The first is easy : can do it.
Second one is more tricky

Thinking:
EM can send :

• if group is relevant or not (with question too)
• If group as some badly answered question

Then:

• Group index option can have
  • with/without relevant index
  • Unclickable group depend on maxtstep
  • Unclickable question depend on last "ALl OK" group

TODO: think same for question (but question are todo actually)

@c_schmitz & @sammousa : please : did I add an the 4 options ?

But then did i set it with multi checkbox or only one selector ?

Group relevance can not be applied to the full question index.
Since group relevance might depend on previous filled in questions / groups.

Basically if you want to have a full group index displayed you must not use relevances that cross group boundaries.

Alternatively we could make an exemption for the relevance equation 0 since that's always irrelevant.

Yes, maybe.

I do some test: i think we can have a real jumpTo with "NO DB save" (because actually this one are not used really).

1. save laststeps
2. EM:jumpto(maxsteps,nosavedb)
3. Look at session group
4. EM:jumpto(laststep)

AFter the session index are filled with all group, stop at first "not validate"

It work, but then i'm sure we can break Server memory because we jump at last page for each user each time.

Hello, I'm having the same problem described here or in 08932, althought in my case, the groups are hidden by a very simple relevance equation based on a prefilled question, by URL; question ubicated in the very first group of the survey.
I am using something like:
localhost:8080/limesurvey/index.php/123456/lang/es/newtest/Y?Rol=R02
In the relevance equation of a group controlled by that question/parameter, the condition included is like:
Rol=='R02'

In 08932, c_schmitz indicated <<Fixing this would mean that all questions needs to be evaluated at all times. This would be way to slow so this is a no-fix. >>. However, as the reporter signaled, perhaps it could be treated as an additional one-time evaluating at the first time the survey loads.

Couldn't be added as a limited, special case of the question index evaluation? Perhaps ignoring (assuming a certain default value) the questions corresponding to step > maxstep. This would also work, I think, with questions prefilled by URL if they belong to an initial group.

Although the logic associated to group's relevance equation could be very complex, the situacions where arises this problem with the full index are probably simpler. If not, we should have used incremental index, isn't it?

Thank very much
Daniel

Hello DenisChenu,
I checked this with the latest version and could not reproduce, so this is most likely fixed for good.
Therefore, I am closing this issue. If you still can reproduce the issue using the latest version, please feel free to re-open the issue.
Thank you!

galads

Think it's more a misunderstanding :)

The screencast :)

If you set group relevance to 0 : group are shown in index if maxstep < step.

There are some issue too with "mandatory question" : user click on last group: but are push to last mandatory question.

This one seems OK

I got it now. Thanks, Denis, I will add it to the backlog

Maybe it's more a new feature about index, or a new plugin system.