Originally filed on https://github.com/mdn/sprints/issues/2148
Just wanted to point out some rendering issues that are caused by translations (like French language for example) which makes some selectors not valid.
Take a look : https://developer.mozilla.org/fr/docs/Web/CSS/line-height
Quoting @schalkneethling
Thank you for reporting this @DamChtlv
The problem is because of:
#Utiliser_les_transitions_pour_accentuer_les_茅l茅ments_d'un_menu ~ * pre[class*=html]' is not a valid selectorMy first guess is that it is the characters such as
茅andd'ucausing the problem
The reasons this happened is one thing. The more pressing thing is that we didn't have any test automation, that uses a real browser, that checks if pages render correctly.
Able to reproduce locally:

It looks like it might actually be characters such as ' as an id such as Relzione_tra_length_e_propriet脿_numeriche does not break it. Specifically referring to the 脿 character in that string.
https://developer.mozilla.org/it/docs/Web/JavaScript/Reference/Global_Objects/Array
So far I have only found the problem with the fr locale
I wonder, should the ID have become Relzione_tra_length_e_proprieta_numeriche (instead of Relzione_tra_length_e_propriet脿_numeriche) by the Wiki post-processing?
Yep, @schalkneethling is right. My bad. We just need to make sure the iframes don't have a ' in their ID.
https://codepen.io/peterbe/pen/vYBrOzj?editors=1111
Sorry, not iframes. These:
<h3 id="Utiliser_les_transitions_pour_accentuer_les_茅l茅ments_d'un_menu">Utiliser les transitions pour accentuer les 茅l茅ments d'un menu</h3>
Well sorry for the French :)
Rewording the ID could be a workaround
Adding the ' in non_url_safe https://github.com/mozilla/kuma/blob/master/kuma/wiki/content.py#L556 could help as well
I came to the same conclusion.
Now I'm just curious what we should do to pages that have "already been allowed" to keep the '.
Well sorry for the French :)
Don't apologize for your beautiful language! It's Kuma who should say sorry.
what we should do to pages that have "already been allowed" to keep the
'
I might have an old crawler of mine that stores the content of ?raw view on files that I could run to detect the pages having such IDs.
Once Kuma would do something differently, a manual save on those pages should be sufficient (they should not be too many)
A drawback is that I have to "choose" which section I want to crawl and I may miss some pages if I don't know the list of sections beforehand.
Another drawback is that it's quite slow to gather all the pages for all the translations for all the sections.
If possible a fallback display in any case would be nice :s
Could it be possible to reopen this one?
Apparently parenthesis have the same issues: https://developer.mozilla.org/fr/docs/Web/CSS/imagefunction (it now contains #Afficher_une_partie_(sprite))
I reworded the heading on this one to avoid ' but used parenthesis and had the same error
You're totally right! Back the drawing board on this one.
An example in English ;) https://developer.mozilla.org/en-US/docs/Web/CSS/CSS_Grid_Layout/Basic_Concepts_of_Grid_Layout
An example in English ;) https://developer.mozilla.org/en-US/docs/Web/CSS/CSS_Grid_Layout/Basic_Concepts_of_Grid_Layout
I scraped that URL locally and did reproduce and was able to prove that my PR fixed it.
I've edited 41 pages of HTML/CSS sections in fr which had a ' (I left line-height alone for now)
if my regexp is correct https://developer.mozilla.org/en-US/docs/Web/CSS/text-transform is the only other page in en-US acros HTML/CSS/JavaScript sections which has brackets in an EmbedLiveSample macro call
There is no more work to do on this. I'll close it.
Most helpful comment
Well sorry for the French :)
Rewording the ID could be a workaround
Adding the
'innon_url_safehttps://github.com/mozilla/kuma/blob/master/kuma/wiki/content.py#L556 could help as well