Please check if this bug still exists in the latest patch release. Thank you.

Same issue for me

Related to condition ? Number compare ? or double scale ?

Capture d’écran du 2021-03-03 14-57-54.png (59,969 bytes)   

Capture d’écran du 2021-03-03 14-57-54.png (59,969 bytes)   

Update survey {Q1_SQ001_0.NAOK } is not updated via javascript

And the proof

Hm, should not be set to "feedback"...

Is this broken on LS3 too? Can't import the LS4 survey there. -.-

The HTML didn't change much since 2018-05-08, so I guess the error must be somewhere else.

LS3 survey with same problem.

The HTML didn't change much since 2018-05-08, so I guess the error must be somewhere else.

It work in 3LTS

It work in 3LTS

Yeah, but it was already converted to Twig in LS3.

LS3 survey with same problem.

Not for me (with gif)

Peek 09-03-2021 15-05.gif (91,416 bytes)   

Peek 09-03-2021 15-05.gif (91,416 bytes)   

Diff between LS3 and LS4 HTML is trivial:

< <col class="dsheader" />

<col class="dsheader" style='width: {{ cellwidth }}%;' />

21c21
< <col class="dsheader" />

<col class="dsheader"  style='width: {{ cellwidth }}%;' />

28c28
< <th class='left-header'>{{ processString(leftheader) }}</th>

        <th class='left-header'>{{ leftheader }}</th>

30c30
< <th class='right-header'>{{ processString(rightheader) }}</th>

        <th class='right-header'>{{ rightheader }}</th>

46c46
< {{ processString(ansrow.question) }}

                {{ ansrow.question }}

72c72
< {{ processString(ddprefix) }}

                    {{ ddprefix }}

94c94
< {{ flatString(processString(lrow.title,1),1) }}

                        {{ flattenText(lrow.title) }}

106c106
< {{ processString(ddsuffix) }}

                    {{ ddsuffix }}

130c130
< <div class="ddprefix ls-input-group-extra">{{ processString(ddprefix) }}</div>

                <div class="ddprefix ls-input-group-extra">{{ ddprefix }}</div>

143c143
< {{ flatString(processString(lrow.title,1),1) }}

                        {{  flattenText(lrow.title) }}

155c155
< <div class="ddsuffix ls-input-group-extra">{{ processString(ddsuffix) }}</div>

                <div class="ddsuffix ls-input-group-extra">{{ ddsuffix }}</div>

Not for me (with gif)

I mean, with same logic. :) Yes, it works in LS3.

Here ?
https://github.com/LimeSurvey/LimeSurvey/blob/12cd8e5a2ca9975a3a45a0b10ca0630344d55e7d/assets/packages/expressions/em_javascript.js#L225-L240

$(this).attr('name') !== $(this).attr('id')

$aQuestionsWithDependencies is empty.

/* select/dropdown item */
$(document).on("change",".select-item select:not([onchange]),.dropdown-item select:not([onchange])",function(event){
    checkconditions($(this).val(), $(this).attr('name').replace("#", "_"), 'select-one', 'change')
});
/* radio/button item */
$(document).on("change",".radio-item :radio:not([onclick]), .button-item :radio:not([onclick])",function(event){
    checkconditions($(this).val(), $(this).attr('name').replace("#", "_"), 'radio', 'click')
});
/* checkbox item */
$(document).on("change",".checkbox-item :checkbox:not([onclick]),.button-item :checkbox:not([onclick])",function(event){
    checkconditions($(this).val(), $(this).attr('name').replace("#", "_"), 'checkbox', 'click')
});
/* hidden item */
$(document).on("updated",".answer-item :hidden",function(event){
    checkconditions($(this).val(), $(this).attr('name').replace("#", "_"), 'equation', 'updated')
});

work : you can assign to me

markdown disable ?

$aQuestionsWithDependencies is empty in LS3 too, checking method GetRelevanceAndTailoringJavaScript.

Did you test https://bugs.limesurvey.org/view.php?id=16610#c62944 ?

Did you test https://bugs.limesurvey.org/view.php?id=16610#c62944 ?

Not yet, one sec...

I do :)

Reason : name (and column) of dual scale are {SGQA}{SQ}#0 and {SGQA}{SQ}#1 , but # broke in ID, then replace by {SGQA}_{SQ}0 and {SGQA}{SQ}_1

Dual scale are the only one with # in name (and column)

This assume "_" are not allowed in question codes, right? Or answer codes.

This assume "_" are not allowed in question codes, right? Or answer codes.

… since years (1.92 i think)

https://github.com/LimeSurvey/LimeSurvey/blob/12cd8e5a2ca9975a3a45a0b10ca0630344d55e7d/application/models/Question.php#L198

We allow bad values in survey done before (if no update of question code)

checkconditions() seem to behave identically in both LS3 and LS4. :|

Guess I could compare how ExprMgr_process_relevance_and_tailoring is generated. :d

@ollehar : seriously ?

The difference is only in javascript … for array_dual_scale
Check https://github.com/LimeSurvey/LimeSurvey/blob/d603a29c5bde04c72c1061eb1b985168c7635e54/application/helpers/qanda_helper.php#L5560

Did you try checkconditions($(this).val(), $(this).attr('name').replace("#", "_"), 'select-one', 'change')

And see https://github.com/LimeSurvey/LimeSurvey/blob/d603a29c5bde04c72c1061eb1b985168c7635e54/assets/scripts/dualscale.js#L22

Denis

Did you try checkconditions($(this).val(), $(this).attr('name').replace("#", "_"), 'select-one', 'change')

No, because this change is not needed in LS3.

LEMval('222593X8X95SQ003#0.NAOK') works in LS3 too but not in LS4.

Thanks for the links.

Oh, you fixed this in 2013 already...

I think we lost https://github.com/LimeSurvey/LimeSurvey/blob/d603a29c5bde04c72c1061eb1b985168c7635e54/assets/scripts/dualscale.js#L22 file in 4.X

Yes, Markus forgot it. I pushed it now to branch bug/16610-conditionals-dual-scale.

Real fix would be to remove # from question name/id, I guess. :d

Real fix would be to remove # from question name/id, I guess. :d

Unsure : my opinion is : use same name than column. Then need to remove the # from column too.

Maybe we need to remove $() and replace it with document.getElementById() then?

@DenisChenu You have time to test my commit in the branch bug/16610-conditionals-dual-scale?

Maybe we need to remove $() and replace it with document.getElementById() then?

? why ? There are currently issue in 3.15 with expression manager using getElementById

And here : we use $(this).attr('name') in javascript event.

i know HTML5 allow # in id , the n we can get $(this).attr('name') === $(this).attr('id') but we need to update all Expressionmanager class …

? why ? There are currently issue in 3.15 with expression manager using getElementById

To avoid collision with # sign.

In 3.X

The active element have name with #
All other element have and not # : id (with value) for each input and for the hidden input used by Expressionmanager javascript

The action was done by dualscale.js
The global on action in em_javascript start in 2015 : https://github.com/LimeSurvey/LimeSurvey/commit/ef6d25f8bacfce6bf3cf1030aeddc9dc7478ae3c#diff-f618aa50447d9c6aa9e3f5967cdb5e410275588d1ef5194613c66217e791f502

For information in 2.6lst : we have a lot of 'onclick="checkconditions($(this).val(), $(this).attr('name'))"' in HTML source.
This make javascript workaround more complex to do …

Expression Manager function seems near same in HTML source of 3.X and 4.X
The input name to used by EM name is done in LEMalias2varName array https://github.com/LimeSurvey/LimeSurvey/blob/610446dac91b6544058ecf5e914c7ac009a79be7/assets/packages/expressions/em_javascript.js#L979-L980

Then:

1. Speed and easy solution #1 : fix and add dualscale.js
2. Speed and easy solution #2 : fix emjavascript to replace # by : make a decision about id contain only _
3. Complex solution (and potential bug) : move from $("#ID") to getElementById … (and unusre it fix this issue)

:D

Capture d’écran du 2021-03-10 15-27-48.png (91,477 bytes)   

Capture d’écran du 2021-03-10 15-27-48.png (91,477 bytes)   

Capture d’écran du 2021-03-10 15-28-00.png (88,070 bytes)   

Capture d’écran du 2021-03-10 15-28-00.png (88,070 bytes)   

PS : id have _ already

Capture d’écran du 2021-03-10 15-34-16.png (85,418 bytes)   

Capture d’écran du 2021-03-10 15-34-16.png (85,418 bytes)   

That's great, but can you also please test my commit? Or read and approve it?

Oups …
Yes sorry :)

Another issue …

But same in 3.X !

Fix is OK

Thanks.

Fix committed to master branch: http://bugs.limesurvey.org/plugin.php?page=Source/view&amp;id=31310

Did you open a Dev mantis for jquery VS javascript AND id VS name ?

Nope. I will start looking at the 192 bugs marked as partially blocking now.

Fixed in Release 4.4.13+210315