RAM consumption on commit fa0087aeb0970e5483f507418019a033e58c1fa3 #101
Comments
|
Hi again, |
|
Finally, waveform freezes, LED green still continuous blinking. Memory still increasing |
|
Thanks for the detailed analysis. I think it is the dataanalyser class' fault. |
|
I have found the issue. The newly introduced |
|
Hi David, |
|
@voloviq Can you please retest? I thing I fixed the ram issue |
|
Hi David, mw@yoga510:~/temp/openhantek_latest/openhantek/build$ OpenHantek I used valgrind to get failure place, mw@yoga510:~/temp/openhantek_latest/openhantek/build$ valgrind OpenHantek To summary, I was unable to see mainwindow at all. Please investigate code to find a leakage. I compiled version from commit e6c6723 |
|
@voloviq I was able to verify the problem and I actually introduced the runtime exception on purpose, because the scopes record length was set, before the available record lengths were requested. |
|
Hi David, now is better I mean OpenHantek start working, I can see waveform main window. Nevertheless when I connected to channel 1 to test generator 1kHz. I was unable to see the square waveform. I see only one line at bottom one line at the top and from time to time one transition. Regulation of timebase, sample rate and frequency base don't give demanded effect. I encounter two times segmentation and one time mw@yoga510:~/temp/openhantek_latest/openhantek/build$ OpenHantek I think there is still something to rework. BR |
|
Unfortunately I can only test with a 6022, which bypasses a lot of the code base, because it doesn't support hardware triggers or features offset commands or internal relays. I will try to find the problem. Was it working when the RAM consumption was high? Or did I break it with that commit? |
|
The Cannot queue arguments of type 'std::vector' message should be generated by this signal/slot pair: solution is to place into registerDockMetaTypes(). I am not sure if my observations are somehow related to this issue, but since one of the latest commits brightness of the waveform spontaneously changes... Might be a manifestation of race condition. |
|
Race conditions shouldn't be possible with the current design. I think it might be related to a color settings override bug. |
…ork with signal/slots. Fixes #101
Signed-off-by: Martin <Ho-Ro@users.noreply.github.com>
Hi David,
I check latest version commit fa0087a. Everythings works fine but after 1hour or 1 and a half Openhantek consumed all of my RAM memory I mean all 8Gbytes. I discovered that because on my desktop I have installed screenlets meter applet with ram usage option sets. Of course after close application system release more than 80% of ram.
So please check maybe some objects are not released and increase continuously.
Besides that issue, everything seems to works well.
Thanks a lot
for Your job
BR
Voloviq
The text was updated successfully, but these errors were encountered: