Automatically convert event descriptions to/from HTML

[NeilRashbrook]

Needed for future calendar work:

• When sending invitations, cancellations and responses, we need to be able to send the description in plain text as well as HTML
• When creating teams meetings, OWA will generate the description in plain text by default, which we will need to upconvert to HTML for the event editor

[benbucksch]

I feel uncomfortable with the subtle differences between Abstract/Message and Calendar/Event .

In email, we're writing .html = , here rawHTML. a possible reason might have been that we never write emails, but we do have to sync back cal events, so maybe that's why you read from rawHTML when syncing to server, and then you write to rawHTML for consistency? This kind of thing would be appreciated as code comments, or as PR comment.

In Message, we clear text in the caller, whereas in Event, the descriptionHTML clears the .text property in the setter.

1. Could you please explain the reasons for the discrepancies (if they are intentional)?
2. what's the plan to make it consistent? (i'd prefer a separate PR for email changes)

[benbucksch]

Needs try/catch, see Abstract/Message.ts

[NeilRashbrook]

Needs try/catch, see Abstract/Message.ts

That's unhelpful, as I don't understand why there's a try/catch there in the first place. (Particularly as there isn't one in the text code, which also performs sanitisation. If that's a mistake on your part, then congratulations, you successfully confused me.)

The front end edits descriptionHTML - obviously this needs to read the sanitised version, but also when the editor changes the value, we need to reset the text automatically; I don't know whether that's possible within the editor. I made the regular text and description accessors clear each other automatically for this reason.

On the other hand, the network and storage may have access to both text and HTML formats and would want to be able to process them separately, so they use the raw accessors for consistency.

[benbucksch]

why there's a try/catch there in the first place

Because sanitizeHTML() and convertTextToHTML() can throw.

[NeilRashbrook]

Are those libraries that bad? Have you filed bugs? (Also convertHTMLToText calls sanitizeHTML, and there's no try/catch there.)

[NeilRashbrook]

I had to rebase the branch because of your merge conflict, but I added the try/catch in a separate commit.

[benbucksch]

1. Could you please explain the reasons for the discrepancies (if they are intentional)?
2. what's the plan to make it consistent? (i'd prefer a separate PR for email changes)

Hey Neil, coud you please address these points?

[NeilRashbrook]

See #449 (comment)

[NeilRashbrook]

what's the plan to make it consistent?

That depends on whether you agree with my reasoning for doing it the way I did.