On first startup, use the splitter values suggested by Dirk:
Top/Bottom: 60/40
Info/Profile: 50/50
List/Map: 60/40
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The profile must be replotted when the dive mode changes.
Weirdly, this was routed via the dive-information tab
(making it inherently non-mobile compatible). Detect
such a change directly in the profile.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
On state change, the splitters were completely emptied and
refilled. Instead try to reuse already existing splitter
slots. This reduces annoying flickering.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The MainWindow has a function to replot the profile. Use that
instead of accessing the profile directly.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The memory management of the quadrant widgets is a total mess:
When setting the widget, the QSplitters take ownership, which
means that they will delete the widget in their destructor.
This is inherently incompatible with singletons, which must
not be deleted.
To avoid all these troubles, remove the widgets from the
QSplitters in the desctructor of the MainWindow. This of
course means that we now have to take care about deletion
of the widgets.
For local widgets use std::unique_ptr, for singletons use
a static variable that is deleted on application exit.
Sadly, for the map widget we can't use a normal singleton,
because the QML MapWidget's memory management is buggy.
Add a comment in the source code explaining this.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
When in planner mode, don't allow the user to change the application
state. This brought us nothing but troubles and inconsistencies.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
There was the "application state", which decided what to show
in the "quadrants" and the "view state" which decided which
quadrant to show. These interacted in a hard-to-grasp way.
The "view state" is used to show the map or dive list in
full screen.
I simply couldn't get these two orthogonal states to interact
properly. Moreover the thing was buggy: If a quadrant was hidden,
the user could still show it, by dragging from the side of the
window, at least under KDE.
To solve these woes, merge the two states into a single
application state. If the widget of a quadrant is set to null,
don't show it. So the four "view states" are now "application
states" where three of the four quadrants are not shown.
This also changes the memory management of the widgets:
widgets that are not shown are now removed from the QSplitter
objects. This makes it possible that the same widget is
shown in *different* quadrants.
While writing this, I stumbled upon a Qt bug, which is known
since 2014:
https://forum.qt.io/topic/43176/qsplitter-sizes-return-0
When restoring the quadrant sizes there was a test whether
the quadrant size is 0. If that was the case, a default size
was set. This seems not to work if the widgets were recently
added. Since this test now always fails, make the quadrants
non-collapsible and thus guarantee that 0 is never saved as
a size.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
When reparenting the statistics widget, QtQuick deletes
the rootNode and all the child nodes. It is unclear whether
this is a bug or intended behavior. In any case, it means
that the pointers to QSG nodes in the chart items become
stale.
To avoid this, delete all chart items in the root node's
destructor, before QtQuick can do anything. It is unclear
from which context this is called (render or UI) and whether
this is even valid. In some tests, it seemed to work.
The difficulty is that all the stale pointers to chart items
have to be deleted as well. All in all, the QSG memory
management is a big nuisance and very brittle.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Somehow three identical lines snuck into commit 0a4e37ee8b ("core/BT: simplify
detection of bluetooth names").
Reported-by: Henrik Brautaset Aronsen <subsurface@henrik.synth.no>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
For Uemis there was string-manipulation that leaked the
temporary string. Use QString instead in order not to have
to bother about such things.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Not that it matters, but there seems to be no reason to allocate
DiveImportedModel on the heap and no reason to leak it after
the download has finished.
Removes a artifactuous comment and fixes a typo.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Including <QObject> should only be necessary in very few
cases, because all widget classes have QObject as a base
class, which means that <QObject> already has to be included.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Instead of that super long if-else if chain, have something more
structured using a table for the common case of prefix based names.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
It would be so much nicer if we could just let libdivecomputer do this,
but the filter function there doesn't quite do things the way we need
them to be. Which is why we have our own function here.
This is a small attempt to rationalize the code that we have to make it
easier to maintain.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Merge with Jef's upstream libdivecomputer updates:
- support new Ratio iX3M 2021 model IDs
- support Mares Horizon, and fix the Mares Genius layout
- add support for Shearwood Sage
- various warning fixes, other minor details
Mark Aqualung i750TC as BLE capable
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Since we no longer show the noisy git updates to the user, it has become
harder for them to know whether a sync to the cloud was successful.
Since a manual sync will never show the new 'what did you change and
here's how you undo it' notification, it seems easy enough to simply
show a status update.
This adds a passive notification with no action button after the user
either uses the main menu or pulling down on the dive list in order to
trigger a manual sync.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The goal is to enable a user experiencing crashes when applying GPS data
to their dive log to make all necessary data available to the
developers. Hopefully the clipboard is large enough to hold all the
data.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The updatePaintNode() function, which is run on the render
thread detected a geometry change and initiated recalculation
of the chart layout.
This means that plotAreaChanged() was called in two different
thread contexts, which is questionable. Instead, hook into
the geometryChanged() function and recalculate the chart items
there.
This fixes a rendering bug, because the old code would first
delete unneeded items and then rerender the chart. Thus, old
grid and tick items were still visible.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
It simplifies reasoning about control flow a lot if it is known
that functions can't be invoked from a different part of the code
base.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This only read accesses the dive and constructs a plot-info
structure. Make the dive parameter const.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This function initializes decompression data from a dive.
The dive is not modified, therefore make it const.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
There was code to create a fake dc in the profile widget in
the case that there are no samples. To my understanding, this
is obsolete, as such fake data is now generated automatically
when adding dives.
If for some reason there really are no samples, quit early
and go into the empty state.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The way the starting time of a new plan was set was bonkers:
1) PlannerWidgets::planDive() invokes DivePlannerPointsModel::
createSimpleDive().
2) createSimpleDive() calls DivePlannerPointsModel::
setupStartTime()
3) setupStartTime() emits a signal startTimeChanged()
4) startTimeChanged is caught by PlannerWidget and sets
the UI field
5) change of the UI field emits a timeChanged() signal which
is connected to DivePlannerPointsModel::setStartTime()
6) setStartTime() sets the time of the plan and displayed_dive
and emits dataChanged()
7) dataChanged() replots the dive()
8) Back in DivePlannerPointsModel::createSimpleDive() the diveplan
start time is overwritten with displayed_dive (the value are
equal owing to 6)
Wow!
But it gets worse:
9) The initial dive plan is set up in createSimpleDive().
Since the profile is drawn in 7) after clearing the displayed_dive
and before constructing the initial plan, the profile is shown
on a dive without samples. It therefore generates a dummy profile.
To make this somewhat less insane, remove the startTimeChanged()
signal in 3), explicitly set the start time of plan and dive to
the one calculated by setupStartTime() and explicitly set the UI
filed in the plannerWidget.
This still indirectly draws the profile via signals in a convoluted
way, but at it straightens out things somewhat. Most importantly,
the profile doesn't have to generate a fake DC.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
While this didn't appear to be needed when dragging the legend with a
mouse, on a touch screen for some reason the drag ended after 30 pixels
either way horizontally (but no apparent limit vertically). By setting
this flag to true, drags on a tablet appear to work as expected.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The labels in bar an pie charts are realized as individual
QSG pixmap nodes with an alpha channel. Sadly, rendering
bright labels onto a transparent background gives very
ugly artifacts.
As a stop gap measure, until the problem is understood,
render on a background with the color of the pie slice
or bar.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Most colors were already collected there, but a few were dispersed
throughout the source files.
For future themeability, move the remaining colors to this common
place.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This is not perfect - the polygon of the confidence area is
calculated even if it is not shown. Oh well.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Up to now, when the user changed the visibility of chart features
(legend, quartiles, labels, etc.) the whole chart was replot.
Instead, only change the visibility status of these items.
After all, this modularity is one of the things the conversion
to QSG was all about.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The scatter plot items shared their textures. These were
std::unique_ptrs and cleaned up on exit. Owing to QSG's
broken memory model, freeing the textures after QApplication
terminated its threads led to crashes. Therefore, leak the
textures. Not satisfying, but ultimately harmless and better
than a crash.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Fix a bug that was fixed in b5c8d0dbb4 and reintroduced in
e7907c494f. Here is the original commit message:
The range for a one-bin chart is [-0.5,0.5], thus the range
in an n-bin chart is [-0.5,n-0.5], not [-0.5,n+0.5].
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The code was wrong, because it deleted the ChartItems in the
main UI thread, not the render thread. This would delete the
QSG nodes in the UI thread and then crash on mobile.
Therefore refactor this part of the code by adding the
items to be deleted to a list that will be deleted by the
render thread.
As a drop in replacement of std::unique_ptr, implement
a silly ChartItemPtr class, which auto-initializes to null.
This turns the deterministic and easily controlled memory
management into a steaming pile of insanity. Obviously,
this can be made much more elegant, but this has to do for now.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
These values were used for items on the QGraphicsScene and
have been replaced by integer values used on the QSG scene.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This one is trivial, since everything is there already:
Replace the QGraphicsSimpleTextItem with a ChartTextItem.
Only few functions have to be renamed.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
All series are converted to QSG. Thus, the pointer to the
QGraphicsView can be removed from the common base class.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Since there are no disk-segment QSG primitives (one could draw
a triangle fan, but that doesn't seem optimal), this draws
into a pixmap and blits that as a QSG node.
Since this is the only series without axis, it needs a function
that returns the size of the plot area. This didn't exist, so
add it.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The original plan to reuse the ChartPixmapItem for the
scatteritems was dumped, because it is unclear if the
textures are shared if generated for each item.
Instead, a new ChartScatterItem was created, where all
items share the same textures (one for highlighted,
one for non-highlighted). This means that the rendering
of the scatter items is now done in the chartitem.cpp
file, which feels like a layering violation. Not good,
but the easiest for now.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This is lazy: Derive from the bar chart item and add whiskers
in the subclassed render() function. The code is ugly, because
the base class function clears the dirty flags and therefore
the derived class has to remember them. Oh well.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
To replace the QGraphicsScene, we need the possibility of
showing and hiding items.
Turns out, the QSG API is completely insane.
Whether an item should be shown is queried by the virtual
function isSubtreeBlocked(), which is supposed to be
overriden by the derived classes.
However, the common nodes for rectangles and pixmaps are
supposed to be created by QQuickWindow, for hardware
optimization. This gives nodes that cannot be derived
from and therefore whether the item is shown or not cannot
be controlled.
There are therefore two distinct cases to consider: The
node is allocated by the code directly or indirectly by
QQuickWindow.
In the latter case, we use a proxy node with the only
purpose of having a "visible" flag and add the obtained
node as a child.
This madness is performed with template trickery to get
unified code.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
To this end, two new ChartItems were added: A "bar" (a rectangle
with a border) and a "text" (multiple lines of text).
It turns out that the text on the bars now looks atrocious.
The reason appears to be that the antialiasing of the font-rendering
does not blend into the alpha channel, but into a supposed
background color? This will have to be investigated.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Render the confidence area and the regression line into a pixmap
and show that using a QSGNode.
It is unclear whether it is preferred to do it this way or to
triangulate the confidence area into triangles to be drawn by
the shader.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
