mudlet/test/functional_tests/MainConsoleSelectionTest.cpp

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

288 lines
12 KiB
C++
Raw Permalink Normal View History

fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
/***************************************************************************
* Copyright (C) 2026 by Mudlet Makers *
* *
* This program is free software; you can redistribute it and/or modify *
* it under the terms of the GNU General Public License as published by *
* the Free Software Foundation; either version 2 of the License, or *
* (at your option) any later version. *
* *
* This program is distributed in the hope that it will be useful, *
* but WITHOUT ANY WARRANTY; without even the implied warranty of *
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the *
* GNU General Public License for more details. *
* *
* You should have received a copy of the GNU General Public License *
* along with this program; if not, write to the *
* Free Software Foundation, Inc., *
* 59 Temple Place - Suite 330, Boston, MA 02111-1307, USA. *
***************************************************************************/
#include <QtTest/QtTest>
infrastructure: use std::chrono literals for time durations (#9493) #### Brief overview of PR changes/additions Convert raw millisecond integer literals at time-duration call sites to `std::chrono` literals, and add `#include <chrono>` to each touched translation unit. Examples: - `QTimer::singleShot(0, ...)` → `QTimer::singleShot(0ms, ...)` - `mpTimerReplay->setInterval(1000)` → `setInterval(1s)` - `mPendingTimer.start(60000)` → `start(1min)` - `QObject::startTimer(50)` → `startTimer(50ms)` - `QTest::qWait(100)` → `QTest::qWait(100ms)` - `QThread::msleep(10)` → `QThread::sleep(10ms)` This is a semantics-preserving refactor - every duration is kept exactly equal to before (e.g. `1000` ms becomes `1s`, `60000` ms becomes `1min`). No behavioural change. #### Motivation for adding to Mudlet Chrono literals make time durations self-documenting and type-safe. `1s` / `100ms` read unambiguously where a bare `1000` / `100` forces the reader to remember each API's unit, and the compiler now rejects unit mismatches. Only genuine duration arguments were converted - loop counts, scroll-line counts, sizes, ports and the like were deliberately left as plain integers. All targeted APIs provide `std::chrono` overloads in the minimum supported Qt (6.8.2): `QTimer::singleShot`/`start`/`setInterval` (5.8), `QObject::startTimer` (5.9), `QThread::sleep(std::chrono::nanoseconds)` (6.6) and `QTest::qWait(std::chrono::milliseconds)` (6.7). #### Other info (issues closed, discussion etc) Test case: the full application builds cleanly and the entire functional `ctest` suite passes. The only failing test is the known, pre-existing `PasswordMigrationTest` LSan exit-leak (GTK3/fontconfig noise), which is unrelated to this change. Assisted-by: Claude:claude-opus-4-8
2026-07-25 20:24:31 +02:00
#include <chrono>
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
#include "Host.h"
#include "MudletInstanceCoordinator.h"
#include "TMainConsole.h"
#include "TTextEdit.h"
#include "TelnetServerStub.h"
#include "ctelnet.h"
#include "dlgConnectionProfiles.h"
#include "mudlet.h"
infrastructure: use std::chrono literals for time durations (#9493) #### Brief overview of PR changes/additions Convert raw millisecond integer literals at time-duration call sites to `std::chrono` literals, and add `#include <chrono>` to each touched translation unit. Examples: - `QTimer::singleShot(0, ...)` → `QTimer::singleShot(0ms, ...)` - `mpTimerReplay->setInterval(1000)` → `setInterval(1s)` - `mPendingTimer.start(60000)` → `start(1min)` - `QObject::startTimer(50)` → `startTimer(50ms)` - `QTest::qWait(100)` → `QTest::qWait(100ms)` - `QThread::msleep(10)` → `QThread::sleep(10ms)` This is a semantics-preserving refactor - every duration is kept exactly equal to before (e.g. `1000` ms becomes `1s`, `60000` ms becomes `1min`). No behavioural change. #### Motivation for adding to Mudlet Chrono literals make time durations self-documenting and type-safe. `1s` / `100ms` read unambiguously where a bare `1000` / `100` forces the reader to remember each API's unit, and the compiler now rejects unit mismatches. Only genuine duration arguments were converted - loop counts, scroll-line counts, sizes, ports and the like were deliberately left as plain integers. All targeted APIs provide `std::chrono` overloads in the minimum supported Qt (6.8.2): `QTimer::singleShot`/`start`/`setInterval` (5.8), `QObject::startTimer` (5.9), `QThread::sleep(std::chrono::nanoseconds)` (6.6) and `QTest::qWait(std::chrono::milliseconds)` (6.7). #### Other info (issues closed, discussion etc) Test case: the full application builds cleanly and the entire functional `ctest` suite passes. The only failing test is the known, pre-existing `PasswordMigrationTest` LSan exit-leak (GTK3/fontconfig noise), which is unrelated to this change. Assisted-by: Claude:claude-opus-4-8
2026-07-25 20:24:31 +02:00
using namespace std::chrono_literals;
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
extern void qInitResources_mudlet();
extern void qInitResources_qm();
extern void qInitResources_additional_splash_screens();
extern void qInitResources_mudlet_fonts_common();
extern void qInitResources_mudlet_fonts_posix();
void initializeQRCResources();
// Regression test for #3922: left-clicking into the main console to give it
// focus must not leave a one-character selection behind. Such a stray
// selection hijacks Ctrl+C away from the command line (TCommandLine prioritises
// any console selection over the input box).
class MainConsoleSelectionTest : public QObject
{
Q_OBJECT
private:
TelnetServerStub* mpServer = nullptr;
const QString mpHostname = "Test-Selection";
QString mpPort; // assigned the stub's actual ephemeral port in init()
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
const QString mpLocalhost = "localhost";
// Send a screenful of text so a click in the middle of the upper pane lands
// on a real, filled line rather than empty space.
QString fillerText() const
{
const QString line = QString(100, QLatin1Char('X'));
QString message;
for (int i = 0; i < 80; ++i) {
message.append(line);
message.append(QStringLiteral("\r\n"));
}
return message;
}
TTextEdit* upperPane() const
{
auto host = mudlet::self()->getActiveHost();
if (!host || !host->mpConsole) {
return nullptr;
}
return host->mpConsole->mUpperPane;
}
void sendMouse(QWidget* w, QEvent::Type type, Qt::MouseButton button, Qt::MouseButtons buttons, const QPointF& localPos)
{
const QPointF globalPos = w->mapToGlobal(localPos.toPoint());
QMouseEvent event(type, localPos, globalPos, button, buttons, Qt::NoModifier);
QApplication::sendEvent(w, &event);
}
private slots:
void initTestCase() { initializeQRCResources(); }
void init()
{
mpServer = new TelnetServerStub(qApp);
mpServer->start(mpLocalhost, 0); // ephemeral OS-assigned port avoids collisions across concurrent test runs
mpPort = QString::number(mpServer->serverPort());
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
mudlet::start();
mudlet::self()->setupConfig();
mudlet::self()->takeOwnershipOfInstanceCoordinator(std::make_unique<MudletInstanceCoordinator>("MudletInstanceCoordinator"));
mudlet::self()->init();
mudlet::self()->setStorePasswordsSecurely(false);
deleteProfileDirectory(mpHostname);
}
// A plain click (press + a move that stays in the same character cell +
// release) must not create a selection.
void test_clickWithoutDragLeavesNoSelection()
{
mpServer->setWelcomeMessage(fillerText());
startProfile(mpHostname, mpLocalhost, mpPort);
QVERIFY2(waitForTextInBuffer(QString(100, QLatin1Char('X'))), "Filler text never reached the buffer");
mudlet::self()->resize(1200, 800);
infrastructure: use std::chrono literals for time durations (#9493) #### Brief overview of PR changes/additions Convert raw millisecond integer literals at time-duration call sites to `std::chrono` literals, and add `#include <chrono>` to each touched translation unit. Examples: - `QTimer::singleShot(0, ...)` → `QTimer::singleShot(0ms, ...)` - `mpTimerReplay->setInterval(1000)` → `setInterval(1s)` - `mPendingTimer.start(60000)` → `start(1min)` - `QObject::startTimer(50)` → `startTimer(50ms)` - `QTest::qWait(100)` → `QTest::qWait(100ms)` - `QThread::msleep(10)` → `QThread::sleep(10ms)` This is a semantics-preserving refactor - every duration is kept exactly equal to before (e.g. `1000` ms becomes `1s`, `60000` ms becomes `1min`). No behavioural change. #### Motivation for adding to Mudlet Chrono literals make time durations self-documenting and type-safe. `1s` / `100ms` read unambiguously where a bare `1000` / `100` forces the reader to remember each API's unit, and the compiler now rejects unit mismatches. Only genuine duration arguments were converted - loop counts, scroll-line counts, sizes, ports and the like were deliberately left as plain integers. All targeted APIs provide `std::chrono` overloads in the minimum supported Qt (6.8.2): `QTimer::singleShot`/`start`/`setInterval` (5.8), `QObject::startTimer` (5.9), `QThread::sleep(std::chrono::nanoseconds)` (6.6) and `QTest::qWait(std::chrono::milliseconds)` (6.7). #### Other info (issues closed, discussion etc) Test case: the full application builds cleanly and the entire functional `ctest` suite passes. The only failing test is the known, pre-existing `PasswordMigrationTest` LSan exit-leak (GTK3/fontconfig noise), which is unrelated to this change. Assisted-by: Claude:claude-opus-4-8
2026-07-25 20:24:31 +02:00
QTest::qWait(100ms);
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
TTextEdit* pane = upperPane();
QVERIFY2(pane, "No upper pane available");
QVERIFY2(pane->width() > 400 && pane->height() > 100, qPrintable(QStringLiteral("Upper pane too small: %1x%2").arg(pane->width()).arg(pane->height())));
pane->unHighlight();
pane->mSelectedRegion = QRegion();
QVERIFY(pane->mSelectedRegion.isEmpty());
// Click in the middle of the pane (well past the timestamp gutter), then
// a move event at the very same pixel - i.e. no movement to a different
// character cell - then release.
const QPointF clickPos = QRectF(pane->rect()).center();
sendMouse(pane, QEvent::MouseButtonPress, Qt::LeftButton, Qt::LeftButton, clickPos);
sendMouse(pane, QEvent::MouseMove, Qt::NoButton, Qt::LeftButton, clickPos);
sendMouse(pane, QEvent::MouseButtonRelease, Qt::LeftButton, Qt::NoButton, clickPos);
QVERIFY2(pane->mSelectedRegion.isEmpty(), "A click with no drag left a stray selection in the console (regression of #3922)");
}
// Control case: a genuine drag across cells must still produce a selection,
// so the fix above does not over-correct.
void test_dragStillSelects()
{
mpServer->setWelcomeMessage(fillerText());
startProfile(mpHostname, mpLocalhost, mpPort);
QVERIFY2(waitForTextInBuffer(QString(100, QLatin1Char('X'))), "Filler text never reached the buffer");
mudlet::self()->resize(1200, 800);
infrastructure: use std::chrono literals for time durations (#9493) #### Brief overview of PR changes/additions Convert raw millisecond integer literals at time-duration call sites to `std::chrono` literals, and add `#include <chrono>` to each touched translation unit. Examples: - `QTimer::singleShot(0, ...)` → `QTimer::singleShot(0ms, ...)` - `mpTimerReplay->setInterval(1000)` → `setInterval(1s)` - `mPendingTimer.start(60000)` → `start(1min)` - `QObject::startTimer(50)` → `startTimer(50ms)` - `QTest::qWait(100)` → `QTest::qWait(100ms)` - `QThread::msleep(10)` → `QThread::sleep(10ms)` This is a semantics-preserving refactor - every duration is kept exactly equal to before (e.g. `1000` ms becomes `1s`, `60000` ms becomes `1min`). No behavioural change. #### Motivation for adding to Mudlet Chrono literals make time durations self-documenting and type-safe. `1s` / `100ms` read unambiguously where a bare `1000` / `100` forces the reader to remember each API's unit, and the compiler now rejects unit mismatches. Only genuine duration arguments were converted - loop counts, scroll-line counts, sizes, ports and the like were deliberately left as plain integers. All targeted APIs provide `std::chrono` overloads in the minimum supported Qt (6.8.2): `QTimer::singleShot`/`start`/`setInterval` (5.8), `QObject::startTimer` (5.9), `QThread::sleep(std::chrono::nanoseconds)` (6.6) and `QTest::qWait(std::chrono::milliseconds)` (6.7). #### Other info (issues closed, discussion etc) Test case: the full application builds cleanly and the entire functional `ctest` suite passes. The only failing test is the known, pre-existing `PasswordMigrationTest` LSan exit-leak (GTK3/fontconfig noise), which is unrelated to this change. Assisted-by: Claude:claude-opus-4-8
2026-07-25 20:24:31 +02:00
QTest::qWait(100ms);
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
TTextEdit* pane = upperPane();
QVERIFY2(pane, "No upper pane available");
pane->unHighlight();
pane->mSelectedRegion = QRegion();
const QPointF startPos = QRectF(pane->rect()).center();
const QPointF endPos = startPos + QPointF(60, 0); // several character cells to the right
sendMouse(pane, QEvent::MouseButtonPress, Qt::LeftButton, Qt::LeftButton, startPos);
sendMouse(pane, QEvent::MouseMove, Qt::NoButton, Qt::LeftButton, endPos);
sendMouse(pane, QEvent::MouseButtonRelease, Qt::LeftButton, Qt::NoButton, endPos);
QVERIFY2(!pane->mSelectedRegion.isEmpty(), "A real drag failed to create a selection");
}
// Regression case for the review fix: once a drag has genuinely selected
// text, dragging back to the original press cell must collapse the extent
// instead of leaving the previous selection frozen.
void test_dragBackToOriginCollapsesSelectionExtent()
{
mpServer->setWelcomeMessage(fillerText());
startProfile(mpHostname, mpLocalhost, mpPort);
QVERIFY2(waitForTextInBuffer(QString(100, QLatin1Char('X'))), "Filler text never reached the buffer");
mudlet::self()->resize(1200, 800);
infrastructure: use std::chrono literals for time durations (#9493) #### Brief overview of PR changes/additions Convert raw millisecond integer literals at time-duration call sites to `std::chrono` literals, and add `#include <chrono>` to each touched translation unit. Examples: - `QTimer::singleShot(0, ...)` → `QTimer::singleShot(0ms, ...)` - `mpTimerReplay->setInterval(1000)` → `setInterval(1s)` - `mPendingTimer.start(60000)` → `start(1min)` - `QObject::startTimer(50)` → `startTimer(50ms)` - `QTest::qWait(100)` → `QTest::qWait(100ms)` - `QThread::msleep(10)` → `QThread::sleep(10ms)` This is a semantics-preserving refactor - every duration is kept exactly equal to before (e.g. `1000` ms becomes `1s`, `60000` ms becomes `1min`). No behavioural change. #### Motivation for adding to Mudlet Chrono literals make time durations self-documenting and type-safe. `1s` / `100ms` read unambiguously where a bare `1000` / `100` forces the reader to remember each API's unit, and the compiler now rejects unit mismatches. Only genuine duration arguments were converted - loop counts, scroll-line counts, sizes, ports and the like were deliberately left as plain integers. All targeted APIs provide `std::chrono` overloads in the minimum supported Qt (6.8.2): `QTimer::singleShot`/`start`/`setInterval` (5.8), `QObject::startTimer` (5.9), `QThread::sleep(std::chrono::nanoseconds)` (6.6) and `QTest::qWait(std::chrono::milliseconds)` (6.7). #### Other info (issues closed, discussion etc) Test case: the full application builds cleanly and the entire functional `ctest` suite passes. The only failing test is the known, pre-existing `PasswordMigrationTest` LSan exit-leak (GTK3/fontconfig noise), which is unrelated to this change. Assisted-by: Claude:claude-opus-4-8
2026-07-25 20:24:31 +02:00
QTest::qWait(100ms);
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
TTextEdit* pane = upperPane();
QVERIFY2(pane, "No upper pane available");
pane->unHighlight();
pane->mSelectedRegion = QRegion();
const QPointF startPos = QRectF(pane->rect()).center();
const QPointF endPos = startPos + QPointF(60, 0); // several character cells to the right
sendMouse(pane, QEvent::MouseButtonPress, Qt::LeftButton, Qt::LeftButton, startPos);
sendMouse(pane, QEvent::MouseMove, Qt::NoButton, Qt::LeftButton, endPos);
QVERIFY2(!pane->mSelectedRegion.isEmpty(), "The initial drag failed to create a selection");
const QRect expandedSelection = pane->mSelectedRegion.boundingRect();
sendMouse(pane, QEvent::MouseMove, Qt::NoButton, Qt::LeftButton, startPos);
sendMouse(pane, QEvent::MouseButtonRelease, Qt::LeftButton, Qt::NoButton, startPos);
const QRect collapsedSelection = pane->mSelectedRegion.boundingRect();
QVERIFY2(collapsedSelection.width() < expandedSelection.width(),
"Dragging back to the press cell left the earlier selection extent frozen");
}
void cleanup()
{
const QString profilePath = mudlet::getMudletPath(enums::profileHomePath, mpHostname);
// Tear down Mudlet (and with it the live cTelnet connection) before the
// stub server it is talking to, so the socket is closed from the client
// side rather than being yanked out from under an active connection when
// the server is destroyed - the latter ordering can flake or crash.
delete mudlet::self();
delete mpServer;
mpServer = nullptr;
deleteDirectory(profilePath);
}
private:
void startProfile(const QString& hostname, const QString& address, const QString& port)
{
infrastructure: use std::chrono literals for time durations (#9493) #### Brief overview of PR changes/additions Convert raw millisecond integer literals at time-duration call sites to `std::chrono` literals, and add `#include <chrono>` to each touched translation unit. Examples: - `QTimer::singleShot(0, ...)` → `QTimer::singleShot(0ms, ...)` - `mpTimerReplay->setInterval(1000)` → `setInterval(1s)` - `mPendingTimer.start(60000)` → `start(1min)` - `QObject::startTimer(50)` → `startTimer(50ms)` - `QTest::qWait(100)` → `QTest::qWait(100ms)` - `QThread::msleep(10)` → `QThread::sleep(10ms)` This is a semantics-preserving refactor - every duration is kept exactly equal to before (e.g. `1000` ms becomes `1s`, `60000` ms becomes `1min`). No behavioural change. #### Motivation for adding to Mudlet Chrono literals make time durations self-documenting and type-safe. `1s` / `100ms` read unambiguously where a bare `1000` / `100` forces the reader to remember each API's unit, and the compiler now rejects unit mismatches. Only genuine duration arguments were converted - loop counts, scroll-line counts, sizes, ports and the like were deliberately left as plain integers. All targeted APIs provide `std::chrono` overloads in the minimum supported Qt (6.8.2): `QTimer::singleShot`/`start`/`setInterval` (5.8), `QObject::startTimer` (5.9), `QThread::sleep(std::chrono::nanoseconds)` (6.6) and `QTest::qWait(std::chrono::milliseconds)` (6.7). #### Other info (issues closed, discussion etc) Test case: the full application builds cleanly and the entire functional `ctest` suite passes. The only failing test is the known, pre-existing `PasswordMigrationTest` LSan exit-leak (GTK3/fontconfig noise), which is unrelated to this change. Assisted-by: Claude:claude-opus-4-8
2026-07-25 20:24:31 +02:00
QTimer::singleShot(0ms, qApp, [hostname, address, port]() {
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
mudlet::self()->startAutoLogin({});
infrastructure: use std::chrono literals for time durations (#9493) #### Brief overview of PR changes/additions Convert raw millisecond integer literals at time-duration call sites to `std::chrono` literals, and add `#include <chrono>` to each touched translation unit. Examples: - `QTimer::singleShot(0, ...)` → `QTimer::singleShot(0ms, ...)` - `mpTimerReplay->setInterval(1000)` → `setInterval(1s)` - `mPendingTimer.start(60000)` → `start(1min)` - `QObject::startTimer(50)` → `startTimer(50ms)` - `QTest::qWait(100)` → `QTest::qWait(100ms)` - `QThread::msleep(10)` → `QThread::sleep(10ms)` This is a semantics-preserving refactor - every duration is kept exactly equal to before (e.g. `1000` ms becomes `1s`, `60000` ms becomes `1min`). No behavioural change. #### Motivation for adding to Mudlet Chrono literals make time durations self-documenting and type-safe. `1s` / `100ms` read unambiguously where a bare `1000` / `100` forces the reader to remember each API's unit, and the compiler now rejects unit mismatches. Only genuine duration arguments were converted - loop counts, scroll-line counts, sizes, ports and the like were deliberately left as plain integers. All targeted APIs provide `std::chrono` overloads in the minimum supported Qt (6.8.2): `QTimer::singleShot`/`start`/`setInterval` (5.8), `QObject::startTimer` (5.9), `QThread::sleep(std::chrono::nanoseconds)` (6.6) and `QTest::qWait(std::chrono::milliseconds)` (6.7). #### Other info (issues closed, discussion etc) Test case: the full application builds cleanly and the entire functional `ctest` suite passes. The only failing test is the known, pre-existing `PasswordMigrationTest` LSan exit-leak (GTK3/fontconfig noise), which is unrelated to this change. Assisted-by: Claude:claude-opus-4-8
2026-07-25 20:24:31 +02:00
QTest::qWait(100ms);
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
QTest::mouseClick(mudlet::self()->mpConnectionDialog->new_profile_button, Qt::LeftButton);
infrastructure: use std::chrono literals for time durations (#9493) #### Brief overview of PR changes/additions Convert raw millisecond integer literals at time-duration call sites to `std::chrono` literals, and add `#include <chrono>` to each touched translation unit. Examples: - `QTimer::singleShot(0, ...)` → `QTimer::singleShot(0ms, ...)` - `mpTimerReplay->setInterval(1000)` → `setInterval(1s)` - `mPendingTimer.start(60000)` → `start(1min)` - `QObject::startTimer(50)` → `startTimer(50ms)` - `QTest::qWait(100)` → `QTest::qWait(100ms)` - `QThread::msleep(10)` → `QThread::sleep(10ms)` This is a semantics-preserving refactor - every duration is kept exactly equal to before (e.g. `1000` ms becomes `1s`, `60000` ms becomes `1min`). No behavioural change. #### Motivation for adding to Mudlet Chrono literals make time durations self-documenting and type-safe. `1s` / `100ms` read unambiguously where a bare `1000` / `100` forces the reader to remember each API's unit, and the compiler now rejects unit mismatches. Only genuine duration arguments were converted - loop counts, scroll-line counts, sizes, ports and the like were deliberately left as plain integers. All targeted APIs provide `std::chrono` overloads in the minimum supported Qt (6.8.2): `QTimer::singleShot`/`start`/`setInterval` (5.8), `QObject::startTimer` (5.9), `QThread::sleep(std::chrono::nanoseconds)` (6.6) and `QTest::qWait(std::chrono::milliseconds)` (6.7). #### Other info (issues closed, discussion etc) Test case: the full application builds cleanly and the entire functional `ctest` suite passes. The only failing test is the known, pre-existing `PasswordMigrationTest` LSan exit-leak (GTK3/fontconfig noise), which is unrelated to this change. Assisted-by: Claude:claude-opus-4-8
2026-07-25 20:24:31 +02:00
QTest::qWait(100ms);
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
QTest::keyClicks(QApplication::focusWidget(), hostname);
infrastructure: use std::chrono literals for time durations (#9493) #### Brief overview of PR changes/additions Convert raw millisecond integer literals at time-duration call sites to `std::chrono` literals, and add `#include <chrono>` to each touched translation unit. Examples: - `QTimer::singleShot(0, ...)` → `QTimer::singleShot(0ms, ...)` - `mpTimerReplay->setInterval(1000)` → `setInterval(1s)` - `mPendingTimer.start(60000)` → `start(1min)` - `QObject::startTimer(50)` → `startTimer(50ms)` - `QTest::qWait(100)` → `QTest::qWait(100ms)` - `QThread::msleep(10)` → `QThread::sleep(10ms)` This is a semantics-preserving refactor - every duration is kept exactly equal to before (e.g. `1000` ms becomes `1s`, `60000` ms becomes `1min`). No behavioural change. #### Motivation for adding to Mudlet Chrono literals make time durations self-documenting and type-safe. `1s` / `100ms` read unambiguously where a bare `1000` / `100` forces the reader to remember each API's unit, and the compiler now rejects unit mismatches. Only genuine duration arguments were converted - loop counts, scroll-line counts, sizes, ports and the like were deliberately left as plain integers. All targeted APIs provide `std::chrono` overloads in the minimum supported Qt (6.8.2): `QTimer::singleShot`/`start`/`setInterval` (5.8), `QObject::startTimer` (5.9), `QThread::sleep(std::chrono::nanoseconds)` (6.6) and `QTest::qWait(std::chrono::milliseconds)` (6.7). #### Other info (issues closed, discussion etc) Test case: the full application builds cleanly and the entire functional `ctest` suite passes. The only failing test is the known, pre-existing `PasswordMigrationTest` LSan exit-leak (GTK3/fontconfig noise), which is unrelated to this change. Assisted-by: Claude:claude-opus-4-8
2026-07-25 20:24:31 +02:00
QTest::qWait(100ms);
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
QTest::keyClick(QApplication::focusWidget(), Qt::Key_Tab);
infrastructure: use std::chrono literals for time durations (#9493) #### Brief overview of PR changes/additions Convert raw millisecond integer literals at time-duration call sites to `std::chrono` literals, and add `#include <chrono>` to each touched translation unit. Examples: - `QTimer::singleShot(0, ...)` → `QTimer::singleShot(0ms, ...)` - `mpTimerReplay->setInterval(1000)` → `setInterval(1s)` - `mPendingTimer.start(60000)` → `start(1min)` - `QObject::startTimer(50)` → `startTimer(50ms)` - `QTest::qWait(100)` → `QTest::qWait(100ms)` - `QThread::msleep(10)` → `QThread::sleep(10ms)` This is a semantics-preserving refactor - every duration is kept exactly equal to before (e.g. `1000` ms becomes `1s`, `60000` ms becomes `1min`). No behavioural change. #### Motivation for adding to Mudlet Chrono literals make time durations self-documenting and type-safe. `1s` / `100ms` read unambiguously where a bare `1000` / `100` forces the reader to remember each API's unit, and the compiler now rejects unit mismatches. Only genuine duration arguments were converted - loop counts, scroll-line counts, sizes, ports and the like were deliberately left as plain integers. All targeted APIs provide `std::chrono` overloads in the minimum supported Qt (6.8.2): `QTimer::singleShot`/`start`/`setInterval` (5.8), `QObject::startTimer` (5.9), `QThread::sleep(std::chrono::nanoseconds)` (6.6) and `QTest::qWait(std::chrono::milliseconds)` (6.7). #### Other info (issues closed, discussion etc) Test case: the full application builds cleanly and the entire functional `ctest` suite passes. The only failing test is the known, pre-existing `PasswordMigrationTest` LSan exit-leak (GTK3/fontconfig noise), which is unrelated to this change. Assisted-by: Claude:claude-opus-4-8
2026-07-25 20:24:31 +02:00
QTest::qWait(100ms);
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
QTest::keyClicks(QApplication::focusWidget(), address);
infrastructure: use std::chrono literals for time durations (#9493) #### Brief overview of PR changes/additions Convert raw millisecond integer literals at time-duration call sites to `std::chrono` literals, and add `#include <chrono>` to each touched translation unit. Examples: - `QTimer::singleShot(0, ...)` → `QTimer::singleShot(0ms, ...)` - `mpTimerReplay->setInterval(1000)` → `setInterval(1s)` - `mPendingTimer.start(60000)` → `start(1min)` - `QObject::startTimer(50)` → `startTimer(50ms)` - `QTest::qWait(100)` → `QTest::qWait(100ms)` - `QThread::msleep(10)` → `QThread::sleep(10ms)` This is a semantics-preserving refactor - every duration is kept exactly equal to before (e.g. `1000` ms becomes `1s`, `60000` ms becomes `1min`). No behavioural change. #### Motivation for adding to Mudlet Chrono literals make time durations self-documenting and type-safe. `1s` / `100ms` read unambiguously where a bare `1000` / `100` forces the reader to remember each API's unit, and the compiler now rejects unit mismatches. Only genuine duration arguments were converted - loop counts, scroll-line counts, sizes, ports and the like were deliberately left as plain integers. All targeted APIs provide `std::chrono` overloads in the minimum supported Qt (6.8.2): `QTimer::singleShot`/`start`/`setInterval` (5.8), `QObject::startTimer` (5.9), `QThread::sleep(std::chrono::nanoseconds)` (6.6) and `QTest::qWait(std::chrono::milliseconds)` (6.7). #### Other info (issues closed, discussion etc) Test case: the full application builds cleanly and the entire functional `ctest` suite passes. The only failing test is the known, pre-existing `PasswordMigrationTest` LSan exit-leak (GTK3/fontconfig noise), which is unrelated to this change. Assisted-by: Claude:claude-opus-4-8
2026-07-25 20:24:31 +02:00
QTest::qWait(100ms);
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
QTest::keyClick(QApplication::focusWidget(), Qt::Key_Tab);
infrastructure: use std::chrono literals for time durations (#9493) #### Brief overview of PR changes/additions Convert raw millisecond integer literals at time-duration call sites to `std::chrono` literals, and add `#include <chrono>` to each touched translation unit. Examples: - `QTimer::singleShot(0, ...)` → `QTimer::singleShot(0ms, ...)` - `mpTimerReplay->setInterval(1000)` → `setInterval(1s)` - `mPendingTimer.start(60000)` → `start(1min)` - `QObject::startTimer(50)` → `startTimer(50ms)` - `QTest::qWait(100)` → `QTest::qWait(100ms)` - `QThread::msleep(10)` → `QThread::sleep(10ms)` This is a semantics-preserving refactor - every duration is kept exactly equal to before (e.g. `1000` ms becomes `1s`, `60000` ms becomes `1min`). No behavioural change. #### Motivation for adding to Mudlet Chrono literals make time durations self-documenting and type-safe. `1s` / `100ms` read unambiguously where a bare `1000` / `100` forces the reader to remember each API's unit, and the compiler now rejects unit mismatches. Only genuine duration arguments were converted - loop counts, scroll-line counts, sizes, ports and the like were deliberately left as plain integers. All targeted APIs provide `std::chrono` overloads in the minimum supported Qt (6.8.2): `QTimer::singleShot`/`start`/`setInterval` (5.8), `QObject::startTimer` (5.9), `QThread::sleep(std::chrono::nanoseconds)` (6.6) and `QTest::qWait(std::chrono::milliseconds)` (6.7). #### Other info (issues closed, discussion etc) Test case: the full application builds cleanly and the entire functional `ctest` suite passes. The only failing test is the known, pre-existing `PasswordMigrationTest` LSan exit-leak (GTK3/fontconfig noise), which is unrelated to this change. Assisted-by: Claude:claude-opus-4-8
2026-07-25 20:24:31 +02:00
QTest::qWait(100ms);
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
QTest::keyClicks(QApplication::focusWidget(), port);
infrastructure: use std::chrono literals for time durations (#9493) #### Brief overview of PR changes/additions Convert raw millisecond integer literals at time-duration call sites to `std::chrono` literals, and add `#include <chrono>` to each touched translation unit. Examples: - `QTimer::singleShot(0, ...)` → `QTimer::singleShot(0ms, ...)` - `mpTimerReplay->setInterval(1000)` → `setInterval(1s)` - `mPendingTimer.start(60000)` → `start(1min)` - `QObject::startTimer(50)` → `startTimer(50ms)` - `QTest::qWait(100)` → `QTest::qWait(100ms)` - `QThread::msleep(10)` → `QThread::sleep(10ms)` This is a semantics-preserving refactor - every duration is kept exactly equal to before (e.g. `1000` ms becomes `1s`, `60000` ms becomes `1min`). No behavioural change. #### Motivation for adding to Mudlet Chrono literals make time durations self-documenting and type-safe. `1s` / `100ms` read unambiguously where a bare `1000` / `100` forces the reader to remember each API's unit, and the compiler now rejects unit mismatches. Only genuine duration arguments were converted - loop counts, scroll-line counts, sizes, ports and the like were deliberately left as plain integers. All targeted APIs provide `std::chrono` overloads in the minimum supported Qt (6.8.2): `QTimer::singleShot`/`start`/`setInterval` (5.8), `QObject::startTimer` (5.9), `QThread::sleep(std::chrono::nanoseconds)` (6.6) and `QTest::qWait(std::chrono::milliseconds)` (6.7). #### Other info (issues closed, discussion etc) Test case: the full application builds cleanly and the entire functional `ctest` suite passes. The only failing test is the known, pre-existing `PasswordMigrationTest` LSan exit-leak (GTK3/fontconfig noise), which is unrelated to this change. Assisted-by: Claude:claude-opus-4-8
2026-07-25 20:24:31 +02:00
QTest::qWait(100ms);
fix: clicking into the main window no longer steals copy from the command line (#9305) #### Brief overview of PR changes/additions - A plain left-click into the main console (e.g. to give it focus) no longer leaves a stray one-character selection behind. - Root cause: `TTextEdit::mouseMoveEvent` began a selection on *any* move while tracking, even when the pointer was still on the same character cell as the press — so the sub-pixel jitter of an ordinary click highlighted one character, and `mouseReleaseEvent` never cleared it. - Fix: a plain left-drag now only starts a selection once the pointer reaches a *different* character cell than the press point. Multi-click (word/line) and Ctrl selections are established in `mousePressEvent` and are untouched. - Adds a functional regression test (`MainConsoleSelectionTest`): a click with no drag leaves no selection, and a genuine drag still selects. #### Motivation for adding to Mudlet That stray one-character selection hijacked copying from the command line — `TCommandLine` prioritises any console selection over the input box on Ctrl+C, so <kbd>Ctrl</kbd>+<kbd>A</kbd> then <kbd>Ctrl</kbd>+<kbd>C</kbd> in the input would copy the single highlighted character instead of the typed command. #### Other info (issues closed, discussion etc) Fixes #3922 #### Test case 1. Type some text into the command-line input box. 2. Left-click once somewhere in the main window text area to focus it (don't drag). 3. <kbd>Ctrl</kbd>+<kbd>A</kbd> in the input box, then <kbd>Ctrl</kbd>+<kbd>C</kbd>. 4. Paste — you should get the full command-line text, **not** a single character from the main window. 5. Confirm a real drag-selection in the main window still works and still wins Ctrl+C as before. Automated: `ctest -R MainConsoleSelectionTest` — RED before the fix (fails at the no-drag assertion), GREEN after. `TOscTest` (console hyperlink-click handling) still passes, confirming no regression to click handling. --------- Signed-off-by: Ethan Hussong <ethan@ethanhussong.com>
2026-07-16 10:36:42 -05:00
QTest::keyClick(QApplication::focusWidget(), Qt::Key_Return);
});
QSignalSpy spy(mudlet::self(), &mudlet::signal_profileLoaded);
if (!spy.wait(5000)) {
QFAIL("Profile took too long to load.");
}
auto host = mudlet::self()->getActiveHost();
if (!host) {
QFAIL("No active host available for the test.");
}
QSignalSpy spy2(&(host->mTelnet), &cTelnet::signal_connected);
if (!spy2.wait(2000)) {
QFAIL("Could not connect with the host.");
}
}
bool waitForTextInBuffer(const QString& text, int timeoutMs = 5000)
{
auto console = mudlet::self()->getActiveHost()->mpConsole;
return QTest::qWaitFor(
[&]() {
for (int i = 0; i <= console->buffer.getLastLineNumber(); ++i) {
if (console->buffer.line(i) == text) {
return true;
}
}
return false;
},
timeoutMs);
}
void deleteProfileDirectory(const QString& profileName)
{
const QString path = mudlet::getMudletPath(enums::profileHomePath, profileName);
deleteDirectory(path);
}
void deleteDirectory(const QString& path)
{
QDir dir(path);
if (!dir.exists()) {
return;
}
dir.removeRecursively();
}
};
void initializeQRCResources()
{
#ifdef INCLUDE_VARIABLE_SPLASH_SCREEN
qInitResources_additional_splash_screens();
#endif
#ifdef INCLUDE_FONTS
qInitResources_mudlet_fonts_common();
#if defined(Q_OS_LINUX) || defined(Q_OS_FREEBSD)
qInitResources_mudlet_fonts_posix();
#endif
#endif
qInitResources_mudlet();
qInitResources_qm();
}
#include "MainConsoleSelectionTest.moc"
QTEST_MAIN(MainConsoleSelectionTest)