We already check with a regex if the message only contains emoji so it is safe
to hardcode the font family as emoji for these messages.
(cherry picked from commit 58213ee3e6)
Improve the handling of text both when sending and receiving.
The main feature is to fix the linked bug (and a host of others that are unreported but similar) which is caused by the fact that we don't properly clean html. This mr does that as per the matrix spec https://spec.matrix.org/v1.5/client-server-api/#mroommessage-msgtypes. So any disallowed tags or attributes are removed and it does the special handling for certain attributes.
Additionally the functions are also designed to cover any other text formatting required, particularly fro received strings.
The receive side is covered by 2 functions `handleRecieveRichText` and `handleRecievePlainText`. The rich/plain in the function name refers to the output type not the input type (both can take plain and rich input), so `handleRecieveRichText` is called to get a string suitable to go in a rich text control and `handleRecievePlainText` for a plain control.
The functions also handle the following some of which was previously handled by `eventToString` in `NeoChatRoom`:
- Strip and reply from the string
- Format any user mentions
- Linkify links in plain strings
- Handle mxc urls in rich text (uses the new `room->makeMediaUrl` functionality from libQuotient)
- `handleRecievePlainText` also deals with markup making `NeoChatRoom->subtitle` redundant
There is also an extensive test suite which defines the behaviour and the best way to review this is probably to look at the tests and decide whether you agree with the expected output given the inputs and/or if there is any missing behaviour.
The final aim especially with the test suite is to give us a framework to make further updates in the future easier and hopefully prevent a new feature breaking old behaviour with the tests.
BUG: 463932 \
BUG: 466330 \
BUG: 466930
This patch fixes two key issues:
* `?` is considered part of the URL, even if there might not be anything past it. I've seen people now putting spaces after URLs to get around this issue.
* `)` is also part of the URL, as well as `:` despite those not being valid characters at the end of URLs.
Make sure that the emote name is added inside the first paragraph rather than as an additional one.
Note alternate to network/neochat!632 because I didn'y realise @tfella had done a patch
CCBUG: 461837
