The undo-code uses owning pointers based on std::unique_ptr to
manage lifetime of C-objects. Since these are generally useful,
move them from the undo-code to the core-code. In fact, this
eliminates one instance of code duplication.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
displayed_dive used to contain the currently displayed (as in
shown on the profile) dive. However, now it is only a "scratch"
dive used by the planner and initialized every time the planner
is started. There is no point in clearing this dive when clearing
the dive data. In fact, the dive should probably be cleared when
the planner finishes.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The current dc global makes no sense on mobile. Therefore,
move the logic of the currently displayed dive computer
to the profile widget and remove the dc_number global
variable.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Don't access the global current_dc, but pass it to the sensor and
tank-use delegates, when the current dive or dive computer changes.
The same pattern is already realized for the tank and weight models.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The dive-equipment tab has a number of "delegates" for editing
tanks sizes, etc. Instead of allocating them, make them subobjects.
The main point here is that, in an upcoming commit, the sensor
delegate will have to be accessed to change the current dive computer.
So far it didn't have a name and therefore was hard to access.
By making it a subobject it also gets a name.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
An attempt at limitting accesses to the globals current_dive and
dc_number. These globals do not make sense on mobile.
The parent widget of the tab-widgets remembers the currently
displayer dive and dive computer and the individual widgets
access these values from there.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Make it possible for the individual tab-widgets to access the
parent widget. In principle this could have been done by
downcasting the pointer returned by parent(), but this makes
it explicit.
The goal here is to store information on the selection,
current dive, etc. without repeating it in every subwidget.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
On selection change, pass down selection (including current
dive and dc) to the tab widgets. Ultimately, this should
remove access to global variables. A number of new accesses
are marked as TODO. They shall be removed in due course.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The signals/slot names for dive selection changes were a mess.
Unify on divesSelected(). Firstly, selectionChanged() is a Qt
thing. Secondly, it is consistent with tripSelected().
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This was very weird: a setSelection() call was always followed
by a selectionChanged() call, though sometimes in convoluted
ways. Notably, the formed was called by the DiveListView, the
lattern then by the MainWindow.
Let's just merge these two functions.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
In DiveListView user actions (select-all, key-press,
mouse-release) were intercepted to send the selection-changed
signal if the selection changed.
However, with the recent cleanups, this can be done
directly from selectionChanged(), since in all cases (at
least the ones I tested), the part of the function that
is responsible for manual selection changes is called
only once.
This avoids quite some complex code flow.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
When manually selecting a trip, the selectionChanged()
virtual function was manually selecting the dives of the
trip and thus ultimately recurse into itself.
So far this seems to work OK, but better to avoid this
recursion by setting the programmaticalSelectionChange
flag.
I'd like to send the selection-changed signal directly
from selectionChanged() and this recursion would lead
to double signals.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The trip selection code was an awkward layering violation.
Whereas dive selections due to dive undo-commands trickled
down via DiveTripModel-->MultiFilterSortModel-->DiveListView,
for trip editing, the DiveListView directly intercepted the
TripEdited signal.
Instead, mimic the dive-selection code. This is a bit longer
but more consistent and logical. The undo/redo of trip changes
is now also a "programmatical" change of the selection.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
After sending a selection-change signal, there follows a current
dive changed signal. Combine these two into a single signal, since
usually the current dive is changed when the selection is changed.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Instead of accessing the global dc_number from the
DivePlannerPointsModel and the CylinderModel, pass them
in the respective initialization functions.
The dc_number global might not make sense on mobile.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
These were not optimal, because they would recalculate the current
dive and divecomputers for every invocation.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
delete_single_dive() is one of those remnants from before the
undo-code. Now it is only called in two contexts:
1) When clearing the whole dive log.
2) When importing dives from the cloud on mobile.
In the first case, the selection is cleared before deleting
the dives.
In the second case, let's just do the same.
Thus, we can remove the last call to the deselect_dive()
function that does some complex calculations concerning
the current dive and divecomputer.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Each of these calls recalculates the current dive and divecomputer.
Instead, collect the dives to be selected/deselected and (de)select
them at once.
This needs some code refactoring in the core, because we need a
function that
1) doesn't send a signal by itself.
2) doesn't clear the trip-selection.
This contains some reorganization of the selection functions
signatures: The filter code is the only caller that keeps the
selected dive and the only caller that cares about whether the
current dive changed. So let only the function that keeps the
selected dive return whether the current dive changed.
It's all very fragile.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
For each selected dive that is hidden by the filter,
unselect_dive() was called, which led to a recalculation
of the current dive and divecomputer.
Instead, collect all deselected dives and deselect them
at the end. Thus, these calculations are performed
only once.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This tries to encapsulate the management of the current dive and
divecomputer in the selection code. The current dive is alreay
set by setSelection(). Add a new parameter to also set the
current divecomputer. If -1 is passed, then the current
computer number is remained. This will allow us to audit the code.
Because for now, the whole "current dive computer" thing seems
to be ill-defined.
This fixes a bug: the dive-computer number wasn't validated
when making a new dive the current dive. The new code has some
drawbacks though: when selecting a whole trip, the validation
will be called for all dives in the trip and thus the dive computer
number will depend on the dive with the lowest amount of dive
computers in the trip. This will need to be fixed.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This removes a constant describing the length of the array.
The enumerated_range code had to be adapted, because the
interaction of C-type arrays with the C++ typesystem is mad.
With C-type arrays, one has to pass a reference to std::declval.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The parser API was very annoying, as a number of tables
to-be-filled were passed in as pointers. The goal of this
commit is to collect all these tables in a single struct.
This should make it (more or less) clear what is actually
written into the divelog files.
Moreover, it should now be rather easy to search for
instances, where the global logfile is accessed (and it
turns out that there are many!).
The divelog struct does not contain the tables as substructs,
but only collects pointers. The idea is that the "divelog.h"
file can be included without all the other files describing
the numerous tables.
To make it easier to use from C++ parts of the code, the
struct implements a constructor and a destructor. Sadly,
we can't use smart pointers, since the pointers are accessed
from C code. Therfore the constructor and destructor are
quite complex.
The whole commit is large, but was mostly an automatic
conversion.
One oddity of note: the divelog structure also contains
the "autogroup" flag, since that is saved in the divelog.
This actually fixes a bug: Before, when importing dives
from a different log, the autogroup flag was overwritten.
This was probably not intended and does not happen anymore.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Strangely enough, half of the infrastructure was
already there, it just wasn't hooked up to a UI
element
Signed-off-by: Robert C. Helling <helling@atdotde.de>
Some DCs only report water type, without salinity level. Subsurface
fixes most of these cases using default levels, but when the type of water
is Sea/Salt, this fix was not saved.
This causes a bit confusion, mainly if the user defines own salinity level.
Signed-off-by: Rafael M. Salvioni <rafael.salvioni@gmail.com>
Fix a bug causing the bluetooth address not being used when downloading
from a 'remembered' dive computer if the device selection is populated.
This specifically excludes MacOS, as 'remembering' bluetooth connections
does not seem to be working there as per
https://github.com/subsurface/subsurface/pull/2158#issuecomment-508933672.
I am not super sure why we are _not_ trying to use the 'remembered'
device if the device selection is populated (`ui.device->currentIndex()
== -1`) - maybe this should be clarified in a comment?
Signed-off-by: Michael Keller <github@ike.ch>
Since the only caller was C++ code, this can be done in
C++ code, which removes memory-management headaches.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Make the time edit respect the configured time format. Also make the
date and time format change when the preferences are changed.
Fixes#3849.
Signed-off-by: Michael Keller <github@ike.ch>
Add titles for the individual topics covered in CONTRIBUTING.md. One
advantage of this is that it makes it possible to link to individual
topics when providing feedback on pull requests.
Signed-off-by: Michael Keller <github@ike.ch>
Format keywords, classes, string examples, etc. in prose of
CODINGSTYLE.md using `inline formatting` with single backticks.
Signed-off-by: Andrei Rybak <rybak.a.v@gmail.com>
Similarly to commit 062533d0a (CONTRIBUTING.md: use dashes instead of
hyphens, 2023-01-14), replace hyphens in prose of CODINGSTYLE.md with
dashes to make the rendered Markdown a bit nicer.
Signed-off-by: Andrei Rybak <rybak.a.v@gmail.com>
The `for` loop in the sample iterates over `serviceUuids`, but the only
other statement in the sample assigns variable named `l`. Fix the name
of the variable in the assignment to be the same as in the loop.
Signed-off-by: Andrei Rybak <rybak.a.v@gmail.com>
Without an blank line, the pseudo-headers of list items are not
separated from the next paragraph. An example is rendered as:
* variable declarations In C code we really ...
instead of intended:
* variable declarations
In C code we really ...
Add missing blank lines between paragraphs inside list items and in
between list items to fix the intended rendering.
Signed-off-by: Andrei Rybak <rybak.a.v@gmail.com>
Samples of code in CODINGSTYLE.md are wrapped in triple-backticks to
render them as blocks of code. For code blocks that are indented within
a list item, Markdown renderer of GitHub treats the leading tab as if it
was four spaces. Rendered code blocks are formatted in a way that
contradicts the code style written down in prose.
Replace leading tabs in indentation with spaces in blocks of sample code
to render them correctly in Markdown lists.
Signed-off-by: Andrei Rybak <rybak.a.v@gmail.com>
In order to support development of the open source firmware of the
OSTC4.
Requires changes in libdivecomputer.
Signed-off-by: Michael Keller <github@ike.ch>
