Broken shortcodes on 0.19 about zola

[amtanq]

Working on this

[jcla1]

Here are some example files - the first one, with zola's current version, produces links where the final bit of the URL is the placeholder (not the whole URL). The second markdown file produces the correct links.

numerik1_ss2021.md
analysis2_ss2024.md

I've checked this again, by deleting the public directory and then doing a clean build with zola v0.19.1.

[Keats]

You had that line the markdown right? If so might be a regression in shortcodes in HTML due to the pulldown-cmark update.

[jcla1]

I am also having this issue. On a markdown page I have multiple a-hrefs and use a shortcode to link to static resources. They now, on most pages, show up as @@ZOLA_SC_PLACEHOLDER@@. Weird is though that it's not on all pages - more recently updated ones (but with the same a-href and shortcode) don't seem to suffer this issue.

[Keats]

Can you share an example md file that has the issue?

[jcla1]

For anyone using zola via homebrew: I found a nice guide that explains how to get an older version of a formula. I'm now back on v0.18.0 and that doesn't have the issue with the placeholder.

[Keats]

Ok I can reproduce it with your files. For some reasons it's the

2021-07-26: **Das Ferientutorium zur Numerik I wird vom 20. September bis zum
24. September täglich von 10 bis 16 Uhr via [WebEx][15] stattfinden. Es wird
veranstaltet von Melinda Hagedorn, die freundlicherweise bereits <a href="{{ get_url(path = "static/numerik1_ss2021/zusatzuebungen.pdf") }}">Aufgaben
zum selbststudium</a> bereit gestellt hat, die im Laufe des Tutoriums dann
besprochen werden können. Eine Anmeldung ist nicht erforderlich.**

paragraph that causes issues. If i remove the HTML tag in there it does render the rest fine 🤔

[LunarEclipse363]

You had that line the markdown right? If so might be a regression in shortcodes in HTML due to the pulldown-cmark update.

IIRC I tracked this down to 13682e1 specifically at some point so yeah.

[Keats]

If anyone is interested, since we have a repro above, we should be able to write tests and fix it somewhat easily. It could be a good first issue!

[amtanq]

Raised a fix: #2601
Edit: Better fix by @clarfonthey -> #2606