

@meshcollider I'm still seeing the same behavior I described above when importing a p2sh-p2wpkh address using just the address and public key.

sipa · 2018-10-18T08:00:21Z

@Sjors You need the redeemscript to make your example solvable (otherwise there is no way for the matcher to go from a P2SH hash to knowing what script the hash is of).

With the private key the weird-evil-but-necessary magic behaviour that causes us to automatically recognize P2WPKH and P2SH-P2WPKH payments to known private keys triggers.

Sjors · 2018-10-18T08:12:49Z

Your wording suggests you're not a fan of adding that weird-evil-but-optional magic behavior to the (very common) case of having one public key?

meshcollider · 2018-10-18T08:19:09Z

@Sjors I'm not really a fan of it, this code is already edge-cased enough without adding more ifs to test for, but I'm happy to add it if you really think its worthwhile.

sipa · 2018-10-18T08:19:48Z

@Sjors No, but I have suggested something else before, which would perhaps also solve your issue, namely that for all single-key schemes you can just provide the pubkey, and importmulti would automatically try the different schemes to see which matches your address/script.

At the very least, something for another PR, though.

meshcollider · 2018-10-18T08:26:18Z

And such a case would be so easy with descriptors too ;)

Sjors · 2018-10-18T09:00:35Z

I'm fine with focusing on descriptors, since those are an order of magnitude more intuitive than the current importmulti syntax.

meshcollider · 2018-10-20T10:47:02Z

Comments addressed + squashed some commits together

sipa · 2018-10-23T01:34:57Z

src/wallet/rpcdump.cpp

-            if (IsValidDestination(dest)) {
-                pwallet->SetAddressBook(dest, label, "receive");
+            // Import into the wallet
+            if (!pwallet->AddWatchOnly(witness_script, timestamp)) {


A small issue still: for P2SH-P2WSH it's unnecessary (and undesirable, I think) to also import the inner P2WSH script as watch-only.

It's not necessary because we're already marking the P2SH-P2WSH script as watch-only, and for solvability all we need is AddCScript, not AddWatchOnly.

It's undesirable because it will cause us to treat payments to the P2WSH address as watchonly IsMine (instead of just the P2SH-P2WSH version).

Good point, fixed

Sjors

I added a few more questions inline, hopefully merely to reduce my own confusion.

My above concerns are addressed in e4a66bb, modulo two tests that @sipa requested and are either missing or I overlooked them.

The added tests only cover whether importmulti returns true, not whether it actually imported anything.

This is done now by checking properties via getaddressinfo.

The added tests are restricted to watchonly and spendable cases, but no solvable-but-not-spendable ones.

Some of the ['solvable'], True) tests now check that ['ismine'], False, but none check for iswatchonly'],False].

Address based import (without key/script) of P2WSH-multisig

Covered in # P2WSH multisig address without scripts or keys and # P2SH-P2WSH 1-of-1 multisig + redeemscript with no private key

For all of P2WPKH, P2WSH-multisig, P2SH-P2WPKH, P2SH-P2WSH-multisig, versions with script/pubkey but no private key, and test that the result is solvable.

I think this one is still missing for P2WSH multisig, between the # P2WSH multisig address without scripts or keys and # Same P2WSH multisig address as above, but now with witnessscript + private keys tests?

P2WPKH with private key, result must be spendable (not just watchonly)

Where "spendable" means ("ismine": true and "iswatchonly": false). This is done.

P2WSH-multisig and P2SH-multisig with more than 1 key, where some of the keys are specified as private keys, and some as public keys.

Missing?

I made a commit that checks solvable for all legacy scenarios as well: Sjors@02d553b

Sjors · 2018-10-23T06:07:50Z

test/functional/wallet_importmulti.py

+        assert_equal(result[0]['success'], True)
+        address_assert = self.nodes[1].getaddressinfo(address['address'])
+        assert_equal(address_assert['iswatchonly'], True)
+        assert_equal(address_assert['solvable'], False)


@sipa can you remind me why a Native Pay-to-Witness-Public-Key-Hash isn't solvable, even though legacy Pay-to-Public-Key-Hash addresses are? Is that just to reduce evil magic like in the P2SH-P2PKH case?
If so, maybe add a comment in the test or in the RPC doc?

In this example it's not solvable because the public key isn't included. It has nothing to do with the script/witness hashing type.

In general, solvability is very easy. Start at the scriptPubKey. Whenever it includes a script hash, we must receive the actual script, and recurse into it. Whenever you encounter a pubkey hash, we must receive the actual public key. A scriptPubKey is solvable whenever you can descend all the way to scripts and pubkeys without any hashes of unknown things.

The only exception to that is that whenever you have a private key X, we also automatically import the P2WPKH and P2SH-P2WPKH versions of it, so no separate importing is needed in that case.

In this example it's not solvable because the public key isn't included.

OK, that makes sense. Actually I misread the legacy test: it does add the public key, which why it's solvable.

sipa · 2018-10-23T20:02:19Z

utACK e4a66bb. As @Sjors points out there are some combinations for which tests can be added, but I think that can be done later.

Please squash fixups?

meshcollider · 2018-10-24T20:29:27Z

Squashed fixups into their respective commits, thanks for the review :)

sipa · 2018-10-25T18:58:25Z

utACK c11875c

instagibbs

utACK

instagibbs · 2018-10-25T19:09:43Z

test/functional/wallet_importmulti.py

+        }])
+        assert_equal(result[0]['success'], True)
+        address_assert = self.nodes[1].getaddressinfo(multi_sig_script['address'])
+        assert_equal(address_assert['solvable'], False)


nit: if "sigsrequired" and "ismine" aren't in results can we assert their absence here?

instagibbs · 2018-10-25T19:12:42Z

test/functional/wallet_importmulti.py

+        }])
+        assert_equal(result[0]['success'], True)
+        address_assert = self.nodes[1].getaddressinfo(multi_sig_script['address'])
+        assert_equal(address_assert['solvable'], True)


nit: ismine?

promag · 2018-10-26T09:28:28Z

src/wallet/rpcwallet.cpp

@@ -3538,6 +3538,7 @@ UniValue getaddressinfo(const JSONRPCRequest& request)
            "  \"address\" : \"address\",        (string) The bitcoin address validated\n"
            "  \"scriptPubKey\" : \"hex\",       (string) The hex-encoded scriptPubKey generated by the address\n"
            "  \"ismine\" : true|false,        (boolean) If the address is yours or not\n"
+            "  \"solvable\" : true|false,      (boolean) If the address is solvable by the wallet\n"


Could have release note.

promag · 2018-10-26T09:53:35Z

src/wallet/rpcdump.cpp

-                    pwallet->UpdateTimeFirstKey(timestamp);
-                }
+        // (P2SH-)P2PK/P2PKH/P2WPKH
+        if (dest.type() == typeid(CKeyID) || dest.type() == typeid(WitnessV0KeyHash)) {


First time using RTTI? Maybe it could be avoided?

BTW, didn't test but I don't see variant::type() in boost 1.47 documentation.

@promag it seems to be there: https://www.boost.org/doc/libs/1_47_0/doc/html/boost/variant.html
I could also do dest.which() == 1 || dest.which() == 4 but that's very unclear, or define a new enum of the types, but that seemed over the top unless there's a good reason to avoid RTTI?

Indeed it's in 1.47. Anyway this is removed in #14565.

achow101 · 2018-10-30T04:35:38Z

utACK 201451b

laanwj · 2018-10-31T16:44:06Z

looks good to me, adding the solvable to getaddressinfo makes a lot of sense, and tests also look good to me
utACK c11875c

@achow101

c11875c Add segwit address tests for importmulti (MeshCollider) 201451b Make getaddressinfo return solvability (MeshCollider) 1753d21 Add release notes for importmulti segwit change (MeshCollider) 353c064 Fix typo in test_framework/blocktools (MeshCollider) f6ed748 Add SegWit support to importmulti with some ProcessImport cleanup (MeshCollider) Pull request description: Add support for segwit to importmulti, supports P2WSH, P2WPKH, P2SH-P2WPKH, P2SH-P2WSH. Adds a new `witnessscript` parameter which must be used for the witness scripts in the relevant situations. Also includes some tests for the various import types. ~Also makes the change in #14019 redundant, but cherry-picks the test from that PR to test the behavior (@achow101).~ Fixes #12253, also addresses the second point in #12703, and fixes #14407 Tree-SHA512: 775a755c524d1c387a99acddd772f677d2073876b72403dcfb92c59f9b405ae13ceedcf4dbd2ee1d7a8db91c494f67ca137161032ee3a2071282eeb411be090a

b985e9c Add release notes for importmulti descriptor support (MeshCollider) fbb5e93 Add test for importing via descriptor (MeshCollider) 9f48053 [wallet] Allow descriptor imports with importmulti (MeshCollider) d2b381c [wallet] Refactor ProcessImport() to call ProcessImportLegacy() (John Newbery) 4cac0dd [wallet] Add ProcessImportLegacy() (John Newbery) a1b25e1 [wallet] Refactor ProcessImport() (John Newbery) Pull request description: ~~Based on #14454 #14565, last two commits only are for review.~~ Best reviewed with `?w=1` Allows a descriptor to be imported into the wallet using `importmulti` RPC. Start and end of range can be specified for ranged descriptors. The descriptor is implicitly converted to old structures on import. Also adds a simple test of a P2SH-P2WPKH address being imported as a descriptor. More tests to come, as well as release notes. Tree-SHA512: 160eb6fd574c4ae5b70e0109f7e5ccc95d9309138603408a1114ceb3c558065409c0d7afb66926bc8e1743c365a3b300c5f944ff18b2451acc0514fbeca1f2b3

Summary: Partial backport of Core [[bitcoin/bitcoin#14454 | PR14454]] bitcoin/bitcoin@f6ed748 bitcoin/bitcoin@c11875c#diff-3dfbfa462305488434b8d8da81f99de7R99-R111 Depends on D6061 Test Plan: ninja ninja check ninja check-functional Reviewers: O1 Bitcoin ABC, #bitcoin_abc, deadalnix Reviewed By: O1 Bitcoin ABC, #bitcoin_abc, deadalnix Subscribers: deadalnix Differential Revision: https://reviews.bitcoinabc.org/D6051

meshcollider added RPC/REST/ZMQ Wallet labels Oct 10, 2018

DocOBITCOIN approved these changes Oct 10, 2018

View reviewed changes

practicalswift reviewed Oct 13, 2018

View reviewed changes

achow101 reviewed Oct 14, 2018

View reviewed changes

meshcollider mentioned this pull request Oct 16, 2018

Allow descriptor imports with importmulti #14491

Merged

sipa reviewed Oct 17, 2018

View reviewed changes

Sjors reviewed Oct 17, 2018

View reviewed changes

Sjors mentioned this pull request Oct 17, 2018

Add P2SH-P2WSH support to listunspent RPC #14481

Merged

instagibbs reviewed Oct 17, 2018

View reviewed changes

DrahtBot mentioned this pull request Oct 20, 2018

Rpc help helper class #14502

Closed

practicalswift mentioned this pull request Oct 22, 2018

Enable flake8 rule E225 which checks for missing whitespace around op… #14534

Closed

sipa reviewed Oct 23, 2018

View reviewed changes

Sjors reviewed Oct 23, 2018

View reviewed changes

achow101 mentioned this pull request Oct 24, 2018

rpc: Require solvability in importmulti if importing more than the scriptPubKey #14558

Closed

meshcollider added 5 commits October 25, 2018 09:30

Add SegWit support to importmulti with some ProcessImport cleanup

f6ed748

Fix typo in test_framework/blocktools

353c064

Add release notes for importmulti segwit change

1753d21

Make getaddressinfo return solvability

201451b

Add segwit address tests for importmulti

c11875c

sipa mentioned this pull request Oct 25, 2018

Overhaul importmulti logic #14565

Merged

instagibbs approved these changes Oct 25, 2018

View reviewed changes

promag reviewed Oct 26, 2018

View reviewed changes

Sjors mentioned this pull request Oct 30, 2018

p2sh-p2wpkh && bech32 support for getkeypool bitcoin-core/HWI#9

Closed

laanwj merged commit c11875c into bitcoin:master Oct 31, 2018

meshcollider deleted the 201810_importmulti_segwit_support branch October 31, 2018 21:40

laanwj removed this from Blockers in High-priority for review Nov 1, 2018

fanquake mentioned this pull request Nov 12, 2018

rpc: importmulti improvements #12703

Closed

joemphilips mentioned this pull request Dec 30, 2018

Update PSBT test MetacoSA/NBitcoin#627

Closed

bitcoinhodler mentioned this pull request Feb 12, 2019

importmulti changes not backward compatible #15392

Closed

bitcoin locked as resolved and limited conversation to collaborators Sep 8, 2021


		CTxDestination pubkey_dest = pubKey.GetID();
		CTxDestination pubkey_dest = pubKey.GetID();

	throw JSONRPCError(RPC_INVALID_PARAMETER, "Invalid scriptPubKey");
	throw JSONRPCError(RPC_INVALID_PARAMETER, "scriptPubKey must be string with script or JSON with address string");

		@@ -809,29 +809,24 @@ UniValue dumpwallet(const JSONRPCRequest& request)
		static UniValue ProcessImport(CWallet * const pwallet, const UniValue& data, const int64_t timestamp) EXCLUSIVE_LOCKS_REQUIRED(pwallet->cs_wallet)

Add SegWit support to importmulti #14454

Add SegWit support to importmulti #14454

Conversation

meshcollider commented Oct 10, 2018 • edited

DrahtBot commented Oct 10, 2018 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

meshcollider commented Oct 14, 2018

achow101 left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Sjors Oct 18, 2018 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

meshcollider commented Oct 14, 2018

sipa commented Oct 15, 2018

meshcollider commented Oct 16, 2018 • edited

sipa commented Oct 17, 2018 • edited

meshcollider commented Oct 17, 2018

sipa left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Sjors left a comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

meshcollider commented Oct 18, 2018

Sjors commented Oct 18, 2018

sipa commented Oct 18, 2018

Sjors commented Oct 18, 2018

meshcollider commented Oct 18, 2018

sipa commented Oct 18, 2018

meshcollider commented Oct 18, 2018

Sjors commented Oct 18, 2018

meshcollider commented Oct 20, 2018

sipa Oct 23, 2018 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Sjors left a comment

Choose a reason for hiding this comment

Sjors Oct 23, 2018 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

sipa commented Oct 23, 2018

meshcollider commented Oct 24, 2018 • edited

sipa commented Oct 25, 2018

meshcollider commented Oct 10, 2018 •

edited

DrahtBot commented Oct 10, 2018 •

edited

Sjors Oct 18, 2018 •

edited

meshcollider commented Oct 16, 2018 •

edited

sipa commented Oct 17, 2018 •

edited

sipa Oct 23, 2018 •

edited

Sjors Oct 23, 2018 •

edited

meshcollider commented Oct 24, 2018 •

edited