ad085f9ba1 multiprocess: Delay wallet client construction (Russell Yanofsky)
Pull request description:
Delay wallet client construction until after logging, thread and other init for two reasons:
- More responsive multiprocess GUI startup. When bitcoin-gui is started this moves the call from bitcoin-gui to bitcoin-node that spawns bitcoin-wallet off of the GUI event thread and onto the background GUI init executor thread.
- Avoids feature_logging.py test failures with bitcoin-node by making bitcoin-wallet logging start after bitcoin-node logging starts,
because the tests are not written to handle the bitcoin-wallet logging init code running first.
This partially reverts commit b266b3e0bf, moving wallet client creation back to the place it was located before.
---
This PR is part of the [process separation project](https://github.com/bitcoin/bitcoin/projects/10).
ACKs for top commit:
laanwj:
code review ACK ad085f9ba1
hebasto:
ACK ad085f9ba1, I have reviewed the code and it looks OK.
Tree-SHA512: 74d957ce2ee096db745c517124f60800185814b06c20db676090e10dce1b90311adbab02865a69731f8c39b9365f9ee14be0830ca1368cac9b474801ea92bad5
d5f985e51f multiprocess: Add new bitcoin-gui, bitcoin-qt, bitcoin-wallet init implementations (Russell Yanofsky)
Pull request description:
Add separate `interfaces::Init` subclasses for `bitcoin-wallet`, `bitcoin-gui`, and `bitcoin-qt` binaries instead of sharing `bitcoind` and `bitcoin-node` init subclasses in different binaries. After this, the new init subclasses can be customized in #10102, so node and wallet code is dropped from the `bitcoin-gui` binary and wallet code is dropped from into the `bitcoin-node` binary.
---
This PR is part of the [process separation project](https://github.com/bitcoin/bitcoin/projects/10).
ACKs for top commit:
lsilva01:
reACK d5f985e
hebasto:
re-ACK d5f985e51f, only suggested changes since my [previous](https://github.com/bitcoin/bitcoin/pull/23006#pullrequestreview-787537444) review.
Tree-SHA512: 6784210bd9ce3a6fbc66852680d0e9bc513c072dc538aeb7f48cb6a41580d3f567ccef04f975ee767a714a4b05d4d87273e94a79abda1b9c25d5ac4bbe752006
451ca244db qt, refactor: Drop intermediate BitcoinApplication::shutdownResult slot (Hennadii Stepanov)
f3a17bbe5f qt: Do not exit and re-enter main event loop during shutdown (Hennadii Stepanov)
b4e0d2c431 qt, refactor: Allocate SendConfirmationDialog instances on heap (Hennadii Stepanov)
332dea2852 qt, refactor: Keep HelpMessageDialog in the main event loop (Hennadii Stepanov)
c8bae37a7a qt, refactor: Keep PSBTOperationsDialog in the main event loop (Hennadii Stepanov)
7fa91e8312 qt, refactor: Keep AskPassphraseDialog in the main event loop (Hennadii Stepanov)
6f6fde30e7 qt, refactor: Keep EditAddressDialog in the main event loop (Hennadii Stepanov)
59f7ba4fd7 qt, refactor: Keep CoinControlDialog in the main event loop (Hennadii Stepanov)
7830cd0b35 qt, refactor: Keep OptionsDialog in the main event loop (Hennadii Stepanov)
13f618818d qt: Add GUIUtil::ShowModalDialogAndDeleteOnClose (Hennadii Stepanov)
Pull request description:
On master (1ef34ee25e) during shutdown `QApplication` exits the main event loop, then re-enter again.
This PR streamlines shutdown process by removing the need to interrupt the main event loop, that is required for #59.
Also, blocking [`QDialog::exec()`](https://doc.qt.io/qt-5/qdialog.html#exec) calls are replaced with safer [`QDialog::show()`](https://doc.qt.io/qt-5/qwidget.html#show), except for `SendConfirmationDialog` as that change is not trivial (marked as TODO).
The [`QDialog::open()`](https://doc.qt.io/qt-5/qdialog.html#open) was not used because the actual modality mode (application modal or window modal) of a dialog depends on whether it has a parent.
This PR does not change behavior, and all touched dialogs are still application modal.
As a follow up, a design research could suggest to make some dialogs window modal.
NOTE for reviewers: quitting app while a dialog is open (e.g., via systray icon menu) must work fine.
ACKs for top commit:
laanwj:
Code review and lighly tested ACK 451ca244db
promag:
ACK 451ca244db, just changed signal to `quitRequested`.
Tree-SHA512: ef01ab6ed803b202e776019a4e1f592e816f7bc786e00574b25a0bf16be2374ddf9db21f0a26da08700df7ef0ab9e879550df46dcfe3b6d940f5ed02ca5f8447
Delay wallet client construction until after logging, thread and other
init for two reasons:
- More responsive multiprocess GUI startup. When bitcoin-gui is started
this moves the call from bitcoin-gui to bitcoin-node that spawns
bitcoin-wallet off of the GUI event thread and onto the background GUI
init executor thread.
- Avoids feature_logging.py test failures with bitcoin-node by making
bitcoin-wallet logging start after bitcoin-node logging starts,
because the tests are not written to handle the bitcoin-wallet logging
init code running first.
This partially reverts commit b266b3e0bf,
moving wallet client creation back to the place it was located before.
Add separate init implementations instead of sharing existing bitcoind
and bitcoin-node ones, so they can start to be differentiated in
upcoming commits with node and wallet code no longer linked into the
bitcoin-gui binary and wallet code no longer linked into the
bitcoin-node binary.
Use interfaces::Init::make* methods instead of interfaces::Make*
functions, so interfaces can be constructed differently in different
executables without having to change any code. (So for example
bitcoin-gui can make an interfaces::Node pointer that communicates with
a bitcoin-node subprocess, while bitcoin-qt can make an interfaces::Node
pointer that starts node code in the same process.)
93cc53a2b2 gui: Unregister wallet notifications before unloading wallets (Russell Yanofsky)
Pull request description:
This change was originally part of both bitcoin/bitcoin#10102 and bitcoin/bitcoin#19101 and is required for both because it avoids the IPC wallet implementation in bitcoin/bitcoin#10102 and the WalletContext implementation in bitcoin/bitcoin#19101 needing to deal with notification objects that have stale pointers to deleted wallets.
ACKs for top commit:
promag:
Code review ACK 93cc53a2b2.
hebasto:
ACK 93cc53a2b2
Tree-SHA512: 805f50a493291ad0f7c48725fbc5058d58ebbdb0770befd51d8aa241209a13f8a46f5982481336ab8338cdc83e9017668089a71deccf1587308e841cf8697825
a02c970eb0 qt, refactor: Revert explicit including QStringBuilder (Hennadii Stepanov)
3fd3a0fc87 qt, build: Optimize string concatenation (Hennadii Stepanov)
Pull request description:
From [Qt docs](https://doc.qt.io/qt-5/qstring.html#more-efficient-string-construction):
> ... multiple uses of the \[`QString`\] '+' operator usually means multiple memory allocations. When concatenating n substrings, where n > 2, there can be as many as n - 1 calls to the memory allocator.
With this PR
> ... the '+' will automatically be performed as the `QStringBuilder` '%' everywhere.
The change in the `src/Makefile.qt.include` file does not justify submitting this PR into the main repo, IMHO.
ACKs for top commit:
laanwj:
Code review ACK a02c970eb0
Talkless:
utACK a02c970eb0, built successfully on Debian Sid with Qt 5.15.2, but did not check if any displayed strings are "wrong" after refactoring.
jarolrod:
ACK a02c970eb0
Tree-SHA512: cbb476ee96f27c3bd6e125efab74d8bf24bbdb4c30576b3feea45e203405f3bf5b497dd7d3e11361fc825fcbf4b893b152921a9efdeaf73b42d1865d85f0ae84
415fb2e1ab GUI/Intro: Move prune setting below explanation (Luke Dashjr)
2a84c6bcf6 GUI/Intro: Estimate max age of backups that can be restored with pruning (Luke Dashjr)
e2dcd957fa GUI/Intro: Rework UI flow to let the user set prune size in GBs (Luke Dashjr)
f2e5a6b54f GUI/Intro: Abstract GUI-to-option into Intro::getPrune (Luke Dashjr)
62932cc686 GUI/Intro: Return actual prune setting from showIfNeeded (Luke Dashjr)
Pull request description:
Moved from https://github.com/bitcoin/bitcoin/pull/18728
ACKs for top commit:
ryanofsky:
Code review ACK 415fb2e1ab. Changes since last review: mb/gib suffixes, constexpr QOverload expected_backup_days tweaks, new moveonly layout commit
jarolrod:
Tested ACK 415fb2e.
Talkless:
tACK 415fb2e1ab, tested on Debian Sid with Qt 5.15.2.
hebasto:
ACK 415fb2e1ab, my unresolved comments are not blockers, and they could be resolved in follow ups.
Tree-SHA512: bd4882a9c08e6a6eb14b7fb6366983db8581425b4949fea212785d34d8fad9e32fb81ca8c8cdbfb2c05ea394aaf5a746ba2cf16623795c7252c3bdb61d455f00
Introduce an android directory under qt and allow one to package bitcoin-qt for Android by running make apk.
Add bitcoin-qt Android build instructions.
This reverts commit 4e1154dfd1.
After bumping Qt version in depends to 5.12.10 in bitcoin/bitcoin#21376,
there are no reasons to use the Fusion style on macOS.
705c1f0648 qt, refactor: Fix 'buttonClicked is deprecated' warnings (Hennadii Stepanov)
c2f4e5ea1d qt, refactor: Fix 'split is deprecated' warnings (Hennadii Stepanov)
8e12d69961 qt, refactor: Fix 'QFlags is deprecated' warnings (Hennadii Stepanov)
fa5749c805 qt, refactor: Fix 'pixmap is deprecated' warnings (Hennadii Stepanov)
b02264cb5d qt, refactor: Fix 'QDateTime is deprecated' warnings (Hennadii Stepanov)
Pull request description:
[What's New in Qt 5.15](https://doc.qt.io/qt-5/whatsnew515.html#deprecated-modules):
> To help preparing for the transition to Qt 6, numerous classes and member functions that will be removed from Qt 6.0 have been marked as deprecated in the Qt 5.15 release.
Fixes#36
ACKs for top commit:
jonasschnelli:
utACK 705c1f0648
promag:
Tested ACK 705c1f0648 on macos with Apple clang version 11.0.3 (clang-1103.0.32.62) and brew qt 5.15.1.
Tree-SHA512: 29e00535b4583ceec0dfb29612e86ee29bdea13651b548c6d22167917a4a10464af49160a12b05151030699f690f437ebb9c4ae9f130f66a722415222165b44f
24bf17602c gui refactor: Inline SplashScreen::ConnectWallet (Russell Yanofsky)
e4f4350471 refactor: Move wallet methods out of chain.h and node.h (Russell Yanofsky)
b266b3e0bf refactor: Create interfaces earlier during initialization (Russell Yanofsky)
Pull request description:
Add WalletClient interface so node interface is cleaner and don't need wallet-specific methods.
The new NodeContext::wallet_client pointer will also be needed to eliminate global wallet variables like ::vpwallets in #19101, because createWallet(), loadWallet(), getWallets(), etc methods called by the GUI need a way to get a reference to the list of open wallets if it is no longer a global variable.
ACKs for top commit:
promag:
Code review ACK 24bf17602c.
MarcoFalke:
ACK 24bf17602c🐚
Tree-SHA512: a70d3776cd6723093db8912028c50075ec5fa0a48b961cb1a945f922658f5363754f8380dbb8378ed128c8c858913024f8264740905b8121a35c0d63bfaed7cf
ca185cf5a1 doc: Document differences in bitcoind and bitcoin-qt locale handling (practicalswift)
Pull request description:
Document differences in `bitcoind` and `bitcoin-qt` locale handling.
Since this seems to be the root cause to the locale dependency issues we've seen over the years I thought it was worth documenting :)
Note that 1.) `QLocale` (used by Qt), 2.) C locale (used by locale-sensitive C standard library functions/POSIX functions and some parts of the C++ standard library such as `std::to_string`) and 3.) C++ locale (used by the C++ input/output library) are three separate things. This comment is about the perhaps surprising interference with the C locale (2) that takes place as part of the Qt initialization.
ACKs for top commit:
hebasto:
re-ACK ca185cf5a1
Tree-SHA512: e51c32f3072c506b0029a001d8b108125e1acb4f2b6a48a6be721ddadda9da0ae77a9b39ff33f9d9eebabe2244c1db09e8502e3e7012d7a5d40d98e96da0dc44
Add WalletClient interface so node interface is cleaner and don't need
wallet-specific methods.
The new NodeContext::wallet_client pointer will also be needed to eliminate
global wallet variables like ::vpwallets, because createWallet(), loadWallet(),
getWallets(), etc methods called by the GUI need a way to get a reference to
the list of open wallets if it is no longer a global variable.
Also tweaks splash screen registration for load wallet events to be delayed
until after wallet client is created.
This is needed to allow bitcoin-gui to connect to existing node process with
-ipcconnect instead of spawning a new process. It's possible to spawn a new
bitcoin-node process without knowing the current data dir or network, but
connecting to an existing bitcoin-node requires knowing the datadir and network
first.
No change in behavior. Replacing references with pointers allows Node interface
creation to be delayed until later during gui startup next commit to support
implementing -ipcconnect option
Change gui code to use gArgs, Params() functions directly instead of going
through interfaces::Node.
Remotely accessing bitcoin-node ArgsManager from bitcoin-gui works fine in
https://github.com/bitcoin/bitcoin/pull/10102, when bitcoin-gui spawns a new
bitcoin-node process and controls its startup, but for bitcoin-gui to support
-ipcconnect option in https://github.com/bitcoin/bitcoin/pull/19461 and connect
to an existing bitcoin-node process, it needs ability to parse arguments itself
before connecting out.
This change also simplifies https://github.com/bitcoin/bitcoin/pull/10102 a
bit, by making the bitcoin-gui -> bitcoin-node startup sequence more similar to
the bitcoin-node -> bitcoin-wallet startup sequence where the parent process
parses arguments and passes them to the child process instead of the parent
process using the child process to parse arguments.
edc316020e test: Remove duplicate NodeContext hacks (Russell Yanofsky)
Pull request description:
Qt tests currently are currently using two NodeContext structs at the same time, one in interfaces::NodeImpl::m_context, and the other in BasicTestingSetup::m_node, and the tests have hacks transferring state between them.
Fix this by getting rid of the NodeImpl::m_context struct and making it a pointer. This way a common BitcoinApplication object can be used for all qt tests, but they can still have their own testing setups.
Non-test code is changing but non-test behavior is still the same as before.
Motivation for this PR is to be able to remove the "std::move(test.m_node.connman)" and mempool hacks for swapping individual NodeContext members in Qt tests, because followup PR #19099 adds yet another member (wallet_client) that needs to be swapped. After this change, the whole NodeContext struct can be swapped instead of individual members, so the workarounds are less fragile and invasive.
ACKs for top commit:
MarcoFalke:
crACK edc316020e🌮
promag:
ACK edc316020e.
Tree-SHA512: c1650e4127f43a4020304ca7c13b5d9122fb5723aacd8fa1cf855d03c6052fcfb7685810aa2a5ef708561015f0022fecaacbad479295104ca45d2c17579466a4
9c69cfe4c5 Add <datadir>/settings.json persistent settings storage. (Russell Yanofsky)
eb682c5700 util: Add ReadSettings and WriteSettings functions (Russell Yanofsky)
Pull request description:
Persistent settings are used in followup PRs #15936 to unify gui settings between bitcoin-qt and bitcoind, and #15937 to add a load_on_startup flag to the loadwallet RPC and maintain a dynamic list of wallets that should be loaded on startup that also can be shared between bitcoind and bitcoin-qt.
ACKs for top commit:
MarcoFalke:
Approach re-ACK 9c69cfe4c5🌾
jnewbery:
utACK 9c69cfe4c5
Tree-SHA512: 39fcc6051717117c9141e934de1d0d3f739484be4685cdf97d54de967c8c816502b4fd0de12114433beaa5c5b7060c810fd8ae4e2b3ce7c371eb729ac01ba2e1
bd315eb5e2 qt: Get rid of cursor in out-of-focus labels (Hennadii Stepanov)
Pull request description:
After clicking on `QLabel` with selectable text the cursor remains forever:
This PR fixes this visual bug.
Earlier attempts to fix this issue:
- #14577
- #14810 (combined with other UX feature)
ACKs for top commit:
promag:
Code review ACK bd315eb5e2.
laanwj:
Tested ACK bd315eb5e2
Tree-SHA512: 6bf89362412e5ce9a4dec6944b62fe44fc31ca49cda7f6e2eb37e847fac9dccb68bca7ac6877b19e42add2333e40d0b4265757ead105ac0a5d28f8ab43b322c3
Qt tests currently are currently using two NodeContext structs at the same
time, one in interfaces::NodeImpl::m_context, and the other in
BasicTestingSetup::m_node, and the tests have hacks transferring state between
them.
Fix this by getting rid of the NodeImpl::m_context struct and making it a
pointer. This way a common BitcoinApplication object can be used for all qt
tests, but they can still have their own testing setups.
Non-test code is changing but non-test behavior is still the same as before.
Motivation for this PR is to be able to remove the
"std::move(test.m_node.connman)" and mempool hacks for swapping individual
NodeContext members in Qt tests, because followup PR #19099 adds yet another
member (wallet_client) that needs to be swapped. After this change, the whole
NodeContext struct can be swapped instead of individual members, so the
workarounds are less fragile and invasive.
Persistent settings are used in followup PRs #15936 to unify gui settings
between bitcoin-qt and bitcoind, and #15937 to add a load_on_startup flag to
the loadwallet RPC and maintain a dynamic list of wallets that should be loaded
on startup that also can be shared between bitcoind and bitcoin-qt.
After #19176, building the gui on Bionic is failing with:
```bash
CXX qt/qt_libbitcoinqt_a-guiutil.o
qt/bitcoin.cpp: In function 'int GuiMain(int, char**)':
qt/bitcoin.cpp:460:35: error: 'Untranslated' was not declared in this scope
node->initError(strprintf(Untranslated("Error parsing command line arguments: %s\n"), error));
```
The merge commit also failed to compile with the same error:
https://travis-ci.org/github/bitcoin/bitcoin/jobs/696627543
4f49d5222e gui, refactor: Register Qt meta types in application constructor (João Barbosa)
Pull request description:
Removes a warning when running `QT_QPA_PLATFORM=cocoa src/qt/test/test_bitcoin-qt`.
ACKs for top commit:
jonasschnelli:
Re utACK 4f49d5222e
hebasto:
ACK 4f49d5222e, tested on macOS 10.15.5.
Tree-SHA512: e931a022ba83cb0ef04d82544ebd9b18242f8fc2b41443afce4d5c4222f222e8b3517bdb484a1a4f61377c5dceca067d8ccf250da3a727299448e54bec33ed6e
8e08d00598 qt: Use parent-child relation to manage lifetime of OptionsModel object (Hennadii Stepanov)
Pull request description:
Both `BitcoinApplication` and `OptionsModel` classes are derived from the `QObject` class, therefore a parent-child relation could be established to manage the lifetime of an `OptionsModel` object:
5236b2e267/src/qt/optionsmodel.cpp (L29-L30)
This PR does not change behavior.
ACKs for top commit:
jonasschnelli:
utACK 8e08d00598
promag:
ACK 8e08d00598.
Tree-SHA512: 0223dddf5ba28b0bfaefeda1b03b4ff95bf7e7d0c1e7b32368171e561813e22129f2a664f09279fa3b4fa63259b7680d55aa3fe66db9c7ae0039b7f529777ec3
96cb597325 gui: Avoid redundant tx status updates (Russell Yanofsky)
Pull request description:
This PR is part of the [process separation project](https://github.com/bitcoin/bitcoin/projects/10).
In `TransactionTablePriv::index`, avoid calling `interfaces::Wallet::tryGetTxStatus` if the status is up to date as of the most recent `NotifyBlockTip` notification. Store height from the most recent notification in a new `ClientModel::cachedNumBlocks` variable in order to check this.
This avoids floods of IPC traffic from `tryGetTxStatus` with #10102 when there are a lot of transactions. It might also make the GUI a little more efficient even when there is no IPC.
ACKs for top commit:
promag:
Code review ACK 96cb597325.
hebasto:
ACK 96cb597325
Tree-SHA512: fce597bf52a813ad4923110d0a39229ea09e1631e0d580ea18cffb09e58cdbb4b111a40a9a9270ff16d8163cd47b0bd9f1fe7e3a6c7ebb19198f049f8dd1aa46
Both BitcoinApplication and OptionsModel classes are derived from the
QObject class, therefore a parent-child relation could be established to
manage the lifetime of an OptionsModel object.
This commit does not change behavior.
In TransactionTablePriv::index, avoid calling
interfaces::Wallet::tryGetTxStatus if the status is up to date as of the most
recent NotifyBlockTip notification. Store height from the most recent
notification in a new ClientModel::cachedNumBlocks variable in order to check
this.
This avoids floods of IPC traffic from tryGetTxStatus with #10102 when there
are a lot of transactions. It might also make the GUI a little more efficient
even when there is no IPC.
af112ab628 qt: Rename SetPrune() to InitializePruneSetting() (Hennadii Stepanov)
b0bfbe5028 refactor: Drop `bool force' parameter (Hennadii Stepanov)
68c9bbe9bc qt: Force set nPruneSize in QSettings after intro (Hennadii Stepanov)
a82bd8fa57 util: Replace magics with DEFAULT_PRUNE_TARGET_GB (Hennadii Stepanov)
Pull request description:
On master (5622d8f315), having `QSettings` set already
```
$ grep nPruneSize ~/.config/Bitcoin/Bitcoin-Qt-testnet.conf
nPruneSize=6
```
enabling prune option in the intro dialog
```
$ ./src/qt/bitcoin-qt -choosedatadir -testnet
```
has no effect:
```
$ grep Prune ~/.bitcoin/testnet3/debug.log
2019-12-08T10:04:41Z Prune configured to target 5722 MiB on disk for block and undo files.
```
---
With this PR:
```
$ grep Prune ~/.bitcoin/testnet3/debug.log
2019-12-08T10:20:35Z Prune configured to target 1907 MiB on disk for block and undo files.
```
This PR has been split of #17453 (the first two commits) as it fixes an orthogonal bug.
Refs:
- https://github.com/bitcoin/bitcoin/pull/17453#discussion_r345424240
- https://github.com/bitcoin/bitcoin/pull/17453#discussion_r350960201
ACKs for top commit:
Sjors:
Code review re-ACK af112ab628
ryanofsky:
Code review ACK af112ab628. Just suggested changes since last review (thanks!)
promag:
Tested ACK af112ab628. Latest suggestions and changes look good to me.
Tree-SHA512: 8ddad34b30dcc2cdcad6678ba8a0b36fa176e4e3465862ef6eee9be0f98d8146705138c9c7995dd8c0990af41078ca743fef1a90ed9240081f052f32ddec72b9
9924bce317 [gui] intro: enable pruning by default unless disk is big (Sjors Provoost)
c8de347a9d [gui] intro: add prune preference (Sjors Provoost)
1bbc49d207 [gui] intro: inform caller if intro was shown (Sjors Provoost)
1957103786 [gui] add explicit prune setter (Sjors Provoost)
1bccf6a52d [node] add forceSetArg to interface (Sjors Provoost)
Pull request description:
This adds a checkbox to the intro screen to enable pruning from the get go.
If the user has plenty of space, it's unchecked by default:
<img width="671" alt="big" src="https://user-images.githubusercontent.com/10217/63641289-10339000-c6ac-11e9-98d7-caf64dff0da6.png">
If the user has insufficient space it's checked by default:
<img width="897" alt="low" src="https://user-images.githubusercontent.com/10217/63641276-d4002f80-c6ab-11e9-9f5b-a53472f814ff.png">
When the user has barely enough space and is likely to need pruning in the near future, this is shown in yellow and we also check the prune box:
<img width="662" alt="medium" src="https://user-images.githubusercontent.com/10217/63641294-1c1f5200-c6ac-11e9-8ecb-6b69e42b1ece.png">
The cut-off for this 10 GB above `m_assumed_blockchain_size` (`=240` in `chainparams.cpp`).
If the user launches the first time with `-prune=...` then we disable the check box and display the correct size (rounded to GB):
<img width="658" alt="Schermafbeelding 2019-08-24 om 20 23 14" src="https://user-images.githubusercontent.com/10217/63641351-09594d00-c6ad-11e9-94fe-fe5ed562e109.png">
The 2 GB default matches the settings default. The user can't change it in the intro screen, but can change it later. I'm tempted to increase that default to 10 GB, and then have the intro screen reduce it if space is really tight.
Tips for testing:
* move your existing data dir elsewhere
* wipe data dir at every restart (behavior is different if it exists)
* launch with `bitcoin-qt -resetguisettings -lang=en` (there's some space issues in different languages)
* fake your free space by changing `intro.cpp` line 90: `freeBytesAvailable = 5000000000; // 5 GB`
* try both testnet and mainnet, because settings are seperate. In particular note how step 7 in `GuiMain` switches where `QTSettings settings` points to; this had me thoroughly confused on testnet, because I was setting them too early.
ACKs for top commit:
jonasschnelli:
Tested ACK 9924bce317
ryanofsky:
utACK 9924bce317. The changes are very logical, and implement the feature in a clean that way that doesn't add a lot of complication and shouldn't interfere with future improvements. I looked at Luke's branch too, and I think there's also a lot of great stuff there that seems fully compatible with this change.
Tree-SHA512: 9523961451c53aebd347716976bc3a4a398f989dc21e9bbbd357060bd11a8f46c435f068bd421bb31ccb08e55445ef67bc347d8d19a4fb8fde9d6d3f9a3bcbb0
ffea41f530 Enable all tests in feature_config_args.py (Hennadii Stepanov)
66f5c17f8a Use CheckDataDirOption() for code uniformity (Hennadii Stepanov)
7e33a18a34 Fix datadir handling in bitcoin-cli (Hennadii Stepanov)
b28dada374 Fix datadir handling in bitcoin-qt (Hennadii Stepanov)
50824093bb Fix datadir handling in bitcoind (Hennadii Stepanov)
740d41ce9f Add CheckDataDirOption() function (Hennadii Stepanov)
c1f325126c Return absolute path early in AbsPathForConfigVal (Hennadii Stepanov)
Pull request description:
Fix#15240, see: https://github.com/bitcoin/bitcoin/issues/15240#issuecomment-487353760Fix#15745
Fix broken `feature_config_args.py` tests (disabled by MarcoFalke@fabe28a0cdcfa13e0e595a0905e3642a960d3077). All test are enabled now.
This PR is alternative to #13621.
User's `$HOME` directory is not touched unnecessarily now.
~To make reviewing easier only `bitcoind` code is modified (neither `bitcoin-cli` nor `bitcoin-qt`).~
Refs:
- https://github.com/bitcoin/bitcoin/issues/15745#issuecomment-479852569 by **laanwj**
- #16220
Top commit has no ACKs.
Tree-SHA512: 4a4cda10e0b67c8f374da0c9567003d2b566d948e7f8550fe246868b5794c15010e88ea206009480b9cd2f737f310a15e984f920730448f99a895893bed351df
