Hi, thanks Pierre for putting together this profiling application! It helped me to think about the whole thing again, and I found an elegant method that works with one regex-based split and without String.replace. It scales linear with both the number of style attributes and the number of variables to replace in the attributes, which makes it superior to the other methods in most cases:
http://dev.openlayers.org/sandbox/ahocevar/sldRenderer/examples/profilingAttributeName.html ${name} feature returned value : foo andreas' method: 16ms returned value : foo pierre's method: 21ms returned value : foo andreas' method 2: 28ms feature2 returned value : foo andreas' method: 30ms returned value : foo pierre's method: 20ms returned value : foo andreas' method 2: 26ms feature3 returned value : foo andreas' method: 51ms returned value : foo pierre's method: 23ms returned value : foo andreas' method 2: 27ms name: ${name}, age: ${age}, address: ${address}, status: ${status} feature returned value : name: foo, age: ${age}, address: ${address}, status: ${status} andreas' method: 17ms returned value : name: foo, age: undefined, address: undefined, status: undefined pierre's method: 69ms returned value : name: foo, age: age, address: address, status: status andreas' method 2: 55ms feature2 returned value : name: foo, age: 20, address: somewhere, status: single andreas' method: 36ms returned value : name: foo, age: 20, address: somewhere, status: single pierre's method: 574ms returned value : name: foo, age: 20, address: somewhere, status: single andreas' method 2: 54ms feature3 returned value : name: foo, age: 20, address: somewhere, status: single andreas' method: 65ms returned value : name: foo, age: 20, address: somewhere, status: single pierre's method: 65ms returned value : name: foo, age: 20, address: somewhere, status: single andreas' method 2: 56ms Regards, Andreas. _______________________________________________ Dev mailing list [email protected] http://openlayers.org/mailman/listinfo/dev
