We had a pattern where doubles were converted to long with
lrint() and then down-cast to a narrower int type.
Because this is unwieldy, introduce a function encapsulating
this.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Now that we have defined addition and subtraction on unit
classes, let's use them in a few examples.
Yes, some of these are a bit pointless, because they are
of the kind
a.mbar - b.mbar => (a-b).mbar
However, these probably should be further simplified
by storing the result in a unit type.
This commit is mostly a proof-of-concept.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The goal here is to add general addition and scalar multiplication
functions to the unit types.
Thereto, we need a CRTP
(https://en.wikipedia.org/wiki/Curiously_recurring_template_pattern)
base class.
However, this breaks compound initialization, so we have to use
named initializers:
weight_t { 2000 } -> weight_t { .grams = 2000 }
The good thing is that this is exactly how these classes were
supposed to be used: make the unit explicit!
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The copy/pasting of dive-sites was fundamentally broken in at least two
ways:
1) The dive-site pointer in struct dive was simply overwritten, which
breaks internal consistency. Also, no dive-site changed signals where
sent.
2) The copied dive-site was stored as a pointer in a struct dive. Thus,
the user could copy a dive, then delete the dive-site and paste.
This would lead to a dangling pointer and ultimately crash the
application.
Fix this by storing the UUID of the dive-site, not a pointer.
To do that, don't store a copy of the dive, but collect all
the data in a `dive_paste_data` structure.
If the dive site has been deleted on paste, do nothing.
Send the appropriate signals on pasting.
The mobile version had an additional bug: It kept a pointer to the
dive to be copied, which might become stale by undo.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Feels natural in a C++ code base.
In analogy to other tables, this creates a struct that derives
from std::vector<>. This is generally frowned upon, but it works
and is the pragmatic thing for now. If someone wants to "fix" that,
they may just do it.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This accesses the global dive_table, so make this explicit.
Since force_fixup_dive() and default_dive() use fixup_dive(),
also move them to struct dive_table.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Before, a non-owning pointer was passed and the dive moved
away from the dive. Instead, let the caller decide if they
still want to keep a copy of the dive, or give up ownership:
In MainWindow and QMLManager new dives are generated, so
one might just as well give up ownership. In contrast,
the planner works on a copy (originally the infamous
"displayed_dive") and now moves the data manually.
This commit also removes duplicate code, by moving the
"create default dive" code from MainWindow and QMLManager
to struct dive.
Finally, determination of the "time zone offset" is not done
in POSIX, since we want to avoid calls form the core into
Qt.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
These functions accessed the global divelog make this explicit.
I'm still not happy about the situation, because these functions
access global state, such as the selection. I think these
should be moved up the call-chain.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This is a messy commit, because the "qPref" system relies
heavily on QString, which means lots of conversions between
the two worlds. Ultimately, I plan to base the preferences
system on std::string and only convert to QString when
pushing through Qt's property system or when writing into
Qt's settings.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Having this as a pointer is an artifact from the C/C++ split.
The triptable header is small enough so that we can
include it directly
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Having this as a pointer is an artifact from the C/C++ split.
The divesitetable header is small enough so that we can
include it directly.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This is a humongous commit, because it touches all parts of the
code. It removes the last user of our horrible TABLE macros, which
simulate std::vector<> in a very clumsy way.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Don't access the global trip_table in an attempt to cut down
on implicit accesses of global variables.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This had to be done simultaneously, because the table macros
do not work properly with C++ objects.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Since struct divecomputer is now fully C++ (i.e. cleans up
after itself), we can simply turn the list of divecomputers
into an std::vector<>. This makes the code quite a bit simpler,
because the first divecomputer was actually a subobject.
Yes, this makes the common case of a single divecomputer a
little bit less efficient, but it really shouldn't matter.
If it does, we can still write a special std::vector<>-
like container that keeps the first element inline.
This change makes pointers-to-divecomputers not stable.
So always access the divecomputer via its index. As
far as I can tell, most of the code already does this.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This is a hairy one, because the sample code is rather tricky.
There was a pattern of looping through pairs of adjacent samples,
for interpolation purposes. Add an range adapter to generalize
such loops.
Removes the finish_sample() function: The code would call
prepare_sample() to start parsing of samples and then
finish_sample() to actuall add it. I.e. a kind of commit().
Since, with one exception, all users of prepare_sample()
called finish_sample() in all code paths, we might just add
the sample in the first place. The exception was sample_end()
in parse.cpp. This brings a small change: samples are now
added, even if they could only be parsed partially. I doubt
that this makes any difference, since it will only happen
for broken divelogs anyway.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This allows us to use non-C member variables. Convert a number
of pointers to unique_ptr<>s.
Code in uemis-downloader.cpp had to be refactored, because
it mixed owning and non-owning pointers. Mad.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
There were a number of free standing functions acting on a
dive-site-table. Make them member functions. This allows
for shorter names. Use the get_idx() function of the base
class, which returns a size_t instead of an int (since that
is what the standard, somewhat unfortunately, uses).
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This makes memory management more simple, as not explicit deletion
is necessary.
A rather large commit, because changing QVector<> to std::vector<>
is propagated up the call chain.
Adds a new range_contains() helper function for collection
types such as std::vector<>. I didn't want to call it
contains(), since we already have a contains function
for strings and let's keep argument overloading simple.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This is a long commit, because it introduces a new abstraction:
a general std::vector<> of std::unique_ptrs<>.
Moreover, it replaces a number of pointers by C++ references,
when the callee does not suppoert null objects.
This simplifies memory management and makes ownership more
explicit. It is a proof-of-concept and a test-bed for
the other core data structrures.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Also remove the UNUSED() macro, as there were no users left.
The macro was silly anyway - there were many falso positives.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Makes the code much nicer to read.
Default initialize cylinder_t to the empty cylinder.
This produces lots of warnings, because most structure are now
not PODs anymore and shouldn't be erased using memset().
These memset()s will be removed one-by-one and replaced by
proper constructors.
The whole ordeal made it necessary to add a constructor to
struct event. To simplify things the whole optimization of
the variable-size event names was removed. In upcoming commits
this will be replaced by std::string anyway.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Since the taxonomy is now a real C++ struct with constructor
and destructor, dive_site has to be converted to C++ as well.
A bit hairy for now, but will ultimately be distinctly simpler.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
- standardise the naming;
- use it consistently;
- apply the 'samples < 50' only when putting manually added dives into
edit mode - everywhere else manually added dives should be treated as
such;
- do not show a warning before editing a manually added dive in planner.
Signed-off-by: Michael Keller <github@ike.ch>
The old code was leaking memory. Use std::unique_ptr<> for
ownership management.
This is still very primitive and divetags are kept during
application lifetime. There should probably be some form
of reference counting. And the taglist should not be global,
but attached to the divelog.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Let's use std::string in the core. Notably, I'd like to make
the numerous main() functions mostly independent of Qt. Some
things will have to remain, such as argument parsing, of course.
This changes the API: instead of returning an error code and
taking a pointer to the actual return-value, return an
std::optional<std::string>> that is set if the function succeeds.
Returning an empty string in the error case might be simpler,
but oh well...
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Quite a bit of fallout in users of this structure.
Conveniently, since git-access.cpp is now C++ we can move
some helpers from the monstrous qthelper.cpp to git-access.cpp.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
When autosync to cloud was enabled, the old code would crash,
because it synced to cloud from a signal handler, which executed
the main loop when checking the cloud connection, which deleted
the object which was causing the original signal. Or something
like that. See discussion in #4119.
To avoid such problems, send a signal via a 'QueuedConnection'
from QMLManager to itself. The slot will be called once the
signal handler terminates and the main event loop retakes
control.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
When changes need saving, the notification text was set quite
deep in the calltree in "saveChangesLocal()". I don't know why
this was put so deep in the call tree. In any case, it prevents
asynchronous saving of the data. Therefore, pull it up to
chnagesNeedSaving().
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The signals of the undo-stack can only be connected after it
was initialized. One of those cases where I would have preferred
a crash over a warning message. The mobile version is extremely
noisy!
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Q_FOREACH and foreach are anachronisms.
Range based for may cause a performance regression: it can
lead to a copy of shared containers (one reason why Qt's
COW containers are broken). However, as long as there is no
user noticeable delay, there is no point in analyzing each case.
And also no point in slapping an 'asConst' on every container
that is looped over.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
To avoid memory management woes. These shouldn't be global
variables, but let's fix that later.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
