Strip HTML from event descriptions before insertion

[benthamite]

Summary

Google Calendar API returns event descriptions as HTML, containing tags like <a>, <br>, <html-blob>, <u>, <ul>/<li>, and HTML entities like &,  , etc. These are currently inserted raw into the :org-gcal: drawer, producing malformed content in Org files.

This causes problems for tools that parse Org files (e.g. org-roam interprets https://meet.google.com/abc</a> as a valid link, inserting a broken path into its database).

Changes

• Add org-gcal--strip-html function that converts HTML descriptions to plain text:
  • <br> → newlines
  • <li> → \n- (Org list items)
  • All other tags → removed
  • Common HTML entities decoded (&, <, >,  , ", ')
  • Consecutive blank lines collapsed
• Apply it when binding desc in org-gcal--update-entry

Round-trip safety

Descriptions posted back to Google Calendar via org-gcal-post-at-point are read back from the Org file as plain text, which Google Calendar accepts. No formatting is lost that wasn't already lost by storing raw HTML in an Org file.

Example

Before:

<html-blob>Click the link:<br><a href="https://meet.jit.si/Weekly">https://meet.jit.si/Weekly</a><br></html-blob>

After:

Click the link:
https://meet.jit.si/Weekly

Fixes #258.

[telotortium]

@benthamite I'm not necessarily opposed to something like this, but I want to allow people to edit their event descriptions locally, including the HTML. Perhaps a per-headline property could be set in the Org file on the events that you want to strip HTML from.

[telotortium]

See my comments - also merge the latest master into this branch.

[benthamite]

@benthamite I'm not necessarily opposed to something like this, but I want to allow people to edit their event descriptions locally, including the HTML. Perhaps a per-headline property could be set in the Org file on the events that you want to strip HTML from.

Fair. How about a user option, org-gcal-strip-html-descriptions, which strips html globally when set to non-nil (defaulting to nil)? I think the typical workflow for people who are bothered by the html in the description is to always disable it, rather than on a per-heading basis.

[telotortium]

@benthamite I'm not necessarily opposed to something like this, but I want to allow people to edit their event descriptions locally, including the HTML. Perhaps a per-headline property could be set in the Org file on the events that you want to strip HTML from.

Fair. How about a user option, org-gcal-strip-html-descriptions, which strips html globally when set to non-nil (defaulting to nil)? I think the typical workflow for people who are bothered by the html in the description is to always disable it, rather than on a per-heading basis.

It should probably at least be a per-calendar option, with a global user-customizable default, so that shared calendars that you import are not corrupted.

[telotortium]

@benthamite Also merge latest master - it has some fixes to the CI

[benthamite]

Done. I've added two user options:

• org-gcal-strip-html-descriptions (boolean, default nil) — the global default.
• org-gcal-strip-html-descriptions-overrides — an alist of (calendar-id . boolean) for per-calendar overrides.

This way users can e.g. strip HTML globally while preserving it for specific shared calendars:

(setopt org-gcal-strip-html-descriptions t)
(setopt org-gcal-strip-html-descriptions-overrides
      '(("shared-calendar@group.calendar.google.com" . nil)))

Also confirmed the branch is up to date with latest master (merged in 09fb52a).

[telotortium]

@benthamite Could you add tests please?

[benthamite]

Added tests and merged latest master:

• org-gcal-test--strip-html: unit tests for HTML-to-text conversion (tags, entities, list items, blank line collapsing)
• org-gcal-test--strip-html-p: predicate tests covering global default nil/t and per-calendar overrides in both directions
• org-gcal-test--update-entry-strip-html: integration test — HTML stripped when enabled
• org-gcal-test--update-entry-preserve-html: integration test — HTML preserved when disabled (default)
• org-gcal-test--update-entry-strip-html-per-calendar: integration test — per-calendar override works

Also merged latest master (picks up PR #279).

[benthamite]

@telotortium, please let me know if there is anything else you’d like me to do.