Stripping signature-like advertisement and reply quote text

Sometimes, DeltaChat fails to strip signature-like text, when it is not correctly prefixed with --.

Examples include:

Sent from my Galaxy

Sent from iPad from John

Sent from Outlook for Android/iOS

as well as “reply quote” text like the following, when it is not written in English:

On Tue, 7 Jan 2025 at 00:11, Stefano <…> wrote:

I feel like all clients could benefit from more accurate “filters” of this kind.

I am not sure how to tag this feature proposal, as I believe this concerns DeltaChat core maybe?

If you have example .eml files saved, you can directly open a core issue, at least Galaxy/iPad signatures should be possible to remove if we have examples.

Any of these messages can be localised to whatever language the sender’s client is set to, though, right?

I don’t know how the message looks like, maybe there is a way to recognize it by some tags in HTML or something like that. Need to look at actual email source to see. Worst case we can strip known spam signatures if it is really just the text added.

2 Likes

Thanks! Just opened Some signature-like text is not stripped from messages · Issue #6420 · deltachat/deltachat-core-rust · GitHub

I noticed DeltaChat does not feature EML export for single messages, nor MBOX export for collection of messages. This would have been handy for this kind of work, but I had to install a separate email client altogether instead. Shall I open a separate feature request?

I talk to people in multiple languages; I can’t set my client to all of them. But maybe I’ve misunderstood.

Same. Even people who can only speak English cannot assume everybody they text with (in English) has their client set in that language.

All I meant is, regardless of what language these pesky phrases are written in, it would be nice not to have to see them.

These are varied enough that a contact-specific function to strip out any selected literal string (or regex, for ones incorporating dates etc.) seems like the only way to do it. Not very efficient or elegant, though.

1 Like