DenisChenu wrote: Like i say : for reporting an issue

Why all the videos and own server, generic theme requirements? It's a JavaScript issue (recalculating style, reapplying style, CPU till 100%). Or the survey file itself contains something.

I can reproduce that issue (testsurvey via demo.sondages.pro/135149 ) in
Firefox and Chrome (both Windows 7 64bit) as well.





LimeSurvey will need to start pushing updates via different tiers. The take it or leave it approach will render installations useless for productive work. And debugging bugs become more difficult, because so many commits are released at once.

1. my demo website is on a old server (more than 11 years of services)
2. my demo link didn't use a core template (myu template have a lot of JS too , can disable but …)

Then LimeSurvey GMBH can always say it's my server or my template … and don't fix the issue.

I think the best way is to use LimeSurvey GMBH server, but here we need to have a «it's better before» …
You can use my 2.6lts for it's better before, but for 3.X issue : use a clean setup on a clean server, OR limesurvey GMBH server (not demo because demo is wiped ech night).

You can use shnoulle.limequery.com/135149?newtest=Y&lang=en for reporting.

PS : strangely : i don't have such big issue on Firefox/Linux or Chromium/Linux : linux POWA \o/

A test with demo.limesurvey.org show the same issues with the LSS export provided.

thanks for discussing the topic !

i set up a fresh install with original vanilla-theme (ajax deactivated) for testing purpose:

p511325.mittwaldserver.info/limesurvey/index.php/135149?lang=en

can you reproduce the time-lag ?


axel

256 MB of memory is fairly low, in my opinion, especially if you want to run bigger studies. LS creates realtively big session files and your memory might fill up quicker as you think.

But this should not be your problem with the loading. Yes, it took quite a while for the whole thing to load. While the question showed up relatively quickly, it took quite a bit until I could start filling the form.

holch wrote: 256 MB of memory is fairly low, in my opinion, especially if you want to run bigger studies. LS creates realtively big session files and your memory might fill up quicker as you think.

Session files are not filling up the memory by default, but the storage.


As showed a few times the issue is in the javascript theme routines. If survey creators monitor their surveys they will see a dropout on questions with this problematic behavior. Mobile device browser will be even more slower (performance, memory limits). The root cause is unknown to me. The question is, if it's an issue inside the provided lss export or a generic questiontype issue.

jelo wrote:
As showed a few times the issue is in the javascript theme routines. If survey creators monitor their surveys they will see a dropout on questions with this problematic behavior. Mobile device browser will be even more slower (performance, memory limits). The root cause is unknown to me. The question is, if it's an issue inside the provided lss export or a generic questiontype issue.

The lss-export is a standard-export from a standard-installation - nothing special in it . Same for the provided link to the test-installation: p511325.mittwaldserver.info/limesurvey/index.php/135149?lang=en

Unfortunately i cannot setup a 2.X installation shortly. With the lss-export and the link, the problematic behaviour in 3.X should be able to monitored/investigated. Anyone who could help?

Thank you !

holch wrote: As showed a few times the issue is in the javascript theme routines. If survey creators monitor their surveys they will see a dropout on questions with this problematic behavior. Mobile device browser will be even more slower (performance, memory limits). The root cause is unknown to me. The question is, if it's an issue inside the provided lss export or a generic questiontype issue.

I spoke with one user who have 100 line array with a lot hidden, and there are issue.

We think of a solutioin to deactivate javascript checking, doing only PHP checking for all expression.

I think it can be done disable the default events on the item.

And can be (sometimes) a solution. Not here since javascript must be done …

Two additional questions

• As there seems to be no quick solution for versions 3.X, I plan to stay with Version 2.X: Do you consider version " limesurvey2.6.7-lts+171212" still as secure or are there known vulnerabilities in it?
• Could you explain "I think it can be done disable the default events on the item." Is there a possible workaround which i could test?

Thank you !

silverlime wrote: As there seems to be no quick solution for versions 3.X, I plan to stay with Version 2.X: Do you consider version " limesurvey2.6.7-lts+171212" still as secure or are there known vulnerabilities in it?

Yes, there are, even in github.com/SondagesPro/LimeSurvey-SondagesPro i'm late on security checking (must look at security fix on 3.X)

silverlime wrote: Could you explain "I think it can be done disable the default events on the item." Is there a possible workaround which i could test?

The javascript checking for action is done here : github.com/LimeSurvey/LimeSurvey/blob/ma...avascript.js#L26-L48

If questionXXX don't need js checking , only PHP checking, i think you can use jquery.off : api.jquery.com/off/ to disable the event check.