This is the equivalent of std::as_const (which requires c++17) or
qAsConst (which requires Qt 5.7), neither of which we have
as minimum dependancies.
By wrapping implicitly shared Qt containers in qgsAsConst we can
safely use c++ for ranged loops instead of Q_FOREACH. (Since
Q_FOREACH's future is shaky)
See https://www.kdab.com/goodbye-q_foreach/ for further details
on why for causes a detach for Qt containers and why Q_FOREACH
is being removed from Qt.
Now that our minimum VS studio version allowed supports std::round,
we should use that in place of Qt's qRound method.
Because:
- it doesn't truncate to int, resulting in unpredictable
behaviour (refs #16925)
- better to stick to standard c++ methods wherever possible,
since they're likely better supported and optimised by the
compilers
- it's a tiny reduction to the barrier for entry to QGIS
development (I'm sick of pointing out the need to use
qRound during PR reviews!)
And use it when we need to clone parameters (instead of more fragile
conversion to and from variants)
This fixes model loading which use algorithms which create python
subclasses of parameter definitions
because using this more consistently throughout the codebase makes it
easier to maintain code.
We also do not want to call the copy constructor on them, using pointers
just makes this more obvious. Further, casting is also something
that's commonly done on pointers and not references.
And if you want a value or a reference, just use QgsGeometry, it's meant
to be handled like this.
1. endless loop when a sqlite-based layer is opened, and
wal+shm files are created. This triggers a loop because
the directoryChanged signal is emitted again and again ...
This was the real blocker.
2. when a new files appears in a directory
the directoryChanged is emitted and OGR/GDAL may fail
to open the file because the file copy was not yet
finished.
This commit fixes 1 but the fix for 2 is only a best effort using
a 100 ms timer: I could not find a definitive solution to this issue,
a file could be legitimately opened in streaming mode and it will
always triggers an error, there is apparently no way to get
the QFileSystemWatcher emit a signal when new files are closed
flusehd and not when they are just created.
