Mobile build -786 small test

[Miika Turkia]

On Sun, Feb 7, 2016 at 11:54 PM, Dirk Hohndel <dirk at hohndel.org> wrote:
> On Sun, Feb 07, 2016 at 10:19:16PM +0100, Jan Mulder wrote:
>> A little round of testing. Version -786-release as from the dailies by Dirk.
>> Used hardware: Wileyfox running Android 6.0.1 latest nightly Cyanogenmod. So
>> not loaded with all kind of vendor bloatware.
>>
>> Overall performance very ok. No crashes. Selecting long dives (without gas
>> data, so skipped interpolation as I mentioned earlier due to a very recent
>> commit) at a very acceptable speed. Editing some dive-notes; worked ok.
>
> That's promising!
>
>> One thing that wonders me (without looking at any code at this point), is
>> the very slow save when editing a dive. Multiple seconds, for effectively
>> writing a file of less than 1Kb. I verified the cloudstorage on filesystem
>> level, and see that only the small saved file has a updated timestamp,
>> suggesting that no other files are accidentally saved. Tested this on a full
>> logbook of approx 750 dives.
>
> Also worth some profiling, I guess. IIRC the way the git save works is
> that it creates the whole git tree in memory and then saves what is not on
> disk. So that may cause the long duration with 750 long dives. But I would
> need to spend more time to remind myself of how exactly this is done...

I just noticed that data size for Subsurface-mobile was 2.55 GB on my
tablet. On my laptop the log's git tree is about 0.5 GB. Unfortunately
I have not figured out how to inspect this on Adnroid device to see
where the additional 2 GB has gone. I did an un-install so now the
size should be sensible, but let's see if it will increase over time.
I have mostly re-installed the APK with "adb install -r
bin/QtApp-debug.apk" so that might be the cause...I wonder if updating
from Google store is much different.

miika