On Tue, Nov 09, 2021 at 12:12:42PM +0100, Theo Buehler wrote: > I'm ok with that. Maybe we should bump even higher, say 8M? Isn't the > main point to have a limit at all? I don't think it needs to be very > tight.
I think it is best to bump the limits to accommodate real world observations. 8 MB objects are likely to require multiple minutes of compute time to decode and marshall around. I'm inclined towards frugality :-) Kind regards, Job
