feat(migration): add type renaming to resolve native asset name conflicts

[Rom1-B]

• Replace : feat(migration): add type renaming to resolve native asset name conflicts #472

---

• I have performed a self-review of my code.
• I have added tests (when available) that prove my fix is effective or that my feature works.
• This change requires a documentation update.

Description

• It fixes !42658
• It is not possible to migrate an asset from the genericobject plugin if its name is similar to an asset in the GLPI core. This fix allows users to rename an asset so they can continue the migration.

Screenshots (if appropriate):

Conflict:

Migrate

[stonebuzz]

Something doesn't seem to be working correctly in the renaming process.

Let's take the KCM object as an example:

I want to rename it to KVM.

After saving the change, GLPI displays a success message indicating that the operation completed successfully:

However, after refreshing the page, I can see that the object has not actually been renamed and is still displayed as KCM.

I then retry the exact same operation, and this time it succeeds:

The object is now correctly renamed to KVM:

It looks like the first attempt returns a success message even though the rename is not actually applied. The change only takes effect after performing the exact same operation a second time.

[stonebuzz]

If a plugin (in this case, Fields) references these objects, it breaks the migration process.

Traitement du plugin « fields »...
Installation des tables MySQL

In DBmysql.php line 417:
                                                                                                                                                                 
  MySQL query error: Table 'sopra.glpi_plugin_genericobject_kcms' doesn't exist (1146) in SQL query "SHOW COLUMNS FROM `glpi_plugin_genericobject_kcms`". 

[stonebuzz]

Hello,

It appears that the foreign keys associated with the renamed GenericObject objects were not renamed accordingly.

This seems to match the issue described in the following support ticket:
https://support.teclib.com/front/ticket.form.php?id=42658

[stonebuzz]

waiting customer feedback

[stonebuzz]

oninput (line 259) triggers a fetch request on every keystroke without debounce. Add a ~300 ms debounce to glpiGenericobjectUpdateCounter to avoid a burst of requests during fast typing.

[Rom1-B]

done

[stonebuzz]

maxlength is set to MAX_TYPE_NAME_LENGTH (25), which caps raw input characters. System names expand digits to words (1 → one), so a 25-char input like a111… produces a system name far longer than 25. The AJAX counter shows the real length, but maxlength gives false assurance — the form submits successfully, then the server rejects it with no inline feedback.

The simplest solution would be to ban numbers and use only letters when renaming.

[Rom1-B]

The real length is already validated both live (counter/warning turn red before you hit 25) and server-side (renameType() rejects and shows an explicit error message if you submit past the limit anyway), so there's no silent failure here. maxlength is just a soft cap, not the source of truth.