Android beta...

[Jan Mulder]

On 06-07-17 19:33, Dirk Hohndel wrote:
> Thanks to outstanding work by Linus, UI work by Tomaz, and great support
> from Alex Blasche from the Qt team, I think we are fairly close to being
> able to launch a strong beta of Subsurface-mobile 2.0 on Android.
> 
> http://subsurface-divelog.org/downloads/test/Subsurface-mobile-4.6.4.350-arm.apk
> 
> With one patch added on top of Qt 5.9.1 for Android I can now successfully
> download from a Suunto EON Steel, a Shearwater Petrel but most importantly
> from a Shearwater Perdix AI. I also have several reports that the Scubapro
> G2 appears to work.
> 

Tested on Android 7.1.2. Version -350 as requested, so a signed one, so 
deleting my development install first.

- A funny cosmetic thing. As I have a relatively big logbook (approx 920 
dives in the cloud, with numerous multi hour dives), it takes some time 
for the initial logbook download to complete. No problem, but the toast 
message (saying ?/60, not sure where the 60 is coming from), is moving 
upwards some pixels on every refresh. So a walking toast message. Looks 
funny, but is not correct obviously.

- the "map it" button is not in the correct theme color (it is blue, 
also in pink theme). More in general. Not sure that the theme color for 
buttons would be. For example, the buttons on the download from DC are 
grey. Consistency would be nice I think, especially because it is a 
major 2.0.

- download from cloud work perfectly.

- the OSTC3 shows correctly as BT and BLE.

- download over BT works correctly. Over BLE not tested (as I know it 
does not work (hopefully, yet)).

- A probably very difficult one to debug/find, and I suspect 
QML/Qt/kirigami based. When I scroll very far down, just by swiping 
multiple times, or by selecting the scrollbar that shows only only when 
scrolling and just after (autohide, seems fine to me), the pull down 
sync gets confused. Ok, confused is a little bit unclear description, 
but still the most accurate I can think of :-) For example: 1) Not being 
able to force the pull down sync, as a pull over the entire screen (up 
to down) is not sufficient distance, 2) pull down sync is activated even 
when not on the top of the divelist. Obviously. 1) and 2) in different 
sessions.

- Still a crash on exit. Seemingly way beyond all relevant Subsurface 
processing, so no real hurt done, but still not nice.
Seems https://bugreports.qt.io/browse/QTBUG-61434

--jan