Commit graph

133 commits

Author SHA1 Message Date
Dirk Hohndel
626a149b7c undo infrastructure: collect the texts for the executed commands
This creates a single string with all the undo texts that are on the stack.

Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2020-03-10 09:25:57 -07:00
Dirk Hohndel
053356c49b build-system: remove pointless distinction between desktop and mobile
It's the same sources, the same library.

Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2020-03-10 09:25:57 -07:00
Dirk Hohndel
41072cf48f core/undo: provide access to the QUndoStack
This is used by the mobile UI to get notifications when the undo
text changes.

Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2020-03-10 09:25:57 -07:00
Berthold Stoeger
57b96490b2 mobile/undo: create EditDive command
Command that just swaps two dives. This is rather complex,
as for example a dive site might be created.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2020-03-10 09:25:57 -07:00
Berthold Stoeger
93bdaa9bb5 undo: implement ApplyGPSFixes undo command
This gets a list of dives with GPS fixes and
1) Adds new dive sites if the dive hasn't a dive site set
2) Edits the location of the dive site

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2020-03-09 12:41:57 -07:00
Berthold Stoeger
c0095f690f mobile/undo: compile undo commands and call undo command for deletion.
First steps towards full undo on mobile.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2020-03-09 12:41:11 -07:00
Dirk Hohndel
7d9e907681 debug: try to capture changes that don't invalidate git cache
At least in those cases where we are sending a divesChanged signal we can
easily check if the cache was properly invalidated. Of course this won't help
in cases where we don't notify the dive list about changes, either.

Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2020-03-07 18:47:34 -08:00
Berthold Stoeger
2a97934db4 Cleanup: Move stringToList to core/qthelper.cpp
The same code was used in desktop and undo commands. Let's unify.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2020-03-06 10:00:13 +01:00
Berthold Stoeger
c495a49b1b Cleanup: use getDiveSelection() for shifting time
There are two cases where dive-times are shifted: in an explicit
dialog and when editing the date/time of a dive.

In each of these cases, the selected dives were collected manually.
Instead use the getDiveSelection() function. Since this returns
a std::vector, change the argument of Command::ShiftTime() to
such a std::vector.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2020-03-06 10:00:13 +01:00
Berthold Stoeger
ee553e059d Filter: move actual filtering loop to core/divefilter.cpp
The DiveFilter class defined the showDive() function to test
whether a dive should be filtered or not. This was used in
DiveTripModel to loop over all dives or all dives affected by
an editing action.

This restricts us in how we do filtering: We can't use indexes
that give us directly the result. To make the filtering more
flexible, move the actual loops that do the filtering to
the DiveFilter class.

The undo-commands likewise called directly the showDive()
function to check whether newly added dives are shown.
Use the new interface here as well.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2020-02-22 15:18:31 -08:00
Berthold Stoeger
e761d00ddd undo: don't add dive to null-trip
In moveDiveToTrip(), the dive was first removed from its old trip
and then added to the new trip. This function is also used to
remove the dive from its trip (by moving it to the "null-trip"
if you whish). Even in that case add_dive_to_trip() was called.
The only reason why this didn't crash is that add_dive_to_trip()
checks whether old and new trip are the same. This is the case
when adding to the "null-trip", since the dive was removed from
the trip just before.

To cut a long story short, to trust on add_dive_to_trip() not
crashing if moving from the null-trip to the null-trip is
way to subtly. If we remove a dive from its trip, don't call
add_dive_to_trip() in the first place.

Reported-by: Dirk Hohndel <dirk@hohndel.org>
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2020-02-19 19:49:30 -08:00
Berthold Stoeger
8391d926c7 Cleanup: remove const bool parameters and return types
These just make no sense. Since the value is copied, it
has no meaning to the caller whether the function can
change the value (and vice versa for return types).

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2020-02-09 12:13:18 -08:00
Berthold Stoeger
9ed886e4be Cleanup: lower-case filenames in core/subsurface-qt/
We tend to use lower-case filenames. Let's do it for these files
as well. Simple search & replace.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2020-02-04 02:16:46 +01:00
Dirk Hohndel
e1cd055111 code cleanup: add empty table structures
It seemed to make sense to combine all three types in one commit.

Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2020-01-10 02:37:03 +09:00
Dirk Hohndel
e9dcac7514 code cleanup: replace use of toStdVector()
Frankly, I find the old API easier...

Also, replace toList() with values()

Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2020-01-10 02:37:03 +09:00
Dirk Hohndel
24e02d878d code cleanup: remove unused member variable
It was initialized, but never referenced.

Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2020-01-10 02:37:03 +09:00
Dirk Hohndel
57f5d4b784 code cleanup: replace deprecated fromStdVector() method
Replacement was only introduced in Qt 5.14

Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2020-01-10 02:37:03 +09:00
Dirk Hohndel
c3614040d9 code cleanup: ensure all fields are initialized
In a total abundance of caution, make sure we don't exit the constructor
leaving fields uninitialized.

Fixes CID 351437

Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2020-01-10 02:37:03 +09:00
Dirk Hohndel
e6210aafc5 code cleanup: initialize all elements in structure
This seems silly. I don't like that warning and would rather disable the warning.

Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2020-01-10 02:37:03 +09:00
willemferguson
d2cf58e07e core: read and write the user-specified salinity
Both XML and git storage are added.

Signed-off-by: willemferguson <willemferguson@zoology.up.ac.za>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2020-01-06 07:00:34 +09:00
Berthold Stoeger
f4f9ba9263 Selection: provide convenience function for single-dive selection
Currently, selecting a single dive or deselecting all dives was
quite awkward: One had to pass in a single-dive vector and the
dive itself (as current dive). Provide a convenience function
that selects a single dive or deselects all dives if null is
passed in.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2019-12-07 10:23:44 -08:00
Berthold Stoeger
e453525afb Translation: explicitly show no dive-count if only one dive edited
We relied upon the translators to remove the parenthesis in cases
like "Edited notes (%n dives)" for n = 1 dives. Dirk doesn't want
that. Therefore, do it in the C++-code.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2019-12-05 10:14:25 -08:00
Berthold Stoeger
72c6b83866 Undo: make weight editing undoable
Implement the EditWeight undo command. Since there is common code
(storage of the old weight), this creates a common base class for
RemoveWeight and EditWeight. The model calls directly into the undo
command, which is somewhat unfortunate as it feels like a layering
violation. It's the easy thing to do for now.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2019-12-05 10:14:25 -08:00
Berthold Stoeger
ab99ca85f1 Cleanup: use free_weightsystem function instead of explicit free
Instead of freeing internal data of the weightsystem structure,
call the free_weightsystem function (which has to be made extern
at first). This makes things more future-proof, should the
weightsystem struct ever be extended.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2019-12-05 10:14:25 -08:00
Berthold Stoeger
dc67876c79 Cleanup: introduce empty_weightsystem constant
To make things more future-proof, introduce an empty_weightsystem
constant. Replace explicit aggragate initialization of empty
weightsystems by this constant.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2019-12-05 10:14:25 -08:00
Berthold Stoeger
b3f530bfb9 Undo: make weight-deletion an undoable action
This one is a bit more complicated than weight adding, because the
multiple-dive case is not well defined. If multiple dives are selected,
this implementation will search for weights that are identical to the
weight deleted in the currently shown dive. The position of the weight
in the list is ignored.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2019-12-05 10:14:25 -08:00
Berthold Stoeger
147a36647c Undo: make adding of weights an undoable action
Introduce an AddWeight undo command. This is modelled after the
numerous dive-edit undo commands. The redo and undo actions are
connected to the WeightModel via two new signals, weightAdded
and weightRemoved.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2019-12-05 10:14:25 -08:00
Berthold Stoeger
7787bfbf9e Selection: move commands/command_private.* to core/selection.*
The file command_private.cpp had functions concerning selections
only. To make these functions accessible from outside the undo
machinery, turn it into a part of the core-library. Currently,
only C++ functions are exported. We might think about also
exporting a C interface.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2019-12-04 13:00:23 +01:00
Berthold Stoeger
a6fa6cdb41 Undo: make profile-editing undoable
Recently, undo of dive-replanning was introduced. Therefore,
it appears logical to do the same thing for editing of the
profile of manually added dives.

For now, use the same undo-command, just change the displayed
text from "replan dive" to "edit profile". Move the fixup dive
call into the undo-command.

Eventually, every action on the profile should be made undoable.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2019-12-03 20:55:58 -08:00
willemferguson
6d7f26f4bf Desktop: add additional star widgets to Information tab
Connect the UI to the underlying dive structure. Enable proper initialisation
and management of star widgets while Information tab is active. Enable undo for
the addtional star widgets.

Signed-off-by: willemferguson <willemferguson@zoology.up.ac.za>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
2019-12-03 20:37:57 -08:00
Berthold Stoeger
b76f207158 Filter: split out filter from model
Split out the actual filtering from the MultiFilterSortModel.
Create a DiveFilter class that does the actual filtering.
Currently, mobile and desktop have their own version of this
class, though ultimately we may want to merge them.

The idea here is that the trip-model and undo-commands have
direct access to the filter-function and thus can take care
of keeping track of the number of shown dives, etc.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2019-11-19 21:13:40 -08:00
Berthold Stoeger
6d6d10f03a Filter: move calculation of shown dives to undo command
The filter-model was catching dives-added / dives-deleted signals
from the models to keep track of the number of shown dives.

To simplify the data flow, do this directly in the undo-command.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2019-11-19 21:13:40 -08:00
Berthold Stoeger
5e29245e68 Refactoring: move undo commands to top level
In the future we might want to use undo-commands for mobile as
well (even if not implementing undo).

Therefore, move the undo-command source from desktop-widgets
to their own commands top-level folder.

Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
2019-11-14 21:02:07 +01:00