QML enums are a bit painful to use, so this uses poor man's emums
instead.
Basically what this changes is that a forced switch to the map doesn't
count as picking the map. That seems obviously correct, as otherwise you
could end up in a situation where a legitimate switch away from the map
is ignored.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This makes no sense and seems crazy. But it also seems to work,
For some reason with the current Kirigami version and Qt 5.13.1
selecting the map page makes the pageStack jump back to the previous
page right away. I cannot find what triggers this behavior.
Since I cannot fix the root cause, I am working around the bug. When we
select the map page we remember that fact and when a different page is
picked with the mapPage being the last page on the stack, we force the
page selection back to the map page. I can imagine countless ways in
which this could go horribly wrong - but right now I can't figure out a
better solution.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
There doesn't appear to be a reason to pop all of the existing pages from the stack.
Just on principle, only close the drawer if it was open.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This had very strange results with the current Kirigami.
Instead set the width of those pages based on our overall column width.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The old calculation was clearly bogus, we'd also get zero columns here.
Instead do a correct calculation of the number of columns and make the
resulting column width a property of the rootItem so we can refer to it
elsewhere.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The hack to remove the action button caused situations where the action
button didn't return. Let's skip that for now. All the other fixes
appear to still be needed.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Remove hidapi from manually built components and use the mxe based one instead.
Remove libzip as that is handled by mxe packages.
Update version of grantlee used to build with qt 5.13.1.
Also hide vscode files from git.
[Dirk Hohndel: combined two commits, cleaned up the commit message and removed
one now incorrect comment line from mxe-based-build.sh]
Signed-off-by: Paul Buxton <paulbuxton.mail@googlemail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The old server certificates where not recognized on some older platform,
so we hardcoded the hex digest of the valid certificate and ignored the
error.
Those certificates have been replaced last week, so there is no point to
this hack anymore - also, we should always show the SSL error, not just
in verbose mode.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The Qt documentation has specific suggestions how to build our own
version of OpenSSL in order to work on Android 5.x:
https://doc.qt.io/qt-5/android-openssl-support.html
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Commit df4fbf7699 ("Android: force different font on OnePlus devices")
inadvertantly added this hunk - let's undo it again.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Instead of deleting the Application explicitly in exit_ui(),
use a unique_ptr to have the instance automatically freed.
This is the pattern given in the Qt-docs:
https://doc.qt.io/qt-5/qapplication.html
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This guarantees that they are actually singletons: there can
only be one application-wide instantiation of these objects.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
To make this class available from QMLManager, the run_ui()
function would create the object and then set a pointer in
QMLManager. It works, but is inconsistent with the rest of
the code. Therefore, make it a classical singleton class,
which is generated on demand.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
GpsListModel was one of those "special" singletons that could
be created explicitly with new. This would make sense if a
parameter were passed to the constructor. We only passed null,
so one might as well turn that into a classical singleton with
default constructor.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
DiveListModel was one of those "special" singletons that could
be created explicitly with new. This would make sense if a
parameter were passed to the constructor. We only passed null,
so one might as well turn that into a classical singleton with
default constructor.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The model was initialized in the global run_ui() function.
Move that into the constructor of the class.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
This hacks the libzip build to not try to find zlib, because it fails
doing so on modern NDK. We just tell it that its there, and be done with
it.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Expose ftdi connection posibilities if its compiled in, not only on
Android. Its useful to be able to test it on a Linux machine to.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
Not everybody have their build dir in the same place, but we can always
find the files relative each other.
Signed-off-by: Anton Lundin <glance@acc.umu.se>
The test if we have to create gas switches wasn't yet aware
of the bailout option.
Reported-by: Dennis Arreborg <dennis@arreborg.eu>
Signed-off-by: Robert C. Helling <helling@atdotde.de>
There was a sad typo in commit eecca6aab0 ("Mobile: replace model-reset
by row-addition in DiveListModel::reload()").
Additionally, we need to also refresh the dive list so that the new
dives are shown.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
We get incorrect changes to a new screen width that causes us to try
draw to a much larger screen than we actually have. Ignore those
changes.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Different Android devices seem to have different patterns of throwing
incorrect width information at us. This seems like a really bad hack,
but for the ones I've seen so far this should give us the right width
information.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
So even setups relying on the wrapper script can do faster rebuilds.
This also cleans up a bug that made passing through the release
parameter fail in the past, and removes overly verbose debugging output.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
DiveImportedModel::recordDives() called add_imported_dives(). But that
actually consumes the dive and dive-site tables. Which in turn will
lead to an inconsistent model.
Properly reset the model by using the consumeTables() function.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The DiveImportedModel and DownloadThread used the same table
of dives and dive sites. This made it very hard to keep the
model consistent: Every modification of the download thread
would make the model inconsistent and could lead to memory
corruption owing to dangling pointers.
Therefore, keep a copy in the model. When updating the model,
use move-semantics, i.e. move the data and reset the tables
of the thread to zero elements.
Since the DiveImportedModel and the DownloadThread are very
tightly integrated, remove the accessor-functions of the
dive and dive-site tables. They fulfilled no purpose
whatsoever as they gave the same access-rights as a public
field.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Even though the returned dive is not const, the table is not
changed, as it only contains pointers.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
The plan is to make the model the authoritative source of
the imported dives. Therefore, access the number of
downloaded dives from there.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
Clearing the table in the thread leaves the model in an inconsistent
state. Don't do that.
Signed-off-by: Berthold Stoeger <bstoeger@mail.tuwien.ac.at>
