Adam Porter

Hi Prot, Thanks, seems like a good idea. I'll merge this for the next version (eventually).

Hi Mark, I've done a bit of exploration, and it should be possible, but it requires some fiddling, as the built-in Emacs functions to analyze SGML/HTML/XML and find the structural...

@meedstrom Thanks for reporting that. Can you reproduce it in a clean Emacs configuration, e.g. using https://github.com/alphapapa/with-emacs.sh ? Hopefully Emacs 29 with tree-sitter will make all this much simpler.

@meedstrom FYI, the fix made for #18 might fix the problem you had with `rjsx-mode`.

The sub/superscript issue is fixed in #217. Not sure if there are other ways that still need to be addressed.

Hi Sergio, Thanks for the well-written report. FWIW I think you could solve two of these problems by applying advice to the function `ement-notify--event-mentions-session-user-p`, i.e. wrap it in the equivalent...

> Do you think it's also possible to use an advice to solve this problem, BTW? I don't see why not. Advice allows you to replace a function with your...

> Wouldn't the simple solution be just ignoring the sender field in ement-room--event-mentions-user-p and/or ement-notify--event-mentions-session-user-p? It doesn't seem like matching on that would be correct in any situation. That's used...

> ement-notify--event-mentions-session-user-p has a separate check for that, (unless (equal (ement-user-id user) (ement-user-id sender)). The problem is indeed that ement-room--event-mentions-user-p doesn't check for that, and acts as if the user...

The best idea I have is to allow the user to specify an "anti-mention" regexp, which would nullify any apparent mention. In this case, the user could set it to...