When an ISNI code is already there, it is now trying to add a new.

Thanks @zas for testing.
Isn't it supposed to add ISNI to a new field when there is already one?
Can't artists have several ISNI, in practice, like they can have several IPI, like recordings can have several ISRC and like works can have several ISWC?

❌ BTW I notice that I have indeed removed the 2 ways of systematic crashing the script but ISNI is no longer set at all.

❌ For example I have an artist with no ISRC and the script just adds a second empty ISNI field.

So I will fix my fix.

But please advise what should be done when:

• Artist already has a different ISNI
• Artist already has the same ISNI
• Artist has no ISNI

When an ISNI code is already there, it is now trying to add a new

What I meant is it tries to add the same ISNI as it was new (so it displays it as a new field, but doesn't actually add it).

About various cases:

• Artist already has a different ISNI
  

Perhaps just warn, but don't add the new one, it might be an error, and we don't know where

• Artist already has the same ISNI
  

Do nothing

• Artist has no ISNI
  

Add the one from WD (if any, of course)

I have fixed the WP URL handling (09f3a01) and the actual filling of ISNI code field from WD (582de28), in this same PR proposal.

The only problem that remains:

MBS new react-hydrate system makes it so that script added ISNI code
will not visually persist if you click "Add ISNI code" button afterwards
or things like that.

I tested these known react-hydrate-resist tricks
but they unfortunately did not work for ISNI codes:

https://github.com/jesus2099/konami-command/blob/a1dd018d84cecee965943dda19617a6807dcedc4/lib/SUPER.js#L163-L172
d4c8183

I put the PR on hold as I will try my react-hydrate workaround once more, tomorrow.
I think I made a mistake when I tried it on ISNI code!

No no, I did already test it well in 582de28 and this react-hydrate workaround doesn't work for ISNI codes, indeed.
I reactivate the PR and there is that slight issue (on some browsers, not on Android Kiwi):
Newly inserted ISNI code (automatically from WD) visually disappears, if you click the Add ISNI code button without having interacted with the new code manually beforehand.

Last update

Problem fixed by c298541 thanks to https://nattaylor.com/blog/2022/userscripts-and-reactjs-forms/
The trick was to let the input event bubbling up!

bubbles: true

Problem fixed by c298541 thanks to https://nattaylor.com/blog/2022/userscripts-and-reactjs-forms/
The trick was to let the input event bubbling up!

bubbles: true

Hi @Jerome-Roy / @loujine
This PR is now ready, I think.
It makes the userscript working, again.
I hope you’re all right, in your cold country.

Not so cold when volcanoes wake up :)

Thanks for your PR, I'll try to take a look this week. I merged some simpler pull requests tonight, ping me if I broke something

I tried to analyse the merge conflict, please tell me if it's OK for you, @loujine. :)

Remaining conflict compare

Repo master branch:

    const isniBlock = document.getElementById(
        'add-isni-code').parentElement.parentElement;
    const fields = isniBlock.getElementsByTagName('input');
    for (const input of fields) {
        existing_isni.push(input.value.split(' ').join(''));
    }
    existing_isni.splice(0, 1); // skip template
    if (existing_isni.includes(isni.split(' ').join(''))) {
        return;
    }
    if (existing_isni.length === 1 && existing_isni[0] === '') {
        document.getElementsByName('edit-artist.isni_codes.0')[0].value = isni;

Repo master branch:

    const isni_fields = document.querySelectorAll('input[name^="edit-artist.isni_codes."]');
    for (const input of isni_fields) {
        if (input.value) {
            existing_isni.push(input.value.split(' ').join(''));
        }
    }
    if (existing_isni.length === 0) {
        (Object.getOwnPropertyDescriptor(Object.getPrototypeOf(isni_fields[0]), 'value').set).call(isni_fields[0], isni);
        isni_fields[0].dispatchEvent(new Event('input', {bubbles: true}));
        $('#newFields').append(
            $('<dt>', {'text': 'New ISNI code added:'})
        ).append(
            $('<dd>', {'text': isni}).css('color', 'green')
        );

Break out

ISNI fields selector

master branch:

    const isniBlock = document.getElementById(
        'add-isni-code').parentElement.parentElement;
    const fields = isniBlock.getElementsByTagName('input');

isniBlock is not used anywhere else, later.

This PR branch:

    const isni_fields = document.querySelectorAll('input[name^="edit-artist.isni_codes."]');

These are doing the same thing, tell me if it's OK to keep this version of the ISNI selector fix, @loujine?

Template

Template no longer exists and we should remove the skip:

master branch:

    existing_isni.splice(0, 1); // skip template

Removed by this PR.

Inform the user of already existing ISNI

master branch does not inform the user:

    if (existing_isni.includes(isni.split(' ').join(''))) {
        return;
    }

This PR informed when same ISNI is already set.

Inform the user of new ISNI set

This PR branch:

        $('#newFields').append(
            $('<dt>', {'text': 'New ISNI code added:'})
        ).append(
            $('<dd>', {'text': isni}).css('color', 'green')
        );

This PR now informs when the new ISNI is being set, like for other attribute changes.

Setting value so that it is not hidden later by react-hydrate

Repo master branch:

        document.getElementsByName('edit-artist.isni_codes.0')[0].value = isni;

Repo master branch:

        (Object.getOwnPropertyDescriptor(Object.getPrototypeOf(isni_fields[0]), 'value').set).call(isni_fields[0], isni);
        isni_fields[0].dispatchEvent(new Event('input', {bubbles: true}));

Now that pages are using react, it is needed to prevent display bugs (disappearing ISNI when giving Add ISNI button), this way.

I solved conflict, keeping my ISNI fields selector.
I hope it's good for you, otherwise, I can amend last commit. :)

I'm glad to hear from you, again! :)

I tested quickly, looks good, thanks so much
The hydration fix looks like the one I needed in the common script last year