trying to make progress with Subsurface-mobile

Jan Mulder jlmulder at xs4all.nl
Mon Mar 20 11:53:33 PDT 2017


On 19-03-17 23:34, Dirk Hohndel wrote:
>>> b) when you tap on a dive that's later in time, it sometimes renders ALL
>>> of the dives (in the background) and seems to be hung for a long time
>>> (depending on how many dives are in your dive list, I guess), before
>>> it finally shows the right one (REALLY bad)
>> Indeed a nasty one that reproduces on my phone. For me, It hangs for about 40 seconds and than the app
>> is killed; logcat shows out of memory, and inspection with top shows indeed steady growth in memory use.

Ok, a little report from my search today for this "enumeration issue". 
The summary is: no relevant progress, just a bigger mystery.

My build environment: Arch Linux, fully on latest stuff like Qt5.8, and 
latest Android NDK and SDK. Testing it on the desktop (the mobile 
build), and on a 7.1.1. Android phone, so also very recent.

Tried the following:

1) bisected through the latest small set of mobile changes. Issue 
reproduces even without any of them.

2) ok, further back to 4.6.3-0, so 78 commits before the version on 
Google Play. Issue reproduces on my build, but not on the 4.6.3-78 from 
the play store. So, the official build is different from mine. Good idea 
would be also to tag the mobile pushes to Google Play in the repo.

3) went further back to v4.6.2 and v4.6,1 and the enumeration issue is 
there as well in my builds. Like I said, a bigger mystery.

Basically, I can only conclude that despite identical source trees, same 
Kirigami 1.1 (but different Qt/NDK/SDK) the reason for the enumeration 
issue is still unclear, and I totally
do not understand why we see this issue now, and not long ago.

best,

--jan


More information about the subsurface mailing list