I was using database session management. I changed it back to file based session management, and I *think* the behaviour isn't manifesting itself any more.

Any one have a guess why db sessions would trigger this behaviour, but file based wouldn't?

I have this exact behaviour randomly on a short one-page survey when user submits the survey. Page is the redrawn with all the responses (17 tickbox) cleared. On second submission this tends to happen less.

Which Limesurvey version did you use?
Which browser was used?
Does it happen with all browsers you tested?
Do you use Ajax mode at theme options?

Same problem here too on azure web app:


LimeSurvey version: 3.15.6
LimeSurvey build: 190108
Operating system: Windows NT RD0003FF43FB00 10.0 build 14393 (Windows Server 2016) i586
PHP version: 7.0.32
Microsoft-IIS/10.0
Web server info: HTTP/1.1 HTTP/2.0
Database driver: sqlsrv
Database version: 12.x

Try disabling AJAX in the survey theme options.

tpartner wrote: Try disabling AJAX in the survey theme options.

I'm still trying to find out, why "disabled" is not the default setting? LS 3 and disabling AJAX mode is the solution to everything

tpartner wrote: Try disabling AJAX in the survey theme options.

AJAX already disabled

jelo wrote:

tpartner wrote: Try disabling AJAX in the survey theme options.

I'm still trying to find out, why "disabled" is not the default setting? LS 3 and disabling AJAX mode is the solution to everything

We are disabling the Ajax mode at our templates by default due to lots of different issues related to that.

What is it actually useful for? What's the benefit? What's the technical difference?

Mazi wrote: We are disabling the Ajax mode at our templates by default due to lots of different issues related to that.

I totally disable ajax mode in SkelVanilla,

I think a real ajax mode must load ONLY updated part (this need big update in PHP not only HTML and JS), current ajax mode load whole page (with head, body etc …) and try to update partial HTML.


But this way : browser must recalculate whole page, sometimes it can make more time to update a page than to clean up whole page, and redisplay (this happen without ajax mode).

I have another case in the German forum.
When answering a LimeSurvey question with a video via the Samsung Internet Browser the error is shown when clicking the forward / submit button.

What is the trigger of the error routine "Please use the LimeSurvey navigation buttons or index."?

Might be good to be able to deactivate this error routine per survey / per question.

Even if the issue is triggered by a bug in the browser it might be better to let the survey flow.

jelo wrote: I have another case in the German forum.
When answering a LimeSurvey question with a video via the Samsung Internet Browser the error is shown when clicking the forward / submit button.
…

Even if the issue is triggered by a bug in the browser it might be better to let the survey flow.

Without ajax mode ?
A bug on browser can not really do this (we never have such error before 3.X).

If it's happen without ajaxmode : i'm happy to have the survey and the vodeo .
Some error with expression manager can do this maybe …

DenisChenu wrote: Without ajax mode ?

Yes, even without AJAX. But the issue is limited to the Samsung Internet Browser.

Looks to me like the detection of the browser button usage is messed up after the videoplaying component got the focus. I would like to see a way to disable this detection routine. If most of your survey participants use such a browser it would be a good thing to make LimeSurvey a bit less clever.

I translated the important parts from the German thread and will inform the threadstarter about this ticket too. bugs.limesurvey.org/view.php?id=14666