Control Spawning of Plot Windows

[Owen Arnold]

The current MantidPlot behaviour is that every time you right-click and Plot (slice, spectrum) you get a new instance of a plot window. This can easily get out of hand. Some users want to re-use the same plot instance rather than generate new windows each time.

We should have a one-time configuration option that allows users to have a 1:1 mapping between the various plot types and their MantidPlot instance. The option should be off by default to preserve existing behaviour.

This is a requirement request from Jon Taylor, from the MantidPlot/Vates meeting 07/Oct/2013.

[Nick Draper]

Bulk move to assigned at the introduction of the triage step

[Federico Montesino Pouzols]

Added config option: ReusePlotInstances (On/Off), re #8091

Changeset: 5f500a75aca8d6d63ccd0e69bcdb7ad583e18cc6

[Federico Montesino Pouzols]

Support option 'ReusePlotInstances' in MantidPlot UI, re #8091

Changeset: 84ba0ed77aeca753277f0ef5d5a5a5487a1fe1a2

[Federico Montesino Pouzols]

Added individual close slice viewer win, re #8091

Changeset: aba0f88fe8a6337cf572b1344bf10fbd4aeababf

[Federico Montesino Pouzols]

tooltip, and try to keep positions, re #8091

Changeset: 9be889d5621dc7a830c2cfbfc7e8416a2216eb0c

[Federico Montesino Pouzols]

Nicer error handling/messaging - spectrum and slice viewers, re #8091

Changeset: 56d66937308d58ef2de2dbdb06ecd35e53473fce

[Federico Montesino Pouzols]

Nicer error handling/messaging - color fill plot, re #8091

Changeset: 4a6f4269bdbab9512e131ad3ec960b06a9d12dcb

[Federico Montesino Pouzols]

comment out, not really possible, cause of plot differences, re #8091

Changeset: 0457dc3b1c3054216636803ebcb889041e7f1414

[Federico Montesino Pouzols]

fixed initialization, re #8091

Changeset: af5c9066eda258436c816ed8751ae816e8da1348

[Federico Montesino Pouzols]

fixed window name issues on windoze, re #8091

Changeset: 8ffc5e7c668412041ae2604327209c7a4885efd1

[Federico M Pouzols]

A new plotting mode has been added that prevents the traditional (annoying) spawning of new windows for every new plot. This new plotting mode is disabled by default, as suggested. The checkbox to enable it can be found in:

View->Preferences->Mantid->Options (tab): "Re-use plot instances for different types of plots"

(http://www.mantidproject.org/MantidPlot:_Preferences_Dialog_Mantid#Options)

The new option has been added in the ConfigService as: "MantidOptions.ReusePlotInstances" = On/Off

When you enable this option, there are now 5 (reused) instances of different types of plots/interfaces that can be shown from the workspaces (dock) window, by right-clicking on workspaces:

• spectrum plot (with/without errors, with/without distribution, etc.)
• color fill plot
• spectrum viewer
• instrument view
• slice viewer

There can only be one spectrum plot at a time, and one color fill plot, etc.

I updated and expanded a bit the description of the general View->Preferences: http://www.mantidproject.org/MantidPlot:_The_View_Menu#View_-.3E_Preferences . And then updated the screenshot of the Preferences Dialog / Options: http://www.mantidproject.org/MantidPlot:_Preferences_Dialog_Mantid#Options

It would be great if Jon Taylor and/or others who were interested in this change could check and provide feedback, as there would certainly be other alternative, plausible ways to address this ticket.

To test:

• check that you can enable/disable the new option (If you hover the mouse you should see a tool tip explanation), and that new windows are created or not depending on your choice.

• With the option disabled you should see the traditional MantidPlot behavior. When you enable it, if for example you plot several color fill plots from different workspaces, all these plots should be shown on a single window.

• check that you can switch between the new "controled" mode and the old "spawning" mode by using the tickbox in preferences / options, at any time.

---

Note the following imperfection:

• The implementation seems sound for the spectrum plots and the color fill plots: new plots re-use the same window.

• The implementation for spectrum viewer, instrument view, and slice viewer is deficient: if there's a previous windows it is closed and destroyed, and a new window is created and shown. This is because a "refresh" or "update underlying data / workspace" functionality is not provided. "Refresh window" methods are required for these three classes that would be able to take a new workspace and show the relevant information without requiring a destroy/new cycle. That functionality is currently entangled in the constructors and other methods. I created three (lowish priority) separate tickets for these, as suggested by Owen. See #10445, #10446, and #10447.

[Dan Nixon]

After having plotted two workspaces with the "Re-use plot instances for different types of plots" option disabled I closed the two plot windows changed the option in Preferences and attempted to plot one of the workspaces again, this gave the following segfault:

Program received signal SIGSEGV, Segmentation fault.
0x00000000006ef4af in QListData::size (this=0x1bc83b0) at /usr/include/qt4/QtCore/qlist.h:98
98	    inline int size() const { return d->end - d->begin; }
(gdb) bt
#0  0x00000000006ef4af in QListData::size (this=0x1bc83b0) at /usr/include/qt4/QtCore/qlist.h:98
#1  0x00000000006f4d74 in QList<Graph*>::size (this=0x1bc83b0) at /usr/include/qt4/QtCore/qlist.h:137
#2  0x0000000000857225 in MultiLayer::setLayersNumber (this=0x1bc8300, n=0) at /home/dan/testing-builds/test1/Code/Mantid/MantidPlot/src/MultiLayer.cpp:1237
#3  0x0000000000647de1 in ApplicationWindow::prepareMultiLayer (this=0x1996000, isNew=@0x7fffffffd1e0: false, window=0x1bc8300, newWindowName=..., clearWindow=true) at /home/dan/testing-builds/test1/Code/Mantid/MantidPlot/src/ApplicationWindow.cpp:2796
#4  0x00000000009d4c9a in MantidUI::plot1D (this=0x1abbb80, toPlot=..., spectrumPlot=true, errs=false, distr=false, style=Graph::Unspecified, plotWindow=0x1bc8300, clearWindow=true) at /home/dan/testing-builds/test1/Code/Mantid/MantidPlot/src/Mantid/MantidUI.cpp:3063
#5  0x00000000009d4843 in MantidUI::plot1D (this=0x1abbb80, toPlot=..., spectrumPlot=true, errs=false, distr=false, plotWindow=0x0, clearWindow=false) at /home/dan/testing-builds/test1/Code/Mantid/MantidPlot/src/Mantid/MantidUI.cpp:2994
#6  0x00000000009a0c14 in MantidDockWidget::plotSpectra (this=0x1cb9a40) at /home/dan/testing-builds/test1/Code/Mantid/MantidPlot/src/Mantid/MantidDock.cpp:1263
#7  0x0000000000ad4f07 in MantidDockWidget::qt_static_metacall (_o=0x1cb9a40, _c=QMetaObject::InvokeMetaMethod, _id=15, _a=0x7fffffffd4f0) at /home/dan/testing-builds/test1-build/MantidPlot/src/Mantid/moc_MantidDock.cxx:108
#8  0x00007ffff33fe87a in QMetaObject::activate(QObject*, QMetaObject const*, int, void**) () from /usr/lib/x86_64-linux-gnu/libQtCore.so.4
#9  0x00007ffff3b5aa62 in QAction::triggered(bool) () from /usr/lib/x86_64-linux-gnu/libQtGui.so.4
#10 0x00007ffff3b5c433 in QAction::activate(QAction::ActionEvent) () from /usr/lib/x86_64-linux-gnu/libQtGui.so.4
#11 0x00007ffff3f8a489 in ?? () from /usr/lib/x86_64-linux-gnu/libQtGui.so.4
#12 0x00007ffff3f8e9b9 in ?? () from /usr/lib/x86_64-linux-gnu/libQtGui.so.4
#13 0x00007ffff3bb050a in QWidget::event(QEvent*) () from /usr/lib/x86_64-linux-gnu/libQtGui.so.4
#14 0x00007ffff3f927bb in QMenu::event(QEvent*) () from /usr/lib/x86_64-linux-gnu/libQtGui.so.4
#15 0x00007ffff3b60e2c in QApplicationPrivate::notify_helper(QObject*, QEvent*) () from /usr/lib/x86_64-linux-gnu/libQtGui.so.4
#16 0x00007ffff3b675dd in QApplication::notify(QObject*, QEvent*) () from /usr/lib/x86_64-linux-gnu/libQtGui.so.4
#17 0x0000000000996476 in MantidApplication::notify (this=0x7fffffffe460, receiver=0x5935b90, event=0x7fffffffdc10) at /home/dan/testing-builds/test1/Code/Mantid/MantidPlot/src/Mantid/MantidApplication.cpp:29
#18 0x00007ffff33ea4dd in QCoreApplication::notifyInternal(QObject*, QEvent*) () from /usr/lib/x86_64-linux-gnu/libQtCore.so.4
#19 0x00007ffff3b66d93 in QApplicationPrivate::sendMouseEvent(QWidget*, QMouseEvent*, QWidget*, QWidget*, QWidget**, QPointer<QWidget>&, bool) () from /usr/lib/x86_64-linux-gnu/libQtGui.so.4
#20 0x00007ffff3bdbcfc in ?? () from /usr/lib/x86_64-linux-gnu/libQtGui.so.4
#21 0x00007ffff3bdb269 in QApplication::x11ProcessEvent(_XEvent*) () from /usr/lib/x86_64-linux-gnu/libQtGui.so.4
#22 0x00007ffff3c02b02 in ?? () from /usr/lib/x86_64-linux-gnu/libQtGui.so.4
#23 0x00007fffe90b2e04 in g_main_context_dispatch () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#24 0x00007fffe90b3048 in ?? () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#25 0x00007fffe90b30ec in g_main_context_iteration () from /lib/x86_64-linux-gnu/libglib-2.0.so.0
#26 0x00007ffff34177be in QEventDispatcherGlib::processEvents(QFlags<QEventLoop::ProcessEventsFlag>) () from /usr/lib/x86_64-linux-gnu/libQtCore.so.4
#27 0x00007ffff3c02bb6 in ?? () from /usr/lib/x86_64-linux-gnu/libQtGui.so.4
#28 0x00007ffff33e90af in QEventLoop::processEvents(QFlags<QEventLoop::ProcessEventsFlag>) () from /usr/lib/x86_64-linux-gnu/libQtCore.so.4
#29 0x00007ffff33e93a5 in QEventLoop::exec(QFlags<QEventLoop::ProcessEventsFlag>) () from /usr/lib/x86_64-linux-gnu/libQtCore.so.4
#30 0x00007ffff33eeb79 in QCoreApplication::exec() () from /usr/lib/x86_64-linux-gnu/libQtCore.so.4
#31 0x0000000000ae4c0c in main (argc=1, argv=0x7fffffffe5f8) at /home/dan/testing-builds/test1/Code/Mantid/MantidPlot/src/main.cpp:219
(gdb) p d
$1 = (QListData::Data *) 0x0

[Federico Montesino Pouzols]

Fix crash when users close plot windows, re #8091

Changeset: c8877c94de15485178ac296f9306d7318ccb2d45

[Federico M Pouzols]

Good catch! This is now properly handled using (guarded) QPointer.

[Dan Nixon]

Merge remote-tracking branch 'origin/feature/8091_control_spawning_of_plot_windows'

Full changeset: ​4b0e772141f60ca801c0c97c6b0d22fa2b28b8ee

[Dan Nixon]

I quite like this, however one thing that would be nice is to be able to enable and disable the controlled/spawning mode on a per feature basis.

For instance I would quite like to only ever have one spectrum viewer instance but also have "unlimited" spectra plot instances. But this is something that could easily be implemented later.

[Federico M Pouzols]

I like this suggestion a lot. I'd add a ticket about it. Let's check with Owen and also see if we get feedback from users.

[Stuart Campbell]

This ticket has been transferred to github issue ​8936