Ideas on combining table refresh with NTIE and sorting logic

[podliashanyk]

Table refresh with NTIE in #87 does not reorder events. Meaning there is no consideration towards the sorting logic (see #81). New events are simply prepended, modified events always stay where they were in the table before NTIE.

[hmpf]

Yeah the python-code should either insert changed events into the correct order or we need to redraw the entire list regularly. Curitz does the latter but has to do that due to how curses works I suspect. The code for this is in curitz itself, zinolib knows nothing about sort order.

[podliashanyk]

NB that inserting events into correct order would also require some advanced fickling with hx-swap and hx-target attributes. We might even need to define a custom event for that (JS). In short: single hx-swap and hx-target on tbody will definitely be replaced by individual ones on each row, and each row would probably listen to the custom event.

Seems easier to have python-code for figuring out the sorting order and then reswapping all of the table rows if there are any changes to sorting. As a plus this option might also work well in tandem with #97 if we land on some solution with reswapping all of the rows there also.

[podliashanyk]

Fixed by using cache in #81