Basic Multiwallet GUI support #11383

luke-jr · 2017-09-21T22:09:11Z

This adds a combobox to the GUI main window and debug window to select the wallet being viewed/used.

jonasschnelli · 2017-09-22T04:32:27Z

src/qt/bitcoin.cpp

-            window->addWallet(BitcoinGUI::DEFAULT_WALLET, walletModel);
-            window->setCurrentWallet(BitcoinGUI::DEFAULT_WALLET);
+            QString WalletName = QString::fromStdString(pwallet->GetName());
+            if (WalletName.endsWith(".dat")) {


Hmm.. not sure if we should hardcode .dat at this point.

It's strictly for stripping.

#11383 (comment)

Hmm.. not sure if we should hardcode .dat at this point.

Note: this is now cleaned up in a later commit ("Get wallet name from WalletModel rather than passing it around")

jonasschnelli · 2017-09-22T04:39:43Z

src/qt/bitcoingui.cpp

@@ -462,6 +462,15 @@ void BitcoinGUI::createToolBars()
        toolbar->addAction(receiveCoinsAction);
        toolbar->addAction(historyAction);
        overviewAction->setChecked(true);
+
+#ifdef ENABLE_WALLET


~~this #ifdef could be avoided/removed, right?~~

Has to bee there I guess because of bool setCurrentWallet(const QString& name); being only defined if ENABLE_WALLET.

jonasschnelli · 2017-09-22T04:55:53Z

Nice!
Works pretty good (played around with it a bit).
I think this may be a first step.

Conceptual issues:

The Combox box without a direct user feedback may lead to selecting the wrong wallet. Android dual/multi-sim works against this by always displaying an icon for the SIM card the action will be executed with. Not 1:1 applicable here... but users should get A) a feedback after switching wallets (are you sure, etc.), B) better visual distinction.
Having the HD / encryption status in the status-bar seems not to be the right place when running multiwallet. It is basically on the same level then the QComboBox.
Wallet name tied to the filename may be not the best choice. Would something speak against having a dedicated walletname per wallet, .. or would this confuse?

Ideas:

Ideally we would have a wallet-navigation-drawer at the left side that consume a more prominent space and make it move obvious which wallet is used.
Critical actions like send-coins should maybe remind the user in the confirmation dialog which wallet it will use.
Users eventually should choose an (coloured) icon per wallet (give a pre-selection of 12 icons) which then will be always visible on a top window level and or displayed when executing critical actions (showing addresses, sending coins, etc.)

Code:

I don't see why the RPC auth argument would be necessary at this point (but I respect your comment that we should discuss that in RPC: Allow rpcauth configs to specify a 4th parameter naming a specific wallet #10615)
I dislike the #Ifdef ENABLE_WALLET clustering. I would reduce it to a minimum and runtime bypass where possible. Especially the #Ifdefs in RPC / http low level code is ugly.

jnewbery · 2017-10-05T22:01:06Z

I think we can fairly easily pull out the contentious rpcauth parts of this to make it more palatable for review. I have a branch here: https://github.com/jnewbery/bitcoin/tree/pr11383.1 which removes the first three commits and makes the one change to rpcconsole.cpp to make this work without rpcauth. It's also rebased on master since there was a merge conflict.

luke-jr · 2017-10-06T02:51:22Z

Rebased, eliminated ENABLE_WALLET stuff added to non-wallet/GUI code, and removed the dependency on rpcauth default wallet stuff.

I did not address @jonasschnelli's GUI comments, because I feel some is best explored separately, as improvements on top of this, while others (confirming wallet changes with a prompt) I think would make the feature annoying to use.

ryanofsky · 2017-10-09T14:19:15Z

src/wallet/rpcwallet.cpp

-static const std::string WALLET_ENDPOINT_BASE = "/wallet/";
-
-CWallet *GetWalletForJSONRPCRequest(const JSONRPCRequest& request)
+void JSONRPCRequestWalletResolver(JSONRPCRequest& jreq, const HTTPRequest& httpreq)


In commit "RPC: Pass wallet through JSONRPCRequest"

Multiwallet rpc calls are broken in this commit (and the whole PR) because JSONRPCRequestWalletResolver is not actually registered or called anywhere. Maybe a commit was lost in the rebase? The problem causes multiwallet.py test to fail with "Wallet file not specified"

ryanofsky · 2017-10-09T14:21:48Z

src/qt/bitcoin.cpp

@@ -252,7 +252,7 @@ public Q_SLOTS:
    QTimer *pollShutdownTimer;
 #ifdef ENABLE_WALLET
    PaymentServer* paymentServer;
-    WalletModel *walletModel;
+    std::vector<WalletModel*> walletModels;


In commit "Qt: Load all wallets into WalletModels"

Might be simpler to make this a vector of unique_ptr, to avoid need for manual deletions. Also since renaming this member anyway, could follow current naming convenetion (m_wallet_models).

Not sure unique_ptr is the right tool for the job here. Renamed.

ryanofsky · 2017-10-09T14:31:08Z

src/qt/walletview.cpp

@@ -95,13 +95,13 @@ void WalletView::setBitcoinGUI(BitcoinGUI *gui)
        connect(this, SIGNAL(message(QString,QString,unsigned int)), gui, SLOT(message(QString,QString,unsigned int)));

        // Pass through encryption status changed signals
-        connect(this, SIGNAL(encryptionStatusChanged(int)), gui, SLOT(setEncryptionStatus(int)));
+        connect(this, SIGNAL(encryptionStatusChanged(int)), gui, SLOT(updateWalletStatus()));


In commit "Qt: Ensure UI updates only come from the currently selected walletView"

This is silently dropping the int status value passed to encryptionStatusChanged, and the int hdEnabled value passed to hdEnabledStatusChanged. Assuming this is intended, you should clean up after this change by deleting the unused parameters from encryptionStatusChanged and hdEnabledStatusChanged declarations and calls.

ryanofsky · 2017-10-09T14:36:56Z

src/qt/rpcconsole.cpp

+        const int wallet_index = ui->WalletSelector->currentIndex();
+        if (wallet_index > 0) {
+            CWalletRef *ppwallet = (CWalletRef*)ui->WalletSelector->itemData(wallet_index).value<void*>();
+            ppwallet = new CWalletRef(*ppwallet);  // Refcount


In commit "Qt: Add wallet selector to debug console"

What does refcount comment mean? Is this a todo suggesting that you want to add refcounting in the future? Should clarify comment or maybe remove it.

ryanofsky · 2017-10-09T14:47:13Z

src/qt/walletmodel.cpp

@@ -65,6 +65,15 @@ WalletModel::~WalletModel()
    unsubscribeFromCoreSignals();
 }

+QString WalletModel::getWalletName() const


In commit "Qt: When multiple wallets are used, include in notifications the name"

Would be nice to follow this up by removing the name parameter passed to BitcoinGUI::addWallet to avoid duplicating .dat stripping logic.

luke-jr · 2017-10-12T07:34:10Z

(Note Travis failure is due to the batch request bug fixed in #11277)

jnewbery · 2017-10-12T18:02:34Z

#11277 is merged. Can you rebase?

luke-jr · 2017-10-12T19:58:25Z

Rebased

Github-Pull: bitcoin#11383 Rebased-From: ad100ff42c2cec31ae4f1fb9c94ad00c4eedaa86

Github-Pull: bitcoin#11383 Rebased-From: 6445a935e6daeadd38c3d8633083ca45fc2cfe1f

jonasschnelli

Lightly tested ACK: 555eec1

Works as intended.
Visually not very elegant, but an acceptable first step.

jonasschnelli · 2017-11-16T20:33:10Z

src/httprpc.cpp

+    PrepareJSONRPCRequestCallbacks.connect(preparer);
+}
+
+void UnregisterJSONRPCRequestPreparer(const JSONRPCRequestPreparer& preparer)


This gets never called, right?

Right, nor does /wallet/ get unregistered... Fixing

jonasschnelli · 2017-11-16T20:49:45Z

src/httprpc.cpp

@@ -236,6 +251,9 @@ bool StartHTTPRPC()
 #ifdef ENABLE_WALLET
    // ifdef can be removed once we switch to better endpoint support and API versioning
    RegisterHTTPHandler("/wallet/", false, HTTPReq_JSONRPC);
+
+    void JSONRPCRequestWalletResolver(JSONRPCRequest& jreq, const HTTPRequest& httpreq);


use externoutside function scope?

Isn't extern not needed (nor desirable in our codebase) for function references?

jonasschnelli · 2017-11-16T20:57:39Z

src/httprpc.cpp


 /** WWW-Authenticate to present with 401 Unauthorized response */
 static const char* WWW_AUTH_HEADER_DATA = "Basic realm=\"jsonrpc\"";

+boost::signals2::signal<void (JSONRPCRequest&, const HTTPRequest&)> PrepareJSONRPCRequestCallbacks;


Isn't the HTTPRequest object unused? Why pack it into the signal, future extensions?

jonasschnelli · 2017-11-16T21:00:21Z

src/qt/rpcconsole.cpp

-                                // in Qt, use always the wallet with index 0 when running with multiple wallets
-                                QByteArray encodedName = QUrl::toPercentEncoding(QString::fromStdString(vpwallets[0]->GetName()));
-                                req.URI = "/wallet/"+std::string(encodedName.constData(), encodedName.length());
+                            CWalletRef * const ppwallet = (CWalletRef*)ppwallet_v;


looks like we should use a C++11 (shared?) pointer here to avoid the manual memory management.

Instead of the whole signal & gluing together the wallet and the RPC server, why not just refactor out GetWalletForJSONRPCRequest and use the same logic here? Seems much more understandable and less code.

promag

Tested, but will look the code. Works as expected.

There are some style issues that could be addressed here or in follow ups:

combo and label in the toolbar use different style (font for instance) than the buttons;
not sure if the right side is the best place in terms of UX;
maybe it should be possible to select wallet from the menu bar (for instance, File -> Wallets -> w1);
window titles could have the wallet name too;
the combo in the console is a bit lost (this time is not in a toolbar);
could show a message in the console when the wallet is changed.

promag · 2017-11-19T22:44:52Z

src/rpc/server.h

@@ -45,8 +47,9 @@ class JSONRPCRequest
    bool fHelp;
    std::string URI;
    std::string authUser;
+    CWallet *wallet;


Do we want wallet here even an opaque pointer?

? This is where it belongs..

Sjors · 2017-11-20T15:07:01Z

I tested the debug console by switching wallets (including no wallet) and using dumpwallet. This seems to work as expected. I also checked that wallet backups from the application menu use the correct wallet.

I agree with @promag's suggestions and added a few below. I don't think any should block this PR, although some might simplify it.

Maybe this is what you were saying, but I have a light preference for reducing UI clutter by removing the dropdowns and instead adding Wallet to the primary menu (between File and Settings). The application title would contain the name of the currently selected wallet, as you suggested. In the Wallet menu, the selected wallet would have a check box, and not be selectable. This has a couple of advantages:

it's more future proof for when there's UI for adding wallets, creating new wallets, etc
you can move Backup Wallet there as well (and rename it to Backup [name of wallet])
you can move Encrypt and change passphrase there (and get rid of Settings menu)
you could open each wallet in its own window (probably not desirable though)
- it would make it easier for users to send funds between wallets, and be more like opening multiple word documents
- I don't know if that's possible (maybe it creates a threading nightmare)
- it makes it more difficult to use the application menu for actions, because you have to clarify which wallet the action refers to (or they need to work nicely with multiple wallets)

I assume #10740 is needed before wallets can be added through the UI?

I like @jonasschnelli's suggestion of allowing each wallet to have a name, so the file name can be inconspicuous. Same for clearly showing the wallet name on the send confirmation screen. I also like the idea of using color, although that can get ugly very quickly.

In case anyone else was confused like me, this how you launch with multiple wallets:
Bitcoin-Qt -wallet=wallet1.dat -wallet=wallet2.dat (and it won't complain if you open a non-existing wallet, e.g. because you left out .dat; it will just create it for you)

Sjors · 2017-11-22T13:04:55Z

src/qt/bitcoingui.cpp

+
+bool BitcoinGUI::multiWallet() const
+{
+    return (WalletSelector->count() > 1);


Using a UI element to count the number of wallets doesn't seem ideal.

I think a better approach is to give QT access to std::vector<CWalletRef> vpwallets or m_wallet_models, and then when the dropdown box (re-)renders itself, it should query that.

That would also avoid passing walletModel instances around like in rpcConsole->addWallet(walletModel);.

Agree with @Sjors

promag · 2017-11-29T00:05:53Z

@luke-jr friendly ping, rebase needed.

laanwj · 2017-12-07T19:05:30Z

As per IRC discussion removing this from high priority for review until review comments get addressed.

luke-jr · 2018-03-02T05:07:35Z

I assume by parse, you mean fabricate...?

The JSONRPCRequest only holds a pointer. This is a good design, unlike making up and then parsing strings internally.

jonasschnelli · 2018-03-02T05:47:36Z

I assume by parse, you mean fabricate...?

Yes. Since the rpcconsole does already fabricate artificial JSONRPCRequests. There is nothing wrong with that. If you want to avoid the artificial JSONRPCRquest creation, you would need to get rid of it completely.

See my commit that...

...remove the unnecessary coupling of server.h and wallet
...removes the signal
...removes "knowhow" of "JSONRPC" from httpserver
...simply continues the JSONRPCRequest fabrication

It also makes this PR more compact (17 additions and 61 deletions.)

jonasschnelli@f8eab48

The other thing that needs fixing is the passing around of pure pointers for CWallet*.
I think, for now, passing around the wallet name could be sufficient (I don't like that,... but seems better then passing around pure pointers). Wallet names are unique for now.

jonasschnelli · 2018-03-02T05:49:19Z

Another tiny conceptual problem is that if one switches the wallet in the main window, it does not automatically switch in the console window (Not sure if it should, but if so, it can be address in a follow up PR).

luke-jr · 2018-03-02T05:53:08Z

I fundamentally disagree with that approach. It is a poor design, and assumes the RPC layer and GUI share some common wallet name table. Perhaps they do for now, but it doesn't excuse the bad design.

jonasschnelli · 2018-03-03T02:32:31Z

I rewrote this PR to avoid the pointer passing and the signal & wallet-server coupling:
https://github.com/bitcoin/bitcoin/compare/master...jonasschnelli:2018/03/multiwallet?expand=1

If I'm alone with my design concerns, then I'm fine dropping it. :)
Would be good the hear other opinions: @ryanofsky @Sjors @promag

Sjors · 2018-03-05T14:49:59Z

@jonasschnelli can you turn that into a PR so it's easier to provide line-by-line feedback? It seems bool BitcoinGUI::multiWallet() still counts UI elements (QComboBox *m_wallet_selector instances). Can that be avoided?

promag · 2018-03-05T17:49:35Z

In the "Sending addresses" and "Receiving addresses" dialogs, it's missing the wallet name. Not sure if the dialog title is the best place.

jonasschnelli · 2018-03-06T04:08:51Z

src/qt/walletview.cpp

+        wallet = walletModel->getWalletName();
+    }
+
+    Q_EMIT incomingTransaction(date, walletModel->getOptionsModel()->getDisplayUnit(), amount, type, address, label, wallet);


Why not remove the bitcoinGui member and always send the wallet name?

When not in multiwallet mode, this causes the "Wallet" line to be omitted from the notification popup.

Since only bitcoingui.cpp listens to the incomingTransaction signal, it could be added or not added there. IMO the signal emitter should not decide wether to show the wallet name or not.

But there is no wallet name in single-wallet mode...

jonasschnelli · 2018-03-07T02:08:43Z

I really like to merge this,... but I would feel pretty uncomfortable to merge code where we unnecessary glue together the RPC/HTTP layer with the wallet as well as manual pointer memory management... this is something that should be fixed in the initial PR IMO rather the count for follow up PR.

luke-jr · 2018-03-07T03:02:52Z

@jonasschnelli This only reduces/removes existing glue between the RPC/HTTP and wallet. It doesn't add any.

The manual pointer management is unavoidable with Qt AFAIK, without resorting to worse, ugly hacks like passing it as a string (which will break as soon as we support wallet unloading).

jonasschnelli · 2018-03-07T03:39:15Z

@jonasschnelli This only reduces/removes existing glue between the RPC/HTTP and wallet. It doesn't add any.

Are you referring to this PR? In this PR you are trying to add a CWallet *wallet; member to JSONRPCRequest which IMO is the wrong direction.

The manual pointer management is unavoidable with Qt AFAIK, without resorting to worse, ugly hacks like passing it as a string (which will break as soon as we support wallet unloading).

Right now, we do identify wallets in multiwallet by it's filename. IMO identifying wallets by it's name is currently superior to passing around CWallet pointer with manual deallocation (delete) (which will break worse in case of dynamic loading/unloading IMO).

jonasschnelli · 2018-03-07T03:47:04Z

The manual pointer management is unavoidable with Qt AFAIK, without resorting to worse, ugly hacks like passing it as a string (which will break as soon as we support wallet unloading).

Right now, we do identify wallets in multiwallet by it's filename. IMO identifying wallets by it's name is currently superior to passing around CWallet pointer with manual deallocation (delete) (which will break worse in case of dynamic loading/unloading IMO).

Clarification: I'm only referring to the RPCConsole here where we simulate a JSONRPCRequest (and therefore need to derive the wallet name anyways to use the the correct URI endpoint).

luke-jr · 2018-03-07T04:36:40Z

Are you referring to this PR? In this PR you are trying to add a CWallet *wallet; member to JSONRPCRequest which IMO is the wrong direction.

Indeed I am. This PR improves the situation by moving the wallet-handling logic out of the RPC/HTTP code, into the wallet-specific module. Having a reference to the wallet for the request, on the request is only reasonable. It is a shortcoming of C++ that it needs to be declared together with the rest of the request class. That could be avoided with a map<string, void*> I suppose, but then we need to do manual allocation of any future refcounted pointer - and the only gain in this is extremely minor.

Right now, we do identify wallets in multiwallet by it's filename.

Only for interfacing to HTTP, an inherently text-based protocol. It makes no sense to do so otherwise, and there are serious shortcomings to this approach. (You could very well end up with a multi-step RPC beginning with one wallet, and finishing in another loaded with the same name after the first was unloaded!)

IMO identifying wallets by it's name is currently superior to passing around CWallet pointer with manual deallocation (delete) (which will break worse in case of dynamic loading/unloading IMO).

This code doesn't do that. The only manual allocation is of a CWalletRef (that is, it is currently a pointer to a pointer). We could store the CWalletRef directly in place of it, but that would break when CWalletRef becomes a refcounting type. The only way to avoid this would be to find a way to teach QVariant to hold a CWalletRef directly.

779c5f9 Qt: hide RPCConsole wallet selector when no wallets are present (Jonas Schnelli) dc6f150 Qt: show wallet name in request dlg in case of multiwallet (Jonas Schnelli) 4826ca4 Qt: show wallet name in send confirmation dlg in case of multiwallet (Jonas Schnelli) cfa4133 GUI: RPCConsole: Log wallet changes (Luke Dashjr) b6d04fc Qt: Get wallet name from WalletModel rather than passing it around (Luke Dashjr) 12d8d26 Qt: When multiple wallets are used, include in notifications the name (Jonas Schnelli) d1ec34a Qt: QComboBox::setVisible doesn't work in toolbars, so defer adding it at all until needed (Luke Dashjr) d49cc70 Qt: Add wallet selector to debug console (Jonas Schnelli) d558f44 Bugfix: RPC: Add missing UnregisterHTTPHandler for /wallet/ (Luke Dashjr) 85d5319 Qt: Ensure UI updates only come from the currently selected walletView (Luke Dashjr) e449f9a Qt: Add a combobox to toolbar to select from multiple wallets (Luke Dashjr) 3dba3c3 Qt: Load all wallets into WalletModels (Luke Dashjr) Pull request description: This is an overhaul of #11383 (plus some additions). It avoids unnecessary coupling of httpserver/jsonrpc and the wallet as well as it avoids pointer pure passing (and pointer deletion) of `CWallet` (plus other minor design changes). Additionally it adds the wallet name to the sendconfirmation and request dialog (in case multiwallet is active) Tree-SHA512: 3d06e18badbc5d1821e488bf1dae463bb0be544cf11b2b618e025812bfdd13c5f39604bb93b4c705313930e7dc4e66f4848b9469ba14871bade58e7a027246a1

jonasschnelli · 2018-03-26T11:59:29Z

Closing in favour of now merged #12610

779c5f9 Qt: hide RPCConsole wallet selector when no wallets are present (Jonas Schnelli) dc6f150 Qt: show wallet name in request dlg in case of multiwallet (Jonas Schnelli) 4826ca4 Qt: show wallet name in send confirmation dlg in case of multiwallet (Jonas Schnelli) cfa4133 GUI: RPCConsole: Log wallet changes (Luke Dashjr) b6d04fc Qt: Get wallet name from WalletModel rather than passing it around (Luke Dashjr) 12d8d26 Qt: When multiple wallets are used, include in notifications the name (Jonas Schnelli) d1ec34a Qt: QComboBox::setVisible doesn't work in toolbars, so defer adding it at all until needed (Luke Dashjr) d49cc70 Qt: Add wallet selector to debug console (Jonas Schnelli) d558f44 Bugfix: RPC: Add missing UnregisterHTTPHandler for /wallet/ (Luke Dashjr) 85d5319 Qt: Ensure UI updates only come from the currently selected walletView (Luke Dashjr) e449f9a Qt: Add a combobox to toolbar to select from multiple wallets (Luke Dashjr) 3dba3c3 Qt: Load all wallets into WalletModels (Luke Dashjr) Pull request description: This is an overhaul of bitcoin#11383 (plus some additions). It avoids unnecessary coupling of httpserver/jsonrpc and the wallet as well as it avoids pointer pure passing (and pointer deletion) of `CWallet` (plus other minor design changes). Additionally it adds the wallet name to the sendconfirmation and request dialog (in case multiwallet is active) Tree-SHA512: 3d06e18badbc5d1821e488bf1dae463bb0be544cf11b2b618e025812bfdd13c5f39604bb93b4c705313930e7dc4e66f4848b9469ba14871bade58e7a027246a1

fanquake added GUI Wallet labels Sep 21, 2017

jonasschnelli reviewed Sep 22, 2017

View reviewed changes

promag mentioned this pull request Sep 25, 2017

Use shared pointer for wallet instances #11402

Closed

luke-jr force-pushed the multiwallet_gui branch from 4a1ef70 to 24ca9cd Compare October 6, 2017 02:46

ryanofsky reviewed Oct 9, 2017

View reviewed changes

luke-jr force-pushed the multiwallet_gui branch from 97bd3e7 to 6445a93 Compare October 12, 2017 19:58

fanquake added this to In progress in Multiwallet support Nov 4, 2017

luke-jr added a commit to bitcoinknots/bitcoin that referenced this pull request Nov 6, 2017

Qt: Rename walletModels to m_wallet_models

51b4432

Github-Pull: bitcoin#11383 Rebased-From: ad100ff42c2cec31ae4f1fb9c94ad00c4eedaa86

luke-jr added a commit to bitcoinknots/bitcoin that referenced this pull request Nov 6, 2017

Qt: Get wallet name from WalletModel rather than passing it around

42ca0f5

Github-Pull: bitcoin#11383 Rebased-From: 6445a935e6daeadd38c3d8633083ca45fc2cfe1f

laanwj added this to Blockers in High-priority for review Nov 16, 2017

luke-jr force-pushed the multiwallet_gui branch from 6445a93 to 555eec1 Compare November 16, 2017 19:47

jonasschnelli approved these changes Nov 16, 2017

View reviewed changes

promag reviewed Nov 19, 2017

View reviewed changes

Sjors mentioned this pull request Nov 20, 2017

[wallet] loadwallet RPC - load wallet at runtime #10740

Merged

Sjors reviewed Nov 22, 2017

View reviewed changes

laanwj removed this from Blockers in High-priority for review Dec 7, 2017

luke-jr added 3 commits December 14, 2017 03:32

Bugfix: RPC: Add missing UnregisterHTTPHandler for /wallet/

cf1eba3

RPC: Add JSONRPCRequest preparation callbacks

bbbc234

RPC: Pass wallet through JSONRPCRequest

76df77f

jonasschnelli reviewed Mar 6, 2018

View reviewed changes

jonasschnelli mentioned this pull request Mar 6, 2018

Multiwallet for the GUI #12610

Merged

laanwj removed this from Blockers in High-priority for review Mar 6, 2018

This was referenced Mar 14, 2018

Refactor: separate gui from wallet and node #10244

Merged

GUI unresponsive during slow operations #10504

Closed

Qt GUI Feature Request: Load wallet file #7675

Closed

MacOSX: wallet.dat content should be stored in Mac OSX Keychain #2545

Closed

jonasschnelli closed this Mar 26, 2018

jonasschnelli removed this from In progress in Multiwallet support Apr 5, 2018

HashUnlimited mentioned this pull request Jul 31, 2018

Multiwallet for the GUI chaincoin/chaincoin#89

Merged

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

Basic Multiwallet GUI support #11383

Basic Multiwallet GUI support #11383

Conversation

luke-jr commented Sep 21, 2017 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jonasschnelli Sep 22, 2017 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

jonasschnelli commented Sep 22, 2017

jnewbery commented Oct 5, 2017

luke-jr commented Oct 6, 2017

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

luke-jr commented Oct 12, 2017

jnewbery commented Oct 12, 2017

luke-jr commented Oct 12, 2017

jonasschnelli 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

promag left a comment • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Sjors commented Nov 20, 2017 • edited

Choose a reason for hiding this comment

Sjors Jan 2, 2018 • edited

Choose a reason for hiding this comment

Choose a reason for hiding this comment

promag commented Nov 29, 2017

laanwj commented Dec 7, 2017

luke-jr commented Mar 2, 2018

jonasschnelli commented Mar 2, 2018

jonasschnelli commented Mar 2, 2018

luke-jr commented Mar 2, 2018

jonasschnelli commented Mar 3, 2018

Sjors commented Mar 5, 2018

promag commented Mar 5, 2018

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

jonasschnelli commented Mar 7, 2018

luke-jr commented Mar 7, 2018

jonasschnelli commented Mar 7, 2018

jonasschnelli commented Mar 7, 2018

luke-jr commented Mar 7, 2018

jonasschnelli commented Mar 26, 2018

luke-jr commented Sep 21, 2017 •

edited

jonasschnelli Sep 22, 2017 •

edited

promag left a comment •

edited

Sjors commented Nov 20, 2017 •

edited

Sjors Jan 2, 2018 •

edited