Closed jwmerrill closed 2 years ago
Update: it turns out the original benchmark wasn't actually exercising renderLatexMathEfficiently
because it doesn't run in case you call the public API .latex()
method with a number rather than a string.
After updating the benchmark, I'm measuring renderLatexMathEfficiently
at 12ms/100 updates and renderLatexMathFromScratch
at 41ms/100 updates.
Adds a new benchmark to measure performance of updating a decimal mathquill. To run the benchmark, do
This benchmark shows that
renderLatexMathEfficiently
is no longer a measurable win, and it has been a source of bugs recently.For reference,
renderLatexMathEfficiently
landed in https://github.com/desmosinc/mathquill/pull/73