OpenSesame not responding for ~10 mn at launch on Windows 7
Hi all,
I have the latest stable version, 3.3.9 Lentiform Loewenfeld (Python 3.7), installed on a Windows 7 system. Since this system is used by several investigators, it usually has no Internet connection to avoid unwanted updates.
Unfortunately, when I launch OpenSesame offline (using conda : https://osdoc.cogsci.nl/3.2/download#anaconda-cross-platform), the window launches but then OpenSesame does not respond for ~10 mn. When I launch it online, this issue does not occur.
As suggested in another post (https://forum.cogsci.nl/discussion/2877/opensesame-3-1-extremely-slow-to-start-the-first-time-and-run-experiments-the-first-time), I tried disabling and removing folders of plugins that can try to connect at launch (OpenScience Framework, analytics, update_checker). Yet, this did not change anything.
Do you have an idea of what I can do to improve the offline launch time, or of how I can monitor the launch to see what is causing this issue ?
Thanks for your help.
Best,
Jessica

Comments
Hi @Jessica Bourgin ,
Since you mention launching 'offline' and 'online', I'm guessing that you don't mean that OpenSesame itself takes this long to start, but that the freeze happens when you start an experiment. Is that correct?
There are many reasons why this might happen, but to trace it down, I would try the following:
-- Sebastiaan
Check out SigmundAI.eu for our OpenSesame AI assistant!
Hi @sebastiaan,
Thanks for your reply. Sorry I wasn't clear, I mean that OpenSesame is long to start, not an experiment.
What happens is the following :
I looked at the Anaconda prompt and did not see any warning or error message appear.
Thanks for your help.
Best regards,
Jessica
Hi @Jessica Bourgin ,
If you pass the
-dcommand line argument (opensesame -d), then OpenSesame will write debugging output to a log file, the location of which will be indicated in the terminal. You should be able to tell where the delay comes from by looking at the timestamps in the debug log. Basically, if you see that two lines are separated by 10 minutes, then that's where things go wrong. Could you do that? You can also just attach the debug log to this discussion, so that I can take a look.— Sebastiaan
Check out SigmundAI.eu for our OpenSesame AI assistant!
Hi @sebastiaan,
Thanks a lot for your response. I just tried with that and it seems that the webbrowser widget is responsible (see debug file attached):
I couldn't find information about this widget. Is it something I can delete or deactivate safely?
I noticed another difference when I run OpenSesame with and without an Internet connection. When I run it offline, the following message appear every 100 ms or so:
Do you know what this means and can it impair the performance while the experiment runs?
Thanks for your help.
Best,
Jessica
Hi @Jessica Bourgin ,
It seems that the delay comes from something that involves the webbrowser component. This is used by various extensions, including the
analyticsextension, theget_startedextension, and theafter_experimentextension (but this is not done at startup). Can you try disabling theget_startedandanalyticsextensions both separately and together to see whether any of these three combinations resolves the issue?Do you know what this means and can it impair the performance while the experiment runs?
It means that the Python kernel in the Jupyter Console keeps crashing and being restarted. And yes, this takes a bit of CPU and therefore could potentially cause delays.
In general, I suspect that there's something on the system interfering with network connections, and that this results in all the things that you run into (which all seem related to network connectivity). A pragmatic workaround for now would be to use
opensesamerun, which has none of these bells and whistles.— Sebastiaan
Check out SigmundAI.eu for our OpenSesame AI assistant!
Hi @sebastiaan,
Thanks a lot for your answer!
Unfortunately, disabling
analyticsorget_started, or both did not solve the issue.Since I don't need specifically to code on this computer, but indeed to run the experiment, I will do as you suggest and use
opensesameruninstead ofopensesame. Thanks for the advice!Best,
Jessica