I noticed the inscriptions.
200 mSa/s and 40 MHz...
In the instructions I found 120 MHz (1 channel) and 60 Mhz (2 channels), as well as 500 mSa/s.
Perhaps it has already been explained why it is written this way and I missed it?
The real specs of the DSO2512G are up to 200MSa/s in single channel mode, and up to 100MSa/s when CH2 is enabled.
In theory there's the OC mode, which boosts the maximum samplerate to 250MSa/s and 125MSa/s, but it's unstable and glitchy, so i don't count that.
About the "40MHz" label, well, if your maximum samplerate is 200MSa/s, you can calculate what's the maximum frequency signals you can accurately represent. There was a debate about this in this same thread.
New version of the app. Changes since last version:
You're fast! Even without electricity. 
Great!

I have a few questions:
Why does the rise time measurement only work with a full cycle?
Wouldn't the measurement accuracy increase with a shorter timebase (screenshot 1)?
When switching on the 20MHz BW limitation I sometimes see more spikes (overshoot) than without. One would actually expect the opposite. (screenshot 2).
In general, however, the limitation works.
How can I recognize the advantage of the Data Buffer 2 source? More accurate measurements?
Thank you!
EDIT: In XY animation, WAV (Avg) source looks the smoothest.
I'm fixing the riseTime issue right now. I will also take a look at the 20MHz filter issue.
DataBuffer2 has up to 4x higher horizontal resolution compared with the normal DataBuffer. This is good because you have more accurate data, more detailed waveform. Because of the screen resolution, which i think is about 1200-1300 pixels from left to right of the graticule, you will not see a big visual difference on screen, but some time-based measurements should be more accurate.
Of course, when you reach 200MSa/s (or 100 if CH2 is enabled), both Databuffers will have the same detail, so the advantage of DataBuffer2 disappears.