No lss or lsa ?
No way to reproduce …

lss attached
The problem is for example for questions 2703 and 2740

Any update about this?
Thanks

From what we have been reviewing the behaviour is happening, but it is to highlight that we believe it may only be happening to questions which are not relevant for the given response (which were not shown during the survey taking process).

Can you confirm this only happens to non-relevant questions?
What kind of export are you using? Individual export or mass response export?

Unfortunately this is not the case.
Some of the answers not included in any report are of mandatory questions.
The problem affects both the Responses -> Survey responses -> View response details function and the export in any format.
I attach the example from one of the submitted answers that are affected by the problem.
The answers stored in the database are shown in the image.
Thanks for the support.

Please, can you include an LSA and point us to a specific response where this is happening?

lsa attached.
I had to import the survey into our testing environment to anonymize tokens so sid and other ids have changed.
Responses affected by the problem are for example 37, 41, 42 but there are many ore.
Not all the answers are "missing" the same values.
Data is correctly stored in the database.

Ohh I understand now!
For specific questions, the response is on the DB, but is showing empty on the individual response view, right?

I thought the row for the whole question was not shown (which could happen if question was not asked).

The response is shown on the queXML export.

image.png (24,587 bytes)   

image.png (24,587 bytes)   

Sorry for not being able to fully explain the problem.
The problem is both in viewing a single response and in the export of the responses in all formats.
I have to provide the export of all responses in csv or Excel format.
Could you please further investigate the problem?
Thanks

I think the is because of the "<" sign used on the answer option.
Can Iask you to remove it and try again? (Just a sa temporary workaround)
Or use "<" as a replacement for "<" (please, try it on a test env)

Thanks very much for the hint.
Changing the answers options, display and export of responses is now working.

Glad workaround works.
Still an issue!

This is a tough one.

The problem is the use of strip_tags.
Responses like "<1 portion/sett" are returned empty, because strip_tags leaves them empty.
That is being used widely.

The question then is: what kind of data is expected on that field? Html? NonHtml?
If we are expecting Html data there, then the "<" should be encoded, which is not.

Then, we can say "great, let's encode it" or have the html editor to encode it
Then the "<" sign, wouldn't be caught by the strip_tags. as it wouldn't be part of a tag.

The issue is that field, the answer options one, is plain text.
People usually don't enter html there.
Most people wouldn't know.

Then we should think about adding an HTML editor.
But there is not enough space!
We can't add a single line html editor!

The we can say, "ok, lets just encode whatever they type in the plain input text".
But if we do that, how would people enter HTML?

OK, maybe we can allow html only by using the popup editor.
But then, people who is not aware of that and already has html in their answeroptions may get their answer options twisted if they just hit save.

So, what we should do?
Hope that what I expose makes sense

@c_schmitz @DenisChenu

You can add < only without XSS activated only. With XSS activated : reencoded as & gt ;

I already ask for a feature to force XSS for superadmin (by options) ? Or maybe a "validate and fix HTML but allow script …"
And sometimes :

We have same issue (but with XSS too) with <img src="myimage.png" alt="myimage"> …

I really think we don't have a clean solution working always …

Else : please no HTML editor …

Maybe : export code if final value is empty ?

I still don't understand the problem. Cannot reproduce because the example survey were deleted.

@c_schmitz, please check testing instructions at https://bugs.limesurvey.org/view.php?id=14743.

It's not exactly duplicate .

14743: is for USER anwser (must be encoded)
This is for Single select HTML part : encode when export ?

From #14743, it says talking about answer options.

At our survey we are asking for age and there is an answer option "< 1".

So...

14743: is for USER anwser (must be encoded)

Looks the same to me. Or at leaset very similar.

1 (free text) User answer : use see and entre strong
2 (single choice) User click on strong

1 Data is strong
2 Data is A1 , answer text is strong

It's clearly NOT the same issue. You can not fix both on same way.

Free text must be encoded in all case : the user ENTER this
Answer text must be filtered

This can be picked up after 18307 and see how exports are working.

This issue is still happening. To be picked up.