Closed jeans11 closed 1 year ago
I haven't tested the exact memory usage lately, but this library is fast because it keeps all rules in memory in an efficiënt way. So that's indeed the instances of Rule.
Sorry chime in after closed, but the primary issue is not the memory usage during operation, it's the memory usage and time spent during initialization building of these efficient structures. What we would need would be way to pre-compute this and ship it in the package, instead of building it every-time.
Ah right, i think that could be done by serializing the data structure to a file and deserialize on initialization. Not sure if that would perform better in terms of initial memory usage, but worth a try. PRs are welcome, i won't get to it.
Hi guys!
I load / create a browscap parser (create once and reuse it) but the load increase of 2G my service memory. Is all the
Rule
that are responsible of the huge memory gap?Thanks