fix: lyon survey post submit duplicate message

[BarbaraOliveira13]

🎩 Description

BUG
After a first survey submission, an 'already answered" message was displayed simultaneously with the success flash message. It is confusing.

---

⚠️ This is the expected behavior by Decidim, so I updated the corresponding tests from .to have_content to .to have_no_content line 79

---

FIX
Override decidim survey show view to fix the UX bug, with adding a condition to hide "already answered" block when success flash message in present: [line 34] && flash[:notice].blank?.

Override 'app/cells/decidim/data_consent/category.erb' and upgrade h3 -> span to fix an accessility test.

Upgrade a test who failed

Testing

1. Log in as a user
2. Go to a survey component
3. Fill and submit the questionnaire
4. ✅ Only the success flash message should appear
5. Navigate back to the survey
6. ✅ Only the "already answered" block should appear (no flash)

📌 Related Issues

[Link your PR to an issue

• Fixes #?](https://github.com/OpenSourcePolitics/intern-tasks/issues/428)
• This is a core Decidim bug, reproduced on nightly.decidim.org — an upstream issue has
  been opened: Survey shows "already answered" block simultaneously with success flash after first submission decidim/decidim#16660

Tasks

• Add specs
• In case of new dependencies or version bump, update related documentation

🛠️ If your PR introduces or updates an Override

• Add the label "override:pending" in your PR
• Add the corresponding label based on the review deadline category (when this override must be checked):
  • WHEN BACKPORT 0.29 ·
  • BEFORE MIGRATION 0.31 ·
  • WAITING FOR PR UPSTREAM ·
  • OTHER
• Add the link to the related Decidim issue or PR (if exist).

📷 Screenshots

Extra information

[Stef-Rousset]

Hello @BarbaraOliveira13 , thanks for your PR, it works well !!

I have some suggestions on the modified files:

• for the modification of CI, we should talk with @moustachu , to make a choice between using the github action or install the latest stable of chrome.
• for the override of the category cell, I think it would be relevant to explain in the description the reason of the override and the line(s) updated
• for the participatory_process_show_spec.rb test, I think it would be interesting to refactor the checks for inert and aria-expanded values with the have_css helper.‎

[Stef-Rousset]

LGTM !