The 5e Shaped Sheet imports data via a JSON attribute. This is parsed and then a bunch of attributes on the sheet are set, triggering the sheetworkers to distribute the data into all the various subsidiary fields. When setting data for characters with a reasonably large number of attributes, such as a Lich with all his spells, the Roll20 side of this seems to be taking a disproportionately long time. Currently, it takes between 10 and 15 seconds to produce a list of dirty attributes the first time around, before any of the sheetworkers fire. The changes made by the sheetworkers then trigger more events, in a round that again takes around 10 seconds to produce a list of dirty attributes. There are several more rounds of cascading events after this, and in total, the import can end up taking between 30 seconds and a minute. For reference, the completed Lich has just under a thousand attributes; the initial round of changes sets around 400 in one go I think. I don't know if that would be considered a lot, but the 10-15 second lag times on detecting each round of dirty attributes seems to be grossly disproportionate to the amount of work being done!
