Changes

Some minor cleaning up.

Only build OAuth (#1627).

Building our own copy of ICU on Linux (#1627).

Build our own copy of ICU on Linux (#1627).

Build our own copy of ICU on Linux (#1627).

Build our own copy of ICU on Linux (#1627).

Build our own copy of ICU on Linux (#1627).

Build our own copy of ICU on Linux (#1627).

Built our own copy of ICU on Linux (#1627).

Built our own copy of ICU on Linux (#1627).

Don't build ICU anymore (#1627).

ICU package: make sure that we can retrieve our list of ICU libraries
(#1627).

ICU package: make sure that we can retrieve our PKGCONFIG directory
(#1627).

Make sure that ICU (and Mesa while we are at it) is available before we
build ICU-dependent libraries (e.g. o2). Also deploy our copy of ICU on
Linux (#1627).

CMake: execute the packaging related code only if really needed.

CMake: execute the packaging related code only if really needed.

CMake: execute the packaging related code only if really needed.

Reverting most of d02504a, b94d322 and 7f5b6c3.

Core plugin: some minor improvement to the reloading of a file.

Some minor cleaning up.

Some minor cleaning up.

Main window: slight improvement to the way we use the Core plugin.

Now track our Simulation Support plugin (#1630).

Simulation Support plugin: removed our handling of the File Handling
interface (#1630).

Central widget: get our Simulation Support plugin to save/rename/reload
a file before our file manager (#1630).

Simulation Support plugin: removed our handling of the File Handling
interface (#1630).

Simulation support: simplified our Simulation interface and put back our
handling of the File Handling interface (#1630).
Indeed, the only thing we need to do (and should be doing) as soon as 
possible is stop the simulation. Once this is done, we can retrieve our 
file details when a SED-ML/CellML file has been reloaded. To retrieve 
our file details just after stopping a simulation will result in our 
file details being wrong since a SED-ML/CellML file will have reloaded 
yet.

Simulation support: make sure that we have a data store before returning
our number of runs (#1630).
This is needed, for example, when a simulation was running and then we 
reloaded a SED-ML file as a result of its corresponding CellML file 
having been modified.

Simulation support: we should really not have to reset our various
arrays in SimulationData::deleteArrays().
Indeed, we should only ever come here once, so…

Simulation Experiment view: a simulation shouldn't be stopped by the
Simulation Experiment view (#1630).
When a simulation needs to be stopped, it should be stopped before doing
anything else (e.g. reloading a file) since it’s run in its own thread.
So, anything else will result in a potential crash of OpenCOR. This is
the reason we now stop a simulation from within our central widget.

Simulation support: properly handle the generation of elapsed time
(#1630).

Simulation Experiment view: only update our progress bar or our tab icon
if needed (#1630).

Graph Panel widget: improved the alignment of a graph panel with its
neighbours.
Indeed, when running the model in issue #1630, the X axis of the first 
graph panel was flashing for a bit before becoming “stable”. This was 
due to us calling xScaleWidget->setMinBorderDist(0, 0), which unlike our
call to yScaleDraw->setMinimumExtent(0.0), we don’t seem to need to 
properly align our graph panels.

Revert "Graph Panel widget: improved the alignment of a graph panel with
its neighbours."
This reverts commit 9920ff9a07c903c8900154eb68ca1eb9702bee5d.

Graph Panel widget: improved the alignment of a graph panel with its
neighbours on macOS.
Indeed, when running the model in issue #1630 on macOS, the X axis of 
the first graph panel was flashing for a bit before becoming "stable". 
This was due to us calling xScaleWidget->setMinBorderDist(0, 0), which 
unlike our call to yScaleDraw->setMinimumExtent(0.0), we don’t seem to 
need to properly align our graph panels (at least, on macOS).

Revert "Graph Panel widget: improved the alignment of a graph panel with
its neighbours on macOS."
This reverts commit c927305cc668a4b5aeb1c05f5ae127db59f2abf6.

Graph Panel widget: improved the alignment of a graph panel with its
neighbours.
Indeed, when running the model in issue #1630, the X axis of the first 
graph panel was flashing on macOS for a bit before becoming "stable", 
due to our call to xScaleWidget->setMinBorderDist(0, 0). So, we had to 
find another way of doing things without that call...

Simulation Experiment view: don’t consider a SED-ML file as modified
under certain conditions (#1626).

macOS: added support for exiting full-screen mode when pressing the Esc
key (#1637).

Slight update to our downloads page (#1637).

Some minor cleaning up.

Some minor cleaning up.

CellML support: have `CellmlFile::runtime()` always return a brand new
runtime (#1630).
The idea is that whoever calls `CellmlFile::runtime()` is responsible 
for deleting that runtime once it’s done with it.

Removed our Simulation interface since we don't need it anymore (thanks
to commit bd5ab3f; #1630).

Removed our Simulation interface since we don't need it anymore (thanks
to commit bd5ab3f; #1630).

Some minor cleaning up.

Removed our Simulation interface since we don't need it anymore (thanks
to commit bd5ab3f; #1630).

CellML support: have `CellmlFile::runtime()` always return a brand new
runtime (#1630).
The idea is that whoever calls `CellmlFile::runtime()` is responsible 
for deleting that runtime once it’s done with it.

Simulation Experiment view: make sure that a simulation is stopped when
a saved file is to be considered as reloaded (#1630).

Simulation Experiment view: prevent the third crash in issue #1630.

Simulation support: actually wait for our worker thread to be done when
asking our simulation to stop (#1630).

Some minor cleaning up.

Simulation Experiment view: prevent the fourth crash in issue #1630.

Some minor cleaning up.

Some minor cleaning up.

Some minor cleaning up.

Simulation Experiment view: prevent the second crash in issue #1630.

Some minor cleaning up.

Some cleaning up.
In the end, because of what we did for issue #1630, we don’t need to 
decide whether we want an elapsed time to be computed or not when asking
for a simulation to stop.

Some minor cleaning up.

Some minor cleaning up.

Some minor cleaning up.

Some minor cleaning up.

Some minor cleaning up.

For testing Jenkins.

For testing Jenkins.

For testing Jenkins.

For testing Jenkins.

Only build QtWebKit (#1627).

QtWebKit library: prevent a warning from CMake (#1627).

QtWebKit library: prevent a warning from CMake.

Using the latest commit of QtWebKit (#1627).

Some minor cleaning up.
Not sure why GIT_TAG was defined on Windows and, indeed, CMake told us 
that this manually-specified variable was not used by the project.

Revert "Some minor cleaning up."
This reverts commit b2fa258764dc2cb90f0e142cdc957b669fbe7362.
Argh, we actually need it for our OpenCOR-specific version of QtWebKit. 
It’s just that I made a mistake when creating the
`v5.212.0-alpha2-72cfbd7` release by using the `5.212` target rather 
than our `opencor` one.

Linux: make sure that we build/retrieve ICU before building/retrieving
QtWebKit (#1627).

Some minor cleaning up.

QtWebKit package: use our copy of ICU on Linux (#1627).

QtWebKit: new Windows and macOS binaries (#1627).

Some minor cleaning up.

QtWebKit: new Linux binaries (#1627).

Stop building QtWebKit only (#1627).

QtWebKit: new Linux binaries (#1627).

QtWebKit package: fixed a small problem with the CMakeLists.txt file
(#1627).

Some minor cleaning up.

Make sure that ICU is handled before QtWebKit (#1627).

Revert part of commit 6640946.

Some minor cleaning up.

ICU package: make sure that we can deploy it even when building it.

ICU package: make sure that we can deploy it even when building it.

ICU package: make sure that we can deploy it even when building it.

Some minor cleaning up.

Don't only build ICU and QtWebKit anymore (#1627).

Some minor cleaning up.

Deploy JPEG and PNG on Linux (#1627).

Slight update to our downloads page (#1627).

Travis CI: now need JPEG and PNG (#1627).

Slight updates to the Oxygen library.

Renamed ULLONG_WIDTH to AttributeNumberWidth since it's defined by gcc
on Ubuntu 18.04 LTS.

Prevent a deprecated warnings for the CellML API on Ubuntu 18.04 LTS.

Make sure that our plugins use RPATH rather than RUNPATH on Linux.

Make sure that OpenCOR itself uses RPATH rather than RUNPATH on Linux.

Properly deploy some system libraries on Linux.

Only build the SUNDIALS package (#1643).

SUNDIALS package: upgraded to version 3.1.1 (#1643).

SUNDIALS package: upgraded to version 3.1.1 (#1643).

SUNDIALS package: new binaries (#1643).

Slight update to our downloads page (#1643).

Don't build the SUNDIALS package only (#1643).

QtWebKit package: updated our macOS binary.

QtWebKit package: slight update to our README.rst file.

Only build the QtWebKit package.

QtWebKit package: slight update to our README.rst file.

Some minor cleaning up.

QtWebKit package: slight update to our README.rst file.

QtWebKit package: slight update to our README.rst file.

QtWebKit package: statically link libjpeg.

QtWebKit package: statically link libpng.

QtWebKit package: statically link ZLIB.

QtWebKit package: slight update to our README.rst file.

QtWebKit package: have it depend on our copy of JPEG, PNG and ZLIB on
Linux and macOS.

Some minor cleaning up.

Some minor cleaning up.

QtWebKit package: use -fPIC for JPEG, PNG and ZLIB (on Linux and macOS).

Some minor cleaning up.

Some minor cleaning up.

QtWebKit package: new Linux and macOS binaries.

QtWebKit package: new Windows binaries.

Stop building only the QtWebKit package.

Some minor cleaning up.

Some minor cleaning up.

Prevent OpenCOR from crashing when selecting Tools | Reset All.
It looks like this might (somehow?) be related to our use of the new 
signal/slot mechanism? At least, that seems to be the only difference 
between the 17 March and 23 April snapshots when it comes to 
MainWindow::updateDockWidgetsVisibility().

CellML Text view: keep track of a CellML file’s documentation (#1642).

Slight update to our downloads page (#1642).

Slight update to our downloads page (#1642).

Only build LLVM+Clang (#1552).

LLVMClang package: upgraded to version 6.0.0 (#1552).

LLVMClang package: upgraded to version 6.0.0 (#1552).

LLVMClang package: upgraded to version 6.0.0 (#1552).

LLVMClang package: upgraded to version 6.0.0 (#1552).

LLVMClang package: new Linux and macOS binaries (#1552).

Stop building on LLVM+Clang (#1552).

Some minor cleaning up.

Simulation Experiment view: make sure that mGraphPanelsWidgetSizes gets
properly initialised upon export a SED-ML file / COMBINE archive.

Slight update to our downloads page (#1552).

LLVMClang package: new Windows binaries (#1552).

Travis CI: upgraded to Qt 5.9.5.

Travis CI: some tests to understand why things are now failing...!?

Travis CI: some tests to understand why things are now failing...!?

Travis CI: some tests to understand why things are now failing...!?

Travis CI: some tests to understand why things are now failing...!?

Travis CI: some tests to understand why things are now failing...!?

Travis CI: optimising the packages we need on Ubuntu.

Travis CI: some tests to understand why things are now failing...!?

Travis CI: optimising the packages we need on Ubuntu.

LLVMClang package: upgraded to version 6.0.0 (#1552).

Built a special version of LLVM+Clang to see whether this is the cause
of our problems on Travis CI.

LLVMClang package: upgraded to version 6.0.0 (#1552).

Some minor cleaning up.

Stopped using Travis CI.
Indeed, we now have everything set up and working nicely on Jenkins. Not
only that, but issue #1552 is building/testing fine locally on Jenkins,
but not on Travis CI (though our Docker Ubuntu 16.04 LTS) and I am
really not sure why. I spent a full day trying to figure out, but to no
avail. So, since Jenkins is working (and also build/test on Windows and
macOS), we might as well only use that, not to mention that we can rely
ccache on Jenkins, thus making our builds faster.

Windows: prevent a warning when using LLVM+Clang.

Text-based editors: can now word-wrap an editor's contents (#1652).

Some minor cleaning up.

Editing view: updated our French translations (#1652).

Text-based editors: remember about word-wrapping (#1652).

Slight update to our downloads page (#1652).

Slight update to our what's new page (#1652).

Slight improvement to our MathML tests.

Prevent a crash when leaving OpenCOR with a simulation still running.

macOS: properly handle the exiting of full-screen mode when pressing the
Esc key (#1637).

Some minor cleaning up.

Some minor cleaning up.

Text-based editors: now highlight all occurrences of a word throughout a
whole file (#1414).

Editor widget: customised our highlighting a bit (#1414).

Editor widget: make sure that our highlighting is cleared whenever
necessary (#1414).

Editor widget: make sure that our position is still the same after some
highlighting (#1414).

Some minor cleaning up (#1414).

Editor widget: make sure that our scrollbars are back to their original
position upon highlighting/replacing all (#1414).

QScintilla support: show our cursor position on our vertical scrollbar
(#1414).

QScintilla widget: added support for the highlighting of lines on the
vertical scrollbar (#1414).

QScintilla editor: simplified our highlighting in the vertical scrollbar
(#1414).
Indeed, it used to be a bit too much in your face, not least because we 
can’t draw the slider “above” our highlights (as Chrome does).

Slight updates to our downloads and what is new pages (#1414).

QScintilla widget: take into account that on Windows and Linux, a
vertical scroll bar has both a up and down arrow button (#1414).

Moved our highlighting support code from QScintillaWidget to
EditorWidget (#1414).

Editor widget: make sure that our scroll bar highlighting is properly
shifted on Windows and Linux (#1414).

Some minor cleaning up (#1414).

Some minor cleaning up.

Some minor cleaning up.

Some minor cleaning up.

Some cleaning up (#1414).

Some minor cleaning up.

Editor widget: make sure that showing our find/replace widget not only
highlight all the occurrences of the current word, but also selects it
(#1414).

Some minor cleaning up.

Some minor cleaning up.

Editor widget: update the highlighted lines in our vertical scroll bar
when adding text (#1414).

Various work on the Editor widget (#1414).
- Put the code for our editor widget’s scroll bar in its own file.
- Added a new file for our editor widget itself
(EditorWidgetEditorWidget).
- Now properly handle the highlighting of all occurrences of a given 
word, i.e. the highlighting on the scroll bar is updated whenever 
adding/removing some text, deleting a selection, etc.

Central widget: removed the need for mRemoteLocalFileNames.

Make sure that we hide our busy widget if we couldn't open a remote
file.

Some minor cleaning up (#1414).

Some minor cleaning up (#1414).

Some minor cleaning up (#1414).

CellML Text editing: handle un/commenting without relying on the
handling of a signal (#1414).

Don't associate an icon with a data store variable anymore (#1657).

Some minor cleaning up.

Editor widget: highlight all occurrences of a word that gets selected by
double clicking on it (#1414).

Editor widget: optimised the highlighting/finding of some text (#1414).

Editor widget: highlight all occurrences of a word that gets selected by
double clicking on it (#1414).

Editor widget: can now find a multiline piece of text (#1414).

Only initialise parameter type to QIcon map when it is first used
(#1657).
If initialisation is done at plugin load time then a CLI application
will crash as QIcon is a GUI class.

Some minor cleaning up (#1657).

Some minor cleaning up.

Removed unnecessary const  and const references.

Removed unnecessary const  and const references.

Editor widget: make sure that we properly handle the use of TAB/BACKTAB.
Also improved our handling of the Esc key.

Editor widget: don't highlight a space-based string.

Central widget: prevent a GUI glitch when changing modes.

Some minor cleaning up.

Revert "Central widget: prevent a GUI glitch when changing modes."
This reverts commit 771708d80c0d78d4765319a2fc0f6d57ef0a69d3.
Indeed, there is definitely no need for this. It’s just that there is 
something between the current codebase and that of 23 April 2018 (i.e. 
our latest snapshot) that is different and that results in the issue 
this commit was trying to address. In fact, there is another issue, 
which is that the Simulation Experiment view flashes a bit when 
switching to it and there are several graph panels, which results in the
N62 SED-ML file being considered modified while it clearly shouldn’t be
(and isn’t using the 2018-04-23 snapshot). So, we need to figure out
what we did that causes those issues…

Simulation Experiment view: now provide a default value for the legend
title of a valid graph (1662).

Slight update to our downloads page (#1662).

Graph Panel widget: prevent a GUI glitch on Windows and Linux.

Some minor cleaning up.

Simulation Experiment view: make sure that mGraphPanelsWidgetSizes gets
properly initialised.

Some minor cleaning up.

Some minor cleaning up.

Prevent a warning on Linux.

Prevent a couple of warnings on Linux.

Some minor cleaning up.

Prevent a warning on Linux.

Some minor cleaning up.

Simulation Experiment view: now automatically assign a colour to a new
graph (#1665).

Graph Panel widget: enable our legend by default (#1665).