GitLab

Useful for debugging tile download or backend issues.

BUG: 429114

We only need the new feature for making things work without installation,
so live without that in older versions.

# Conflicts:
#	CMakeLists.txt

This doesn't cover basic JNI types yet, but the API can easily be extended
to cover those now as well.

Not relevant at this point, but will matter when eventually splitting this
out into its own framework.

This allows us to remove the last remaining usage of context properties,
and also exposes the new NetworkStatus API to QML.

This needs further ECM changes to be deployed correctly without a make
install first.

With this we can defer map downloads to when we are connected to an
non-metered network.

Android backend still to be done.

This avoids entangling app code and code that eventually should move to
KF5 further here, in particular for the upcoming network status work.

This requires making SolidExtras a dynamic library on Android for its AAR
to be picked up by androiddeployqt. For this to work without requiring a
make install first, a few ECM changes are needed though.

At least for FairEmail this fails on the attachments as we seem to lack
read permissions for those. Either way, this also still has to be fed into
the extractor engine.

Should also become useful for a generalizing JNI array functions.

This should now give all necessary building blocks to reassemble MIME
message here and feed them into the extractor.

Still needs to be further generalized to work for other types than strings.

Not usable for testing yet, but good enough to compile on non-Android.

Make this more scalable going forward, and helps to ensure this is
externally extendable in a similar fashion.

This is implicitly done right now by the usage in the app, but when moving
out of the itinerary repo to a standalone framework eventually this will
no longer be the case.

The JNI and Java parts even build on non-Android using the QtAndroid mock
API, making this less likely to break.

Allows reordering header locations, and is needed for installation
eventually anyway.

Simplifies writing test code running on the host. With a bit of extending
the JNI mock objects we might even be able to build this on the host
eventually.

Prepares this for handling different types of intents, such as content
shared by e.g. email applications. For text emails we already get the
relevant information here, attachments are still missing though.

Still missing support for decoding the multi-element EXTRA_STREAMs used
by FairEmail for attachments though.

Shouldn't matter here yet, as that type isn't used in this file, but this
will have weird side-effects when that changes.

With an unfortunately set up include search path order, the wrong one (in
either direction) can be picked up, with bizarre consequences. So better
avoid that problem entirely.