Performance of the chat are plummeting

I fear it is getting worse and worse everyday

It makes it barely usable.

Are we kind of stuck with the following dilemma?

1. unbearably slow indexDB
2. localStorage overflow

@MattJ @XaviFP A fresh indexedDB + sessionStorage before the converse initialization seems to avoid this kind of issues. Still, is it safe to do?

I refer to the comment I made after speaking with JC: #176 (comment 35145)

In summary he recommends: 1) clear history on reconnect, 2) stick with localStorage. Obviously the ultimate choice is you to you guys, and if indexedDB seems to improve things here, great.

Clearing sessionStorage entirely will mess up RAI if the client resumes an existing session (if you clear everything related to the old session, this probably won't happen).

But in principle, if clearing local caches seems to improve stuff, there is no harm in doing it.

Performances looks alright using IndexedDB + Websockets + Clear IndexedDB & sessionStorage before every initialization.

With localStorage, we usually overflow its max size on less than one session, leading to a lot of inconsistencies.

Don't know for @alexbourlier But I also feels that we have less "I-already-read-this-unread". Still, all my own messages are still marked as unread. #189 (closed)

But in principle, if clearing local caches seems to improve stuff, there is no harm in doing it.

Good to know

Well, on my side I have no unread. My circles are never bold. So... pretty consistent UX so far :-)

Closing in favor of #211 (closed)

closed