with the adittion of gpslistmodel/location, the libraries
qt-models had a direct dependency on subsurface-core, and
subsurface-core had a direct dependency on qt-models, this is
bad.
Moving a bit of code around I'v managed to clean this out, and
also to clear a bit of uneeded code (GpsTracker and gpsTracker where
basically the same thing.)
Signed-off-by: Tomaz Canabrava <tomaz.canabrava@intel.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
When starting ssrf-mobile, and the local divelist is identical to the
cloud version (so same sha1), a save of local modifications to the
cloud should be possible. This fix enables this by pretending
that a suppressed load from cloud on start is still a valid load
from cloud.
Signed-off-by: Jan Mulder <jlmulder@xs4all.nl>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
So far the user can't interact with these. We should implement an ability to
visualize the GPS fix and to delete it.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
We already keep those units around so the backend HTML exporter can match them.
So it makes perfect sense to also apply them in the mobile app.
Fixes#987
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
A couple of members could potentially have been used uninitialized - it's
possible that this was the cause for people overwriting data in cloud storage
with empty dive lists.
See #985
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This fixes two issues. In general, after edits the dive list wasn't updated so
it showed data inconsistent with what the dive details showed (clearly bogus).
Even more annoyingly, when we change the date or time of a dive it could
obviously move around in the dive list. So we need to resort the dive table and
recreate the dive list. For really long dive lists this is possibly overkill,
but in my testing this seemed very quick and much easier than trying to
manually get this right, even in the case where the list wasn't resorted.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This is a bit more complicated because we are asking the user to edit the text
field instead of giving them a date and time picker. This is not a great
choice, but let's run with it for now.
One downside is that the user is likely going to edit the date "Oct 29" -> "Oct
25" without adjusting the day of the week. And if we then try to parse that Qt
correctly complains about an invalid date. So we hack around this by removing
the day of the week from both the format and the date entered (which of course
now will break things if the user did, in fact, adjust the day of the week). As
I said, not a great solution.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
If we loaded the dive list from cache and then try to figure out if the remote
repository had anything different, we were being super stupid if the SHA was
identical... we had already cleared the dive list by the time we decided that
we didn't need to load things. Granted, the model was still populated (oops),
but the backend data structure was cleared and accesses to it (e.g., when
drawing the profile) would cause things to crash.
The helper function duplicates some code, but trying to not duplicate the code
made things even harder to read.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This gives the user a bit more information about the progress of setting
the cloud credentials. IMO, especially the information that the
credentials are invalid is crucial for the user experience.
Signed-off-by: Miika Turkia <miika.turkia@gmail.com>
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This still isn't the user experience that I want, but at least now it's not a
static map image anymore but an interactive map.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This isn't quite perfect yet. If it takes too long to get the GPS fix (i.e., if
you save it before you get the fix), this will simply fail and not store a
position. But in normal conditions (you check the box, you edit the data, you
save), especially when outside on a dive boat, this should work fine.
For the other cases we need to implement some kind of callback to still collect
the data.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Getting closer to being able to really edit / add dives in the mobile UI.
This works for manually added dives - needs a bit more thought for dives
downloaded from dive computers as we don't necessarily want to change the
maxdepth in conflict with the samples.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This is a little hacky as it changes the units based on user input - a little
crude but works in the typical cases.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This is a simple way to deal with a "no network" situation. But this isn't
ideal, yet, if there is a slow network as the dive list will be reset again
once the sync from the cloud finishes. So there is some more thought needed to
make this work "mostly as expected".
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Instead of doing the silly "onEditingFinished" we get the strings from the QML
components at the time we commit the change. Much more logical, much more
straight forward, no issues with the TextArea not having an onEditingFinished
signal.
This still has a few open todos: the temperatures aren't parsed, the edit
screen is missing depth and duration, we can't edit the dive time (and it isn't
passed in on the commit). But it's progress.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Now we at least start out with the corret date, time and number. This still
isn't functional as a lot of the data aren't used and the way you save the data
is completely silly, but it's another step in the right direction.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This mainly happens because add dive is completely broken right now, but in
general it seems to be good policy not to blindly dereference this pointer...
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
I'm not sure this is the best way to do this - QML should be able to get to the
model data directly (I hope?). But this seems to work and I need it to make Add
Dive be semi-correct.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This is of course stupid and NOT what we want to do, but one could argue it's
better than nothing (well, not sure, whatever). If we have a GPS location
associated with a dive and you tap on the location name when showing the dive
details, it opens a static image of a satellite map with a marker for the dive
site.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
At least one tester cannot retrieve their web user id. This should help us
collect more data and figure out why this fails.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Now the message should make more sense. First it tells you that it's looking
for dives. Then you get some progress during the git download, and error
messages if things failed.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
So this has a lot of caveats:
- right now it only works for buddy, divemaster and suit
- you have to actually exit the field with your cursor or the change
doesn't take - that's ridiculous, there must be a far more clever way to
do this
- because I use the onEditingFinished handler I can't do this for the
Notes (so here's another reason why I KNOW that this is the wrong way to
do this)
But it shows in principle how this could be done and once someone who
actually knows what they are doing gets their hands on the code I'm
optimistic that this can be morphed into something much more useful.
It does tie together the changes made in the previous commits so that both
clicking around on the dive list gives the expected results and synching
the data back to the cloud actually works.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Well, at least it doesn't what you think it does. Let's use our little
helper to actually compare the strings.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
We really need to verify that the credentials are valid before trying to access
our backend resources. Trying to do so in a clean manner caused quite a bit of
changes to how we retrieve the webservice userid and how we load the dive list
from cloud storage.
So instead of accessing the network resources directly, this adds a handler
function that first checks the validity of the credentials (by using the
rederict handler on the cloud server), and only calls the function that does
the actual work (looks up the web service userid, loads the dives) if that
succeeds.
Right now there is no good user feedback mechanism - this just gets logged on
the log page. But this is a massive improvement if there are issues with
network connectivity or if the user mistyped their credentials.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
The logging to the UI didn't work anymore since the message area had been
removed in commit 8646934ba3 ("Simple DiveList as initial Page").
This way all the updates simply land on the Log page.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This makes for a much more friendly first use experience:
Open Subsurface-mobile, enter your cloud credentials, tap on Save and you see
your dive list.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This prevents people from overwriting a perfectly fine repository with an empty
one. Typically happens when you first enter your cloud credentials and then
don't Load Dives right away.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
get_error_message() clears the error message in the process, so calling it
twice in a row does not do what you might think it does.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
Instead of directly using the status output for the QML UI, set up the
function used to display messages to the user as part of the constructor.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
This should actually not be in the mobile section at all. This needs to be
available on the desktop as well.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
It's just not user friendly to have two different user IDs for two
different web services that we provide. Instead in the following commits
we'll add a way to retrieve the location service web service userid with
your cloud storage user id.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
With this Subsurface-mobile should be able to mostly replace the companion
app. This needs some more testing and fine tuning (for example the minimum
time / distance should be configurable, there should be a location name),
but I think the hard part is done now.
Signed-off-by: Dirk Hohndel <dirk@hohndel.org>
